librarian-bots/model_cards_with_metadata · Datasets at Hugging Face

modelId stringlengths 5 139	author stringlengths 2 42	last_modified timestamp[us, tz=UTC]date 2020-02-15 11:33:14 2025-07-13 18:27:38	downloads int64 0 223M	likes int64 0 11.7k	library_name stringclasses 518 values	tags listlengths 1 4.05k	pipeline_tag stringclasses 55 values	createdAt timestamp[us, tz=UTC]date 2022-03-02 23:29:04 2025-07-13 18:27:10	card stringlengths 11 1.01M
Ryukijano/masked-lm-tpu	Ryukijano	2023-08-30T12:30:04Z	5	0	transformers	[ "transformers", "tf", "roberta", "fill-mask", "generated_from_keras_callback", "base_model:FacebookAI/roberta-base", "base_model:finetune:FacebookAI/roberta-base", "license:mit", "autotrain_compatible", "endpoints_compatible", "region:us" ]	fill-mask	2023-08-30T07:41:30Z	--- license: mit base_model: roberta-base tags: - generated_from_keras_callback model-index: - name: Ryukijano/masked-lm-tpu results: [] --- <!-- This model card has been generated automatically according to the information Keras had access to. You should probably proofread and complete it, then remove this comment. --> # Ryukijano/masked-lm-tpu This model is a fine-tuned version of [roberta-base](https://huggingface.co/roberta-base) on an unknown dataset. It achieves the following results on the evaluation set: - Train Loss: 5.8422 - Train Accuracy: 0.0344 - Validation Loss: 5.8152 - Validation Accuracy: 0.0340 - Epoch: 48 ## Model description More information needed ## Intended uses & limitations More information needed ## Training and evaluation data More information needed ## Training procedure ### Training hyperparameters The following hyperparameters were used during training: - optimizer: {'name': 'AdamWeightDecay', 'learning_rate': {'class_name': 'WarmUp', 'config': {'initial_learning_rate': 0.0001, 'decay_schedule_fn': {'class_name': 'PolynomialDecay', 'config': {'initial_learning_rate': 0.0001, 'decay_steps': 111625, 'end_learning_rate': 0.0, 'power': 1.0, 'cycle': False, 'name': None}, '__passive_serialization__': True}, 'warmup_steps': 5875, 'power': 1.0, 'name': None}}, 'decay': 0.0, 'beta_1': 0.9, 'beta_2': 0.999, 'epsilon': 1e-08, 'amsgrad': False, 'weight_decay_rate': 0.001} - training_precision: float32 ### Training results \| Train Loss \| Train Accuracy \| Validation Loss \| Validation Accuracy \| Epoch \| \|:----------:\|:--------------:\|:---------------:\|:-------------------:\|:-----:\| \| 10.2437 \| 0.0000 \| 10.1909 \| 0.0000 \| 0 \| \| 10.1151 \| 0.0001 \| 9.9763 \| 0.0016 \| 1 \| \| 9.8665 \| 0.0107 \| 9.6535 \| 0.0215 \| 2 \| \| 9.5331 \| 0.0230 \| 9.2992 \| 0.0223 \| 3 \| \| 9.2000 \| 0.0231 \| 8.9944 \| 0.0222 \| 4 \| \| 8.9195 \| 0.0229 \| 8.7450 \| 0.0224 \| 5 \| \| 8.6997 \| 0.0231 \| 8.6124 \| 0.0219 \| 6 \| \| 8.5689 \| 0.0229 \| 8.4904 \| 0.0222 \| 7 \| \| 8.4525 \| 0.0230 \| 8.3865 \| 0.0223 \| 8 \| \| 8.3594 \| 0.0230 \| 8.3069 \| 0.0221 \| 9 \| \| 8.2662 \| 0.0231 \| 8.2092 \| 0.0224 \| 10 \| \| 8.1956 \| 0.0231 \| 8.1208 \| 0.0222 \| 11 \| \| 8.1285 \| 0.0229 \| 8.0806 \| 0.0219 \| 12 \| \| 8.0345 \| 0.0234 \| 8.0030 \| 0.0220 \| 13 \| \| 7.9960 \| 0.0228 \| 7.9144 \| 0.0224 \| 14 \| \| 7.9065 \| 0.0231 \| 7.8661 \| 0.0221 \| 15 \| \| 7.8449 \| 0.0229 \| 7.7873 \| 0.0219 \| 16 \| \| 7.7673 \| 0.0232 \| 7.6903 \| 0.0229 \| 17 \| \| 7.6868 \| 0.0242 \| 7.6129 \| 0.0243 \| 18 \| \| 7.6206 \| 0.0250 \| 7.5579 \| 0.0246 \| 19 \| \| 7.5231 \| 0.0258 \| 7.4564 \| 0.0254 \| 20 \| \| 7.4589 \| 0.0262 \| 7.4136 \| 0.0255 \| 21 \| \| 7.3658 \| 0.0269 \| 7.2941 \| 0.0265 \| 22 \| \| 7.2832 \| 0.0274 \| 7.1998 \| 0.0270 \| 23 \| \| 7.2035 \| 0.0275 \| 7.1203 \| 0.0271 \| 24 \| \| 7.1116 \| 0.0280 \| 7.0582 \| 0.0269 \| 25 \| \| 7.0099 \| 0.0287 \| 6.9567 \| 0.0287 \| 26 \| \| 6.9296 \| 0.0294 \| 6.8759 \| 0.0287 \| 27 \| \| 6.8524 \| 0.0296 \| 6.8272 \| 0.0285 \| 28 \| \| 6.7757 \| 0.0300 \| 6.7311 \| 0.0291 \| 29 \| \| 6.7031 \| 0.0304 \| 6.6316 \| 0.0305 \| 30 \| \| 6.6361 \| 0.0306 \| 6.5744 \| 0.0307 \| 31 \| \| 6.5578 \| 0.0312 \| 6.4946 \| 0.0312 \| 32 \| \| 6.4674 \| 0.0319 \| 6.4212 \| 0.0314 \| 33 \| \| 6.4096 \| 0.0322 \| 6.3557 \| 0.0320 \| 34 \| \| 6.3614 \| 0.0321 \| 6.3093 \| 0.0322 \| 35 \| \| 6.2754 \| 0.0329 \| 6.2240 \| 0.0326 \| 36 \| \| 6.2609 \| 0.0326 \| 6.2114 \| 0.0321 \| 37 \| \| 6.1866 \| 0.0329 \| 6.1645 \| 0.0320 \| 38 \| \| 6.1470 \| 0.0330 \| 6.1193 \| 0.0323 \| 39 \| \| 6.0936 \| 0.0329 \| 6.0600 \| 0.0324 \| 40 \| \| 6.0625 \| 0.0330 \| 6.0282 \| 0.0323 \| 41 \| \| 6.0062 \| 0.0335 \| 5.9649 \| 0.0329 \| 42 \| \| 5.9731 \| 0.0339 \| 5.9661 \| 0.0330 \| 43 \| \| 5.9460 \| 0.0335 \| 5.9259 \| 0.0330 \| 44 \| \| 5.9206 \| 0.0338 \| 5.8926 \| 0.0333 \| 45 \| \| 5.8734 \| 0.0343 \| 5.8471 \| 0.0340 \| 46 \| \| 5.8663 \| 0.0341 \| 5.8561 \| 0.0337 \| 47 \| \| 5.8422 \| 0.0344 \| 5.8152 \| 0.0340 \| 48 \| ### Framework versions - Transformers 4.32.1 - TensorFlow 2.12.0 - Tokenizers 0.13.3
PlaceReporter99/Utility_Bot_Chat	PlaceReporter99	2023-08-30T12:23:38Z	0	0	transformers	[ "transformers", "conversational", "en", "dataset:fka/awesome-chatgpt-prompts", "license:mit", "endpoints_compatible", "region:us" ]	text-generation	2023-08-30T12:18:00Z	--- license: mit datasets: - fka/awesome-chatgpt-prompts language: - en metrics: - accuracy library_name: transformers pipeline_tag: conversational ---
Norod78/SDXL-jojoso_style-Lora	Norod78	2023-08-30T12:22:51Z	520	3	diffusers	[ "diffusers", "text-to-image", "stable-diffusion", "lora", "en", "base_model:stabilityai/stable-diffusion-xl-base-1.0", "base_model:adapter:stabilityai/stable-diffusion-xl-base-1.0", "region:us" ]	text-to-image	2023-08-17T12:33:25Z	--- base_model: stabilityai/stable-diffusion-xl-base-1.0 instance_prompt: jojoso style tags: - text-to-image - stable-diffusion - lora - diffusers widget: - text: jojoso style cute dog - text: jojoso style picure of a Pokemon - text: jojoso style Cthulhu rising from the sea in a great storm - text: jojoso style girl with a pearl earring by johannes vermeer inference: true language: - en --- # Trigger words Use "jojoso style" in your prompts # Examples The girl with a pearl earring by johannes vermeer, Very detailed, clean, high quality, sharp image, jojoso style ![The girl with a pearl earring by johannes vermeer](https://huggingface.co/Norod78/SDXL-jojoso_style-Lora/resolve/main/Examples/00189-20230816194930-558-the%20girl%20with%20a%20pearl%20earring%20by%20johannes%20vermeer%2C%20Very%20detailed%2C%20clean%2C%20high%20quality%2C%20sharp%20image%2C%20jojoso%20style%20_lora_SDXL-jojo.jpeg) Cthulhu rising from the sea in a great storm, Very detailed, clean, high quality, sharp image, jojoso style ![Cthulhu rising from the sea in a great storm](https://huggingface.co/Norod78/SDXL-jojoso_style-Lora/resolve/main/Examples/00224-20230816201509-557-Cthulhu%20rising%20from%20the%20sea%20in%20a%20great%20storm%2C%20Very%20detailed%2C%20clean%2C%20high%20quality%2C%20sharp%20image%2C%20jojoso%20style%20_lora_SDXL-jojoso_st.jpeg)
umerah/Task3	umerah	2023-08-30T12:14:26Z	10	0	transformers	[ "transformers", "pytorch", "gpt2", "text-generation", "generated_from_trainer", "base_model:distilbert/distilgpt2", "base_model:finetune:distilbert/distilgpt2", "license:apache-2.0", "autotrain_compatible", "text-generation-inference", "endpoints_compatible", "region:us" ]	text-generation	2023-08-30T09:05:51Z	--- license: apache-2.0 base_model: distilgpt2 tags: - generated_from_trainer model-index: - name: Task3 results: [] --- <!-- This model card has been generated automatically according to the information the Trainer had access to. You should probably proofread and complete it, then remove this comment. --> # Task3 This model is a fine-tuned version of [distilgpt2](https://huggingface.co/distilgpt2) on the None dataset. It achieves the following results on the evaluation set: - Loss: 3.0246 ## Model description More information needed ## Intended uses & limitations More information needed ## Training and evaluation data More information needed ## Training procedure ### Training hyperparameters The following hyperparameters were used during training: - learning_rate: 0.0001 - train_batch_size: 16 - eval_batch_size: 16 - seed: 42 - optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08 - lr_scheduler_type: linear - num_epochs: 10 ### Training results \| Training Loss \| Epoch \| Step \| Validation Loss \| \|:-------------:\|:-----:\|:-----:\|:---------------:\| \| 3.1753 \| 1.0 \| 1741 \| 3.0775 \| \| 3.0194 \| 2.0 \| 3482 \| 3.0338 \| \| 2.9194 \| 3.0 \| 5223 \| 3.0157 \| \| 2.8401 \| 4.0 \| 6964 \| 3.0063 \| \| 2.7765 \| 5.0 \| 8705 \| 3.0064 \| \| 2.7266 \| 6.0 \| 10446 \| 3.0093 \| \| 2.6817 \| 7.0 \| 12187 \| 3.0105 \| \| 2.6458 \| 8.0 \| 13928 \| 3.0156 \| \| 2.6195 \| 9.0 \| 15669 \| 3.0205 \| \| 2.5997 \| 10.0 \| 17410 \| 3.0246 \| ### Framework versions - Transformers 4.32.1 - Pytorch 2.0.1+cu118 - Datasets 2.14.4 - Tokenizers 0.13.3
cvapict/yhi-message-type-paraphrase-mpnet-base-v2	cvapict	2023-08-30T12:11:46Z	4	0	sentence-transformers	[ "sentence-transformers", "pytorch", "mpnet", "setfit", "text-classification", "arxiv:2209.11055", "license:apache-2.0", "region:us" ]	text-classification	2023-08-30T12:11:04Z	--- license: apache-2.0 tags: - setfit - sentence-transformers - text-classification pipeline_tag: text-classification --- # cvapict/yhi-message-type-paraphrase-mpnet-base-v2 {'accuracy': 0.8536585365853658} This is a [SetFit model](https://github.com/huggingface/setfit) that can be used for text classification. The model has been trained using an efficient few-shot learning technique that involves: 1. Fine-tuning a [Sentence Transformer](https://www.sbert.net) with contrastive learning. 2. Training a classification head with features from the fine-tuned Sentence Transformer. ## Usage To use this model for inference, first install the SetFit library: ```bash python -m pip install setfit ``` You can then run inference as follows: ```python from setfit import SetFitModel # Download from Hub and run inference model = SetFitModel.from_pretrained("cvapict/yhi-message-type-paraphrase-mpnet-base-v2") # Run inference preds = model(["i loved the spiderman movie!", "pineapple on pizza is the worst 🤮"]) ``` ## BibTeX entry and citation info ```bibtex @article{https://doi.org/10.48550/arxiv.2209.11055, doi = {10.48550/ARXIV.2209.11055}, url = {https://arxiv.org/abs/2209.11055}, author = {Tunstall, Lewis and Reimers, Nils and Jo, Unso Eun Seo and Bates, Luke and Korat, Daniel and Wasserblat, Moshe and Pereg, Oren}, keywords = {Computation and Language (cs.CL), FOS: Computer and information sciences, FOS: Computer and information sciences}, title = {Efficient Few-Shot Learning Without Prompts}, publisher = {arXiv}, year = {2022}, copyright = {Creative Commons Attribution 4.0 International} } ```
vinayaksodar/ppo-LunarLander-v2	vinayaksodar	2023-08-30T12:10:20Z	0	0	stable-baselines3	[ "stable-baselines3", "LunarLander-v2", "deep-reinforcement-learning", "reinforcement-learning", "model-index", "region:us" ]	reinforcement-learning	2023-08-30T12:10:00Z	--- library_name: stable-baselines3 tags: - LunarLander-v2 - deep-reinforcement-learning - reinforcement-learning - stable-baselines3 model-index: - name: PPO results: - task: type: reinforcement-learning name: reinforcement-learning dataset: name: LunarLander-v2 type: LunarLander-v2 metrics: - type: mean_reward value: 267.79 +/- 9.72 name: mean_reward verified: false --- # PPO Agent playing LunarLander-v2 This is a trained model of a PPO agent playing LunarLander-v2 using the [stable-baselines3 library](https://github.com/DLR-RM/stable-baselines3). ## Usage (with Stable-baselines3) TODO: Add your code ```python from stable_baselines3 import ... from huggingface_sb3 import load_from_hub ... ```
bigscience/bloom-3b-intermediate	bigscience	2023-08-30T12:06:50Z	20	1	transformers	[ "transformers", "pytorch", "safetensors", "bloom", "text-generation", "ak", "ar", "as", "bm", "bn", "ca", "code", "en", "es", "eu", "fon", "fr", "gu", "hi", "id", "ig", "ki", "kn", "lg", "ln", "ml", "mr", "ne", "nso", "ny", "or", "pa", "pt", "rn", "rw", "sn", "st", "sw", "ta", "te", "tn", "ts", "tum", "tw", "ur", "vi", "wo", "xh", "yo", "zh", "zhs", "zht", "zu", "arxiv:1909.08053", "arxiv:2110.02861", "arxiv:2108.12409", "license:bigscience-bloom-rail-1.0", "autotrain_compatible", "text-generation-inference", "endpoints_compatible", "region:us" ]	text-generation	2022-08-07T17:47:50Z	--- license: bigscience-bloom-rail-1.0 language: - ak - ar - as - bm - bn - ca - code - en - es - eu - fon - fr - gu - hi - id - ig - ki - kn - lg - ln - ml - mr - ne - nso - ny - or - pa - pt - rn - rw - sn - st - sw - ta - te - tn - ts - tum - tw - ur - vi - wo - xh - yo - zh - zhs - zht - zu --- # <span style="color:red"><b>WARNING:</b> The checkpoints on this repo are not fully trained model. Evaluations of intermediary checkpoints and the final model will be added when conducted (see below).</span> # <p>BLOOM LM<br/> _BigScience Large Open-science Open-access Multilingual Language Model_ <br/>Model Card</p> <img src="https://assets.website-files.com/6139f3cdcbbff3a68486761d/613cd8997b270da063e230c5_Tekengebied%201-p-500.png" alt="BigScience Logo" width="200"/> Version 1.3 / 11.July.2022 - Available intermediary checkpoints - global steps: + `1000`, `10000`, `50000`, `100000`, `150000`, `200000`, `250000`, `300000` You can check the available checkpoints by clicking on the branches section of the repo # How to load a specific version We use `git tags` to load a model in a specific version (eg. `global_step1000`): ```python from transformers import AutoModelForCausalLM model = AutoModelForCausalLM.from_pretrained( "bigscience/bloom-2b5-intermediate", revision="global_step1000", torch_dtype="auto", ) ``` # Table of Contents 1. [Model Details](#model-details) 2. [Uses](#uses) 3. [Training Data](#training-data) 4. [Risks and Limitations](#risks-and-limitations) 5. [Evaluation](#evaluation) 6. [Recommendations](#recommendations) 7. [Glossary and Calculations](#glossary-and-calculations) 8. [More Information](#more-information) 9. [Model Card Authors](#model-card-authors) --- # Model Details BLOOM is a type of language model, which is a probability distribution over sequences of words. Specifically, BLOOM is a Large Language Model (LLM), meaning that it is trained on vast amounts of text data using industrial-scale computational resources. As such, the model is able to capture the statistical tendencies of words, phrases, sentences, and larger spans of text that it is exposed to in the training data. ## Basics This section provides information about the model type, version, license, funders, release date, developers, and contact information. It is useful for anyone who wants to reference the model. <details> <summary>Click to expand</summary> Developed by: BigScience ([website](https://bigscience.huggingface.co)) All collaborators are either volunteers or have an agreement with their employer. (Further breakdown of participants forthcoming.) Model Type: Transformer-based Language Model Version: 1.0.0 Languages: Multiple; see [training data](#training-data) License: RAIL License v1.0 ([link](https://huggingface.co/spaces/bigscience/license)) Release Date Estimate: Monday, 11.July.2022 Send Questions to: [email protected] Cite as: BigScience, _BigScience Language Open-science Open-access Multilingual (BLOOM) Language Model_. International, May 2021-May 2022 Funded by: * The French government. * Hugging Face ([website](https://huggingface.co)). * Organizations of contributors. (Further breakdown of organizations forthcoming.) </details> ## Technical Specifications This section includes details about the model objective and architecture, and the compute infrastructure. It is useful for people interested in model development. <details> <summary>Click to expand</summary> Please see [the BLOOM training README](https://github.com/bigscience-workshop/bigscience/tree/master/train/tr11-176B-ml#readme) for full details on replicating training. ### Model Architecture and Objective * Modified from Megatron-LM GPT2 (see [paper](https://arxiv.org/abs/1909.08053), [BLOOM Megatron code](https://github.com/bigscience-workshop/Megatron-DeepSpeed)): * Decoder-only architecture * Layer normalization applied to word embeddings layer (`StableEmbedding`; see [code](https://github.com/facebookresearch/bitsandbytes), [paper](https://arxiv.org/pdf/2110.02861.pdf)) * ALiBI positional encodings (see [paper](https://arxiv.org/pdf/2108.12409.pdf)), with GeLU activation functions * 176 billion parameters: * 70 layers, 112 attention heads * Hidden layers are 14336-dimensional * Sequence length of 2048 tokens used (see [BLOOM tokenizer](https://huggingface.co/bigscience/tokenizer), [tokenizer description](#tokenization)) Objective Function: Cross Entropy with mean reduction (see [API documentation](https://pytorch.org/docs/stable/generated/torch.nn.CrossEntropyLoss.html#torch.nn.CrossEntropyLoss)). ### Compute infrastructure Jean Zay Public Supercomputer, provided by the French government (see [announcement](https://www.enseignementsup-recherche.gouv.fr/fr/signature-du-marche-d-acquisition-de-l-un-des-supercalculateurs-les-plus-puissants-d-europe-46733)). #### Hardware * 384 A100 80GB GPUs (48 nodes) * Additional 32 A100 80GB GPUs (4 nodes) in reserve * 8 GPUs per node Using NVLink 4 inter-gpu connects, 4 OmniPath links * CPU: AMD * CPU memory: 512GB per node * GPU memory: 640GB per node * Inter-node connect: Omni-Path Architecture (OPA) * NCCL-communications network: a fully dedicated subnet * Disc IO network: shared network with other types of nodes #### Software * Megatron-DeepSpeed ([Github link](https://github.com/bigscience-workshop/Megatron-DeepSpeed)) * DeepSpeed ([Github link](https://github.com/microsoft/DeepSpeed)) * PyTorch (pytorch-1.11 w/ CUDA-11.5; see [Github link](https://github.com/pytorch/pytorch)) * apex ([Github link](https://github.com/NVIDIA/apex)) </details> --- # Training This section provides information about the training data, the speed and size of training elements, and the environmental impact of training. It is useful for people who want to learn more about the model inputs and training footprint. <details> <summary>Click to expand</summary> ## Training Data This section provides a high-level overview of the training data. It is relevant for anyone who wants to know the basics of what the model is learning. Details for each dataset are provided in individual [Data Cards](https://huggingface.co/spaces/bigscience/BigScienceCorpus). Training data includes: - 45 natural languages - 12 programming languages - In 1.5TB of pre-processed text, converted into 350B unique tokens (see [the tokenizer section](#tokenization) for more.) ### Languages The pie chart shows the distribution of languages in training data. ![pie chart showing the distribution of languages in training data](https://github.com/bigscience-workshop/model_card/blob/main/assets/data/pie_chart.svg?raw=true) The following tables shows the further distribution of Niger-Congo & Indic languages and programming languages in the training data. Distribution of Niger Congo and Indic languages. \| Niger Congo \| Percentage \| \| Indic \| Percentage \| \|----------------\|------------ \|------ \|-----------\|------------\| \| Chi Tumbuka \| 0.00002 \| \| Assamese \| 0.01 \| \| Kikuyu \| 0.00004 \| \| Odia \| 0.04 \| \| Bambara \| 0.00004 \| \| Gujarati \| 0.04 \| \| Akan \| 0.00007 \| \| Marathi \| 0.05 \| \| Xitsonga \| 0.00007 \| \| Punjabi \| 0.05 \| \| Sesotho \| 0.00007 \| \| Kannada \| 0.06 \| \| Chi Chewa \| 0.0001 \| \| Nepali \| 0.07 \| \| Setswana \| 0.0002 \| \| Telugu \| 0.09 \| \| Northern Sotho \| 0.0002 \| \| Malayalam \| 0.10 \| \| Fon \| 0.0002 \| \| Urdu \| 0.10 \| \| Kirundi \| 0.0003 \| \| Tamil \| 0.20 \| \| Wolof \| 0.0004 \| \| Bengali \| 0.50 \| \| Kuganda \| 0.0004 \| \| Hindi \| 0.70 \| \| Chi Shona \| 0.001 \| \| Isi Zulu \| 0.001 \| \| Igbo \| 0.001 \| \| Xhosa \| 0.001 \| \| Kinyarwanda \| 0.003 \| \| Yoruba \| 0.006 \| \| Swahili \| 0.02 \| Distribution of programming languages. \| Extension \| Language \| Number of files \| \|----------------\|------------\|-----------------\| \| java \| Java \| 5,407,724 \| \| php \| PHP \| 4,942,186 \| \| cpp \| C++ \| 2,503,930 \| \| py \| Python \| 2,435,072 \| \| js \| JavaScript \| 1,905,518 \| \| cs \| C# \| 1,577,347 \| \| rb \| Ruby \| 6,78,413 \| \| cc \| C++ \| 443,054 \| \| hpp \| C++ \| 391,048 \| \| lua \| Lua \| 352,317 \| \| go \| GO \| 227,763 \| \| ts \| TypeScript \| 195,254 \| \| C \| C \| 134,537 \| \| scala \| Scala \| 92,052 \| \| hh \| C++ \| 67,161 \| \| H \| C++ \| 55,899 \| \| tsx \| TypeScript \| 33,107 \| \| rs \| Rust \| 29,693 \| \| phpt \| PHP \| 9,702 \| \| c++ \| C++ \| 1,342 \| \| h++ \| C++ \| 791 \| \| php3 \| PHP \| 540 \| \| phps \| PHP \| 270 \| \| php5 \| PHP \| 166 \| \| php4 \| PHP \| 29 \| ### Preprocessing Tokenization: The BLOOM tokenizer ([link](https://huggingface.co/bigscience/tokenizer)), a learned subword tokenizer trained using: - A byte-level Byte Pair Encoding (BPE) algorithm - A simple pre-tokenization rule, no normalization - A vocabulary size of 250,680 It was trained on a subset of a preliminary version of the corpus using alpha-weighting per language. ## Speeds, Sizes, Times Training logs: [Tensorboard link](https://huggingface.co/tensorboard/bigscience/tr11-176B-ml-logs/) - Dates: - Started 11th March, 2022 11:42am PST - Estimated end: 5th July, 2022 - Checkpoint size: - Bf16 weights: 329GB - Full checkpoint with optimizer states: 2.3TB - Training throughput: About 150 TFLOP per GPU per second - Number of epochs: 1 - Estimated cost of training: Equivalent of $2-5M in cloud computing (including preliminary experiments) - Server training location: Île-de-France, France ## Environmental Impact The training supercomputer, Jean Zay ([website](http://www.idris.fr/eng/jean-zay/jean-zay-presentation-eng.html)), uses mostly nuclear energy. The heat generated by it is reused for heating campus housing. Estimated carbon emissions: (Forthcoming.) Estimated electricity usage: (Forthcoming.) </details> --- # Uses This section addresses questions around how the model is intended to be used, discusses the foreseeable users of the model (including those affected by the model), and describes uses that are considered out of scope or misuse of the model. It is useful for anyone considering using the model or who is affected by the model. <details> <summary>Click to expand</summary> ## Intended Use This model is being created in order to enable public research on large language models (LLMs). LLMs are intended to be used for language generation or as a pretrained base model that can be further fine-tuned for specific tasks. Use cases below are not exhaustive. ### Direct Use - Text generation - Exploring characteristics of language generated by a language model - Examples: Cloze tests, counterfactuals, generations with reframings ### Downstream Use - Tasks that leverage language models include: Information Extraction, Question Answering, Summarization ### Misuse and Out-of-scope Use This section addresses what users ought not do with the model. See the [BLOOM License](https://huggingface.co/spaces/bigscience/license), Attachment A, for detailed usage restrictions. The below list is non-exhaustive, but lists some easily foreseeable problematic use cases. #### Out-of-scope Uses Using the model in [high-stakes](#high-stakes) settings is out of scope for this model. The model is not designed for [critical decisions](#critical-decisions) nor uses with any material consequences on an individual's livelihood or wellbeing. The model outputs content that appears factual but is not correct. Out-of-scope Uses Include: - Usage in biomedical domains, political and legal domains, or finance domains - Usage for evaluating or scoring individuals, such as for employment, education, or credit - Applying the model for critical automatic decisions, generating factual content, creating reliable summaries, or generating predictions that must be correct #### Misuse Intentionally using the model for harm, violating [human rights](#human-rights), or other kinds of malicious activities, is a misuse of this model. This includes: - Spam generation - Disinformation and influence operations - Disparagement and defamation - Harassment and abuse - [Deception](#deception) - Unconsented impersonation and imitation - Unconsented surveillance - Generating content without attribution to the model, as specified in the [RAIL License, Use Restrictions](https://huggingface.co/spaces/bigscience/license) ## Intended Users ### Direct Users - General Public - Researchers - Students - Educators - Engineers/developers - Non-commercial entities - Community advocates, including human and civil rights groups ### Indirect Users - Users of derivatives created by Direct Users, such as those using software with an [intended use](#intended-use) - Users of [Derivatives of the Model, as described in the License](https://huggingface.co/spaces/bigscience/license) ### Others Affected (Parties Prenantes) - People and groups referred to by the LLM - People and groups exposed to outputs of, or decisions based on, the LLM - People and groups whose original work is included in the LLM </details> --- # Risks and Limitations This section identifies foreseeable harms and misunderstandings. <details> <summary>Click to expand</summary> Model may: - Overrepresent some viewpoints and underrepresent others - Contain stereotypes - Contain [personal information](#personal-data-and-information) - Generate: - Hateful, abusive, or violent language - Discriminatory or prejudicial language - Content that may not be appropriate for all settings, including sexual content - Make errors, including producing incorrect information as if it were factual - Generate irrelevant or repetitive outputs </details> --- # Evaluation This section describes the evaluation protocols and provides the results. <details> <summary>Click to expand</summary> ## Metrics This section describes the different ways performance is calculated and why. Includes: \| Metric \| Why chosen \| \|--------------------\|--------------------------------------------------------------------\| \| [Perplexity](#perplexity) \| Standard metric for quantifying model improvements during training \| \| Cross Entropy [Loss](#loss) \| Standard objective for language models. \| And multiple different metrics for specific tasks. _(More evaluation metrics forthcoming upon completion of evaluation protocol.)_ ## Factors This section lists some different aspects of what BLOOM models. Its focus is on those aspects that are likely to give rise to high variance in model behavior. - Language, such as English or Yoruba - Domain, such as newswire or stories - Demographic characteristics, such as gender or nationality ## Results Results are based on the [Factors](#factors) and [Metrics](#metrics). Train-time Evaluation: As of 25.May.2022, 15:00 PST: - Training Loss: 2.0 - Validation Loss: 2.2 - Perplexity: 8.9 (More evaluation scores forthcoming.) </details> --- # Recommendations This section provides information on warnings and potential mitigations. <details> <summary>Click to expand</summary> - Indirect users should be made aware when the content they're working with is created by the LLM. - Users should be aware of [Risks and Limitations](#risks-and-limitations), and include an appropriate age disclaimer or blocking interface as necessary. - Models trained or finetuned downstream of BLOOM LM should include an updated Model Card. - Users of the model should provide mechanisms for those affected to provide feedback, such as an email address for comments. </details> --- # Glossary and Calculations This section defines common terms and how metrics are calculated. <details> <summary>Click to expand</summary> - <a name="loss">Loss:</a> A calculation of the difference between what the model has learned and what the data shows ("groundtruth"). The lower the loss, the better. The training process aims to minimize the loss. - <a name="perplexity">Perplexity:</a> This is based on what the model estimates the probability of new data is. The lower the perplexity, the better. If the model is 100% correct at predicting the next token it will see, then the perplexity is 1. Mathematically this is calculated using entropy. - <a name="high-stakes">High-stakes settings:</a> Such as those identified as "high-risk AI systems" and "unacceptable risk AI systems" in the European Union's proposed [Artificial Intelligence (AI) Act](https://artificialintelligenceact.eu/annexes/). - <a name="critical-decisions">Critical decisions:</a> Such as those defined in [the United States' proposed Algorithmic Accountability Act](https://www.congress.gov/117/bills/s3572/BILLS-117s3572is.pdf). - <a name="human-rights">Human rights:</a> Includes those rights defined in the [Universal Declaration of Human Rights](https://www.un.org/sites/un2.un.org/files/2021/03/udhr.pdf). - <a name="personal-data-and-information">Personal Data and Personal Information:</a> Personal data and information is defined in multiple data protection regulations, such as "[personal data](https://gdpr-info.eu/issues/personal-data/)" in the [European Union's General Data Protection Regulation](https://gdpr-info.eu); and "personal information" in the Republic of South Africa's [Protection of Personal Information Act](https://www.gov.za/sites/default/files/gcis_document/201409/3706726-11act4of2013popi.pdf), The People's Republic of China's [Personal information protection law](http://en.npc.gov.cn.cdurl.cn/2021-12/29/c_694559.htm). - <a name="sensitive-characteristics">Sensitive characteristics:</a> This includes specifically protected categories in human rights (see [UHDR, Article 2](https://www.un.org/sites/un2.un.org/files/2021/03/udhr.pdf)) and personal information regulation (see GDPR, [Article 9; Protection of Personal Information Act, Chapter 1](https://www.gov.za/sites/default/files/gcis_document/201409/3706726-11act4of2013popi.pdf)) - <a name="deception">Deception:</a> Doing something to intentionally mislead individuals to believe something that is false, such as by creating deadbots or chatbots on social media posing as real people, or generating text documents without making consumers aware that the text is machine generated. </details> --- # More Information This section provides links to writing on dataset creation, technical specifications, lessons learned, and initial results. <details> <summary>Click to expand</summary> ## Dataset Creation Blog post detailing the design choices during the dataset creation: https://bigscience.huggingface.co/blog/building-a-tb-scale-multilingual-dataset-for-language-modeling ## Technical Specifications Blog post summarizing how the architecture, size, shape, and pre-training duration where selected: https://bigscience.huggingface.co/blog/what-language-model-to-train-if-you-have-two-million-gpu-hours More details on the architecture/optimizer: https://github.com/bigscience-workshop/bigscience/tree/master/train/tr11-176B-ml Blog post on the hardware/engineering side: https://bigscience.huggingface.co/blog/which-hardware-to-train-a-176b-parameters-model Details on the distributed setup used for the training: https://github.com/bigscience-workshop/bigscience/tree/master/train/tr11-176B-ml Tensorboard updated during the training: https://huggingface.co/bigscience/tr11-176B-ml-logs/tensorboard#scalars&tagFilter=loss ## Lessons Insights on how to approach training, negative results: https://github.com/bigscience-workshop/bigscience/blob/master/train/lessons-learned.md Details on the obstacles overcome during the preparation on the engineering side (instabilities, optimization of training throughput, so many technical tricks and questions): https://github.com/bigscience-workshop/bigscience/blob/master/train/tr11-176B-ml/chronicles.md ## Initial Results Initial prompting experiments using interim checkpoints: https://huggingface.co/spaces/bigscience/bloom-book </details> --- # Model Card Authors Ordered roughly chronologically and by amount of time spent. Margaret Mitchell, Giada Pistilli, Yacine Jernite, Ezinwanne Ozoani, Marissa Gerchick, Nazneen Rajani, Sasha Luccioni, Irene Solaiman, Maraim Masoud, Somaieh Nikpoor, Carlos Muñoz Ferrandis, Stas Bekman, Christopher Akiki, Danish Contractor, David Lansky, Angelina McMillan-Major, Tristan Thrush, Suzana Ilić, Gérard Dupont, Shayne Longpre, Manan Dey, Stella Biderman, Douwe Kiela, Emi Baylor, Teven Le Scao, Aaron Gokaslan, Julien Launay, Niklas Muennighoff
SHENMU007/neunit_BASE_V13.5.6	SHENMU007	2023-08-30T11:59:02Z	75	0	transformers	[ "transformers", "pytorch", "tensorboard", "speecht5", "text-to-audio", "1.1.0", "generated_from_trainer", "zh", "dataset:facebook/voxpopuli", "base_model:microsoft/speecht5_tts", "base_model:finetune:microsoft/speecht5_tts", "license:mit", "endpoints_compatible", "region:us" ]	text-to-audio	2023-08-30T07:34:28Z	--- language: - zh license: mit base_model: microsoft/speecht5_tts tags: - 1.1.0 - generated_from_trainer datasets: - facebook/voxpopuli model-index: - name: SpeechT5 TTS Dutch neunit results: [] --- <!-- This model card has been generated automatically according to the information the Trainer had access to. You should probably proofread and complete it, then remove this comment. --> # SpeechT5 TTS Dutch neunit This model is a fine-tuned version of [microsoft/speecht5_tts](https://huggingface.co/microsoft/speecht5_tts) on the VoxPopuli dataset. ## Model description More information needed ## Intended uses & limitations More information needed ## Training and evaluation data More information needed ## Training procedure ### Training hyperparameters The following hyperparameters were used during training: - learning_rate: 1e-05 - train_batch_size: 8 - eval_batch_size: 8 - seed: 42 - gradient_accumulation_steps: 4 - total_train_batch_size: 32 - optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08 - lr_scheduler_type: linear - lr_scheduler_warmup_steps: 500 - training_steps: 4000 ### Training results ### Framework versions - Transformers 4.31.0.dev0 - Pytorch 2.0.1+cu117 - Datasets 2.12.0 - Tokenizers 0.13.3
cvapict/yhi-message-type-all-distilroberta-v1	cvapict	2023-08-30T11:45:48Z	5	0	sentence-transformers	[ "sentence-transformers", "pytorch", "roberta", "setfit", "text-classification", "arxiv:2209.11055", "license:apache-2.0", "region:us" ]	text-classification	2023-08-30T11:44:20Z	--- license: apache-2.0 tags: - setfit - sentence-transformers - text-classification pipeline_tag: text-classification --- # cvapict/yhi-message-type-all-distilroberta-v1 {'accuracy': 0.8373983739837398} This is a [SetFit model](https://github.com/huggingface/setfit) that can be used for text classification. The model has been trained using an efficient few-shot learning technique that involves: 1. Fine-tuning a [Sentence Transformer](https://www.sbert.net) with contrastive learning. 2. Training a classification head with features from the fine-tuned Sentence Transformer. ## Usage To use this model for inference, first install the SetFit library: ```bash python -m pip install setfit ``` You can then run inference as follows: ```python from setfit import SetFitModel # Download from Hub and run inference model = SetFitModel.from_pretrained("cvapict/yhi-message-type-all-distilroberta-v1") # Run inference preds = model(["i loved the spiderman movie!", "pineapple on pizza is the worst 🤮"]) ``` ## BibTeX entry and citation info ```bibtex @article{https://doi.org/10.48550/arxiv.2209.11055, doi = {10.48550/ARXIV.2209.11055}, url = {https://arxiv.org/abs/2209.11055}, author = {Tunstall, Lewis and Reimers, Nils and Jo, Unso Eun Seo and Bates, Luke and Korat, Daniel and Wasserblat, Moshe and Pereg, Oren}, keywords = {Computation and Language (cs.CL), FOS: Computer and information sciences, FOS: Computer and information sciences}, title = {Efficient Few-Shot Learning Without Prompts}, publisher = {arXiv}, year = {2022}, copyright = {Creative Commons Attribution 4.0 International} } ```
Shivam098/my_awesome_opus_books_model	Shivam098	2023-08-30T11:40:07Z	12	0	transformers	[ "transformers", "pytorch", "tensorboard", "safetensors", "t5", "text2text-generation", "generated_from_trainer", "dataset:opus100", "base_model:google-t5/t5-small", "base_model:finetune:google-t5/t5-small", "license:apache-2.0", "model-index", "autotrain_compatible", "text-generation-inference", "endpoints_compatible", "region:us" ]	text2text-generation	2023-08-24T11:36:29Z	--- license: apache-2.0 base_model: t5-small tags: - generated_from_trainer datasets: - opus100 metrics: - bleu model-index: - name: my_awesome_opus_books_model results: - task: name: Sequence-to-sequence Language Modeling type: text2text-generation dataset: name: opus100 type: opus100 config: en-ps split: train args: en-ps metrics: - name: Bleu type: bleu value: 8.2239 --- <!-- This model card has been generated automatically according to the information the Trainer had access to. You should probably proofread and complete it, then remove this comment. --> # my_awesome_opus_books_model This model is a fine-tuned version of [t5-small](https://huggingface.co/t5-small) on the opus100 dataset. It achieves the following results on the evaluation set: - Loss: 3.1766 - Bleu: 8.2239 - Gen Len: 7.6785 ## Model description More information needed ## Intended uses & limitations More information needed ## Training and evaluation data More information needed ## Training procedure ### Training hyperparameters The following hyperparameters were used during training: - learning_rate: 2e-05 - train_batch_size: 16 - eval_batch_size: 16 - seed: 42 - optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08 - lr_scheduler_type: linear - num_epochs: 2 ### Training results \| Training Loss \| Epoch \| Step \| Validation Loss \| Bleu \| Gen Len \| \|:-------------:\|:-----:\|:----:\|:---------------:\|:------:\|:-------:\| \| 3.5442 \| 1.0 \| 3957 \| 3.2616 \| 9.5837 \| 6.9205 \| \| 3.3951 \| 2.0 \| 7914 \| 3.1766 \| 8.2239 \| 7.6785 \| ### Framework versions - Transformers 4.32.0 - Pytorch 2.0.1+cu118 - Datasets 2.14.4 - Tokenizers 0.13.3
cvapict/yhi-message-type-all-mpnet-base-v2	cvapict	2023-08-30T11:39:14Z	4	0	sentence-transformers	[ "sentence-transformers", "pytorch", "mpnet", "setfit", "text-classification", "arxiv:2209.11055", "license:apache-2.0", "region:us" ]	text-classification	2023-08-30T11:37:48Z	--- license: apache-2.0 tags: - setfit - sentence-transformers - text-classification pipeline_tag: text-classification --- # cvapict/yhi-message-type-all-mpnet-base-v2 {'accuracy': 0.8292682926829268} This is a [SetFit model](https://github.com/huggingface/setfit) that can be used for text classification. The model has been trained using an efficient few-shot learning technique that involves: 1. Fine-tuning a [Sentence Transformer](https://www.sbert.net) with contrastive learning. 2. Training a classification head with features from the fine-tuned Sentence Transformer. ## Usage To use this model for inference, first install the SetFit library: ```bash python -m pip install setfit ``` You can then run inference as follows: ```python from setfit import SetFitModel # Download from Hub and run inference model = SetFitModel.from_pretrained("cvapict/yhi-message-type-all-mpnet-base-v2") # Run inference preds = model(["i loved the spiderman movie!", "pineapple on pizza is the worst 🤮"]) ``` ## BibTeX entry and citation info ```bibtex @article{https://doi.org/10.48550/arxiv.2209.11055, doi = {10.48550/ARXIV.2209.11055}, url = {https://arxiv.org/abs/2209.11055}, author = {Tunstall, Lewis and Reimers, Nils and Jo, Unso Eun Seo and Bates, Luke and Korat, Daniel and Wasserblat, Moshe and Pereg, Oren}, keywords = {Computation and Language (cs.CL), FOS: Computer and information sciences, FOS: Computer and information sciences}, title = {Efficient Few-Shot Learning Without Prompts}, publisher = {arXiv}, year = {2022}, copyright = {Creative Commons Attribution 4.0 International} } ```
dkqjrm/20230830152959	dkqjrm	2023-08-30T11:34:30Z	105	0	transformers	[ "transformers", "pytorch", "tensorboard", "bert", "text-classification", "generated_from_trainer", "dataset:super_glue", "license:apache-2.0", "autotrain_compatible", "endpoints_compatible", "region:us" ]	text-classification	2023-08-30T06:30:26Z	--- license: apache-2.0 tags: - generated_from_trainer datasets: - super_glue metrics: - accuracy model-index: - name: '20230830152959' results: [] --- <!-- This model card has been generated automatically according to the information the Trainer had access to. You should probably proofread and complete it, then remove this comment. --> # 20230830152959 This model is a fine-tuned version of [bert-large-cased](https://huggingface.co/bert-large-cased) on the super_glue dataset. It achieves the following results on the evaluation set: - Loss: 0.7271 - Accuracy: 0.5 ## Model description More information needed ## Intended uses & limitations More information needed ## Training and evaluation data More information needed ## Training procedure ### Training hyperparameters The following hyperparameters were used during training: - learning_rate: 0.0007 - train_batch_size: 16 - eval_batch_size: 8 - seed: 11 - optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08 - lr_scheduler_type: linear - num_epochs: 80.0 ### Training results \| Training Loss \| Epoch \| Step \| Validation Loss \| Accuracy \| \|:-------------:\|:-----:\|:-----:\|:---------------:\|:--------:\| \| No log \| 1.0 \| 340 \| 0.7372 \| 0.5 \| \| 0.7583 \| 2.0 \| 680 \| 0.7264 \| 0.5 \| \| 0.7541 \| 3.0 \| 1020 \| 0.7485 \| 0.5 \| \| 0.7541 \| 4.0 \| 1360 \| 0.7304 \| 0.5 \| \| 0.7509 \| 5.0 \| 1700 \| 0.7363 \| 0.5 \| \| 0.7455 \| 6.0 \| 2040 \| 0.7360 \| 0.5 \| \| 0.7455 \| 7.0 \| 2380 \| 0.7240 \| 0.5 \| \| 0.7493 \| 8.0 \| 2720 \| 0.7806 \| 0.5 \| \| 0.7515 \| 9.0 \| 3060 \| 0.7284 \| 0.5 \| \| 0.7515 \| 10.0 \| 3400 \| 0.7375 \| 0.5 \| \| 0.7489 \| 11.0 \| 3740 \| 0.7340 \| 0.5 \| \| 0.7452 \| 12.0 \| 4080 \| 0.7255 \| 0.5 \| \| 0.7452 \| 13.0 \| 4420 \| 0.8158 \| 0.5 \| \| 0.7463 \| 14.0 \| 4760 \| 0.7256 \| 0.5 \| \| 0.75 \| 15.0 \| 5100 \| 0.7364 \| 0.5 \| \| 0.75 \| 16.0 \| 5440 \| 0.7260 \| 0.5 \| \| 0.7411 \| 17.0 \| 5780 \| 0.7320 \| 0.5 \| \| 0.7482 \| 18.0 \| 6120 \| 0.7296 \| 0.5 \| \| 0.7482 \| 19.0 \| 6460 \| 0.7312 \| 0.5 \| \| 0.7424 \| 20.0 \| 6800 \| 0.7405 \| 0.5 \| \| 0.7374 \| 21.0 \| 7140 \| 0.7258 \| 0.5 \| \| 0.7374 \| 22.0 \| 7480 \| 0.7609 \| 0.5 \| \| 0.7388 \| 23.0 \| 7820 \| 0.7608 \| 0.5 \| \| 0.7382 \| 24.0 \| 8160 \| 0.7239 \| 0.5 \| \| 0.7385 \| 25.0 \| 8500 \| 0.7315 \| 0.5 \| \| 0.7385 \| 26.0 \| 8840 \| 0.7472 \| 0.5 \| \| 0.7392 \| 27.0 \| 9180 \| 0.7863 \| 0.5 \| \| 0.737 \| 28.0 \| 9520 \| 0.7261 \| 0.5 \| \| 0.737 \| 29.0 \| 9860 \| 0.7403 \| 0.5 \| \| 0.7322 \| 30.0 \| 10200 \| 0.7245 \| 0.5 \| \| 0.7359 \| 31.0 \| 10540 \| 0.7239 \| 0.5 \| \| 0.7359 \| 32.0 \| 10880 \| 0.7555 \| 0.5 \| \| 0.7368 \| 33.0 \| 11220 \| 0.7239 \| 0.5 \| \| 0.7349 \| 34.0 \| 11560 \| 0.7380 \| 0.5 \| \| 0.7349 \| 35.0 \| 11900 \| 0.7279 \| 0.5 \| \| 0.7367 \| 36.0 \| 12240 \| 0.7263 \| 0.5 \| \| 0.7343 \| 37.0 \| 12580 \| 0.7252 \| 0.5 \| \| 0.7343 \| 38.0 \| 12920 \| 0.7299 \| 0.5 \| \| 0.7347 \| 39.0 \| 13260 \| 0.7344 \| 0.5 \| \| 0.7315 \| 40.0 \| 13600 \| 0.7247 \| 0.5 \| \| 0.7315 \| 41.0 \| 13940 \| 0.7277 \| 0.5 \| \| 0.7324 \| 42.0 \| 14280 \| 0.7246 \| 0.5 \| \| 0.7319 \| 43.0 \| 14620 \| 0.7289 \| 0.5 \| \| 0.7319 \| 44.0 \| 14960 \| 0.7297 \| 0.5 \| \| 0.7321 \| 45.0 \| 15300 \| 0.7389 \| 0.5 \| \| 0.7304 \| 46.0 \| 15640 \| 0.7245 \| 0.5 \| \| 0.7304 \| 47.0 \| 15980 \| 0.7306 \| 0.5 \| \| 0.7316 \| 48.0 \| 16320 \| 0.7239 \| 0.5 \| \| 0.7325 \| 49.0 \| 16660 \| 0.7312 \| 0.5 \| \| 0.7311 \| 50.0 \| 17000 \| 0.7241 \| 0.5 \| \| 0.7311 \| 51.0 \| 17340 \| 0.7239 \| 0.5 \| \| 0.7292 \| 52.0 \| 17680 \| 0.7244 \| 0.5 \| \| 0.732 \| 53.0 \| 18020 \| 0.7250 \| 0.5 \| \| 0.732 \| 54.0 \| 18360 \| 0.7239 \| 0.5 \| \| 0.727 \| 55.0 \| 18700 \| 0.7257 \| 0.5 \| \| 0.7277 \| 56.0 \| 19040 \| 0.7247 \| 0.5 \| \| 0.7277 \| 57.0 \| 19380 \| 0.7256 \| 0.5 \| \| 0.7296 \| 58.0 \| 19720 \| 0.7269 \| 0.5 \| \| 0.7284 \| 59.0 \| 20060 \| 0.7340 \| 0.5 \| \| 0.7284 \| 60.0 \| 20400 \| 0.7257 \| 0.5 \| \| 0.7269 \| 61.0 \| 20740 \| 0.7254 \| 0.5 \| \| 0.7266 \| 62.0 \| 21080 \| 0.7240 \| 0.5 \| \| 0.7266 \| 63.0 \| 21420 \| 0.7245 \| 0.5 \| \| 0.7274 \| 64.0 \| 21760 \| 0.7261 \| 0.5 \| \| 0.7278 \| 65.0 \| 22100 \| 0.7240 \| 0.5 \| \| 0.7278 \| 66.0 \| 22440 \| 0.7492 \| 0.5 \| \| 0.7252 \| 67.0 \| 22780 \| 0.7254 \| 0.5 \| \| 0.727 \| 68.0 \| 23120 \| 0.7241 \| 0.5 \| \| 0.727 \| 69.0 \| 23460 \| 0.7244 \| 0.5 \| \| 0.7253 \| 70.0 \| 23800 \| 0.7242 \| 0.5 \| \| 0.7249 \| 71.0 \| 24140 \| 0.7247 \| 0.5 \| \| 0.7249 \| 72.0 \| 24480 \| 0.7280 \| 0.5 \| \| 0.7244 \| 73.0 \| 24820 \| 0.7267 \| 0.5 \| \| 0.7232 \| 74.0 \| 25160 \| 0.7284 \| 0.5 \| \| 0.7234 \| 75.0 \| 25500 \| 0.7272 \| 0.5 \| \| 0.7234 \| 76.0 \| 25840 \| 0.7294 \| 0.5 \| \| 0.7233 \| 77.0 \| 26180 \| 0.7257 \| 0.5 \| \| 0.7222 \| 78.0 \| 26520 \| 0.7284 \| 0.5 \| \| 0.7222 \| 79.0 \| 26860 \| 0.7276 \| 0.5 \| \| 0.7225 \| 80.0 \| 27200 \| 0.7271 \| 0.5 \| ### Framework versions - Transformers 4.26.1 - Pytorch 2.0.1+cu118 - Datasets 2.12.0 - Tokenizers 0.13.3
eliept1/ppo-Pyramids	eliept1	2023-08-30T11:24:47Z	0	0	ml-agents	[ "ml-agents", "tensorboard", "onnx", "Pyramids", "deep-reinforcement-learning", "reinforcement-learning", "ML-Agents-Pyramids", "region:us" ]	reinforcement-learning	2023-08-30T11:24:21Z	--- library_name: ml-agents tags: - Pyramids - deep-reinforcement-learning - reinforcement-learning - ML-Agents-Pyramids --- # ppo Agent playing Pyramids This is a trained model of a ppo agent playing Pyramids using the [Unity ML-Agents Library](https://github.com/Unity-Technologies/ml-agents). ## Usage (with ML-Agents) The Documentation: https://unity-technologies.github.io/ml-agents/ML-Agents-Toolkit-Documentation/ We wrote a complete tutorial to learn to train your first agent using ML-Agents and publish it to the Hub: - A short tutorial where you teach Huggy the Dog 🐶 to fetch the stick and then play with him directly in your browser: https://huggingface.co/learn/deep-rl-course/unitbonus1/introduction - A longer tutorial to understand how works ML-Agents: https://huggingface.co/learn/deep-rl-course/unit5/introduction ### Resume the training ```bash mlagents-learn <your_configuration_file_path.yaml> --run-id=<run_id> --resume ``` ### Watch your Agent play You can watch your agent playing directly in your browser 1. If the environment is part of ML-Agents official environments, go to https://huggingface.co/unity 2. Step 1: Find your model_id: eliept1/ppo-Pyramids 3. Step 2: Select your .nn /.onnx file 4. Click on Watch the agent play 👀
pumaML/turkishReviews-ds-mini	pumaML	2023-08-30T11:21:06Z	61	0	transformers	[ "transformers", "tf", "gpt2", "text-generation", "generated_from_keras_callback", "base_model:openai-community/gpt2", "base_model:finetune:openai-community/gpt2", "license:mit", "autotrain_compatible", "endpoints_compatible", "region:us" ]	text-generation	2023-08-30T11:15:23Z	--- license: mit base_model: gpt2 tags: - generated_from_keras_callback model-index: - name: pumaML/turkishReviews-ds-mini results: [] --- <!-- This model card has been generated automatically according to the information Keras had access to. You should probably proofread and complete it, then remove this comment. --> # pumaML/turkishReviews-ds-mini This model is a fine-tuned version of [gpt2](https://huggingface.co/gpt2) on an unknown dataset. It achieves the following results on the evaluation set: - Train Loss: 9.1942 - Validation Loss: 9.2669 - Epoch: 2 ## Model description More information needed ## Intended uses & limitations More information needed ## Training and evaluation data More information needed ## Training procedure ### Training hyperparameters The following hyperparameters were used during training: - optimizer: {'name': 'AdamWeightDecay', 'learning_rate': {'class_name': 'WarmUp', 'config': {'initial_learning_rate': 5e-05, 'decay_schedule_fn': {'class_name': 'PolynomialDecay', 'config': {'initial_learning_rate': 5e-05, 'decay_steps': -896, 'end_learning_rate': 0.0, 'power': 1.0, 'cycle': False, 'name': None}, '__passive_serialization__': True}, 'warmup_steps': 1000, 'power': 1.0, 'name': None}}, 'decay': 0.0, 'beta_1': 0.9, 'beta_2': 0.999, 'epsilon': 1e-08, 'amsgrad': False, 'weight_decay_rate': 0.01} - training_precision: mixed_float16 ### Training results \| Train Loss \| Validation Loss \| Epoch \| \|:----------:\|:---------------:\|:-----:\| \| 10.3292 \| 9.9955 \| 0 \| \| 9.6809 \| 9.6477 \| 1 \| \| 9.1942 \| 9.2669 \| 2 \| ### Framework versions - Transformers 4.32.1 - TensorFlow 2.12.0 - Datasets 2.14.4 - Tokenizers 0.13.3
anshulsc/q-Taxi-v3	anshulsc	2023-08-30T11:17:54Z	0	0	null	[ "Taxi-v3", "q-learning", "reinforcement-learning", "custom-implementation", "model-index", "region:us" ]	reinforcement-learning	2023-08-30T11:17:52Z	--- tags: - Taxi-v3 - q-learning - reinforcement-learning - custom-implementation model-index: - name: q-Taxi-v3 results: - task: type: reinforcement-learning name: reinforcement-learning dataset: name: Taxi-v3 type: Taxi-v3 metrics: - type: mean_reward value: 7.56 +/- 2.71 name: mean_reward verified: false --- # Q-Learning Agent playing1 Taxi-v3 This is a trained model of a Q-Learning agent playing Taxi-v3 . ## Usage ```python model = load_from_hub(repo_id="anshulsc/q-Taxi-v3", filename="q-learning.pkl") # Don't forget to check if you need to add additional attributes (is_slippery=False etc) env = gym.make(model["env_id"]) ```
anshulsc/q-FrozenLake-v1-4x4-noSlippery	anshulsc	2023-08-30T11:11:13Z	0	0	null	[ "FrozenLake-v1-4x4-no_slippery", "q-learning", "reinforcement-learning", "custom-implementation", "model-index", "region:us" ]	reinforcement-learning	2023-08-30T11:06:45Z	--- tags: - FrozenLake-v1-4x4-no_slippery - q-learning - reinforcement-learning - custom-implementation model-index: - name: q-FrozenLake-v1-4x4-noSlippery results: - task: type: reinforcement-learning name: reinforcement-learning dataset: name: FrozenLake-v1-4x4-no_slippery type: FrozenLake-v1-4x4-no_slippery metrics: - type: mean_reward value: 1.00 +/- 0.00 name: mean_reward verified: false --- # Q-Learning Agent playing1 FrozenLake-v1 This is a trained model of a Q-Learning agent playing FrozenLake-v1 . ## Usage ```python model = load_from_hub(repo_id="anshulsc/q-FrozenLake-v1-4x4-noSlippery", filename="q-learning.pkl") # Don't forget to check if you need to add additional attributes (is_slippery=False etc) env = gym.make(model["env_id"]) ```
unrealheinrich/vegetables	unrealheinrich	2023-08-30T11:07:16Z	0	0	null	[ "onnx", "en", "license:mit", "region:us" ]	null	2023-08-30T11:02:02Z	--- license: mit language: - en metrics: - f1 ---
dkqjrm/20230830151725	dkqjrm	2023-08-30T11:06:41Z	104	0	transformers	[ "transformers", "pytorch", "tensorboard", "bert", "text-classification", "generated_from_trainer", "dataset:super_glue", "license:apache-2.0", "autotrain_compatible", "endpoints_compatible", "region:us" ]	text-classification	2023-08-30T06:17:54Z	--- license: apache-2.0 tags: - generated_from_trainer datasets: - super_glue metrics: - accuracy model-index: - name: '20230830151725' results: [] --- <!-- This model card has been generated automatically according to the information the Trainer had access to. You should probably proofread and complete it, then remove this comment. --> # 20230830151725 This model is a fine-tuned version of [bert-large-cased](https://huggingface.co/bert-large-cased) on the super_glue dataset. It achieves the following results on the evaluation set: - Loss: 0.7242 - Accuracy: 0.5 ## Model description More information needed ## Intended uses & limitations More information needed ## Training and evaluation data More information needed ## Training procedure ### Training hyperparameters The following hyperparameters were used during training: - learning_rate: 0.0005 - train_batch_size: 16 - eval_batch_size: 8 - seed: 11 - optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08 - lr_scheduler_type: linear - num_epochs: 80.0 ### Training results \| Training Loss \| Epoch \| Step \| Validation Loss \| Accuracy \| \|:-------------:\|:-----:\|:-----:\|:---------------:\|:--------:\| \| No log \| 1.0 \| 340 \| 0.7462 \| 0.4984 \| \| 0.7527 \| 2.0 \| 680 \| 0.7272 \| 0.5047 \| \| 0.7493 \| 3.0 \| 1020 \| 0.7466 \| 0.5 \| \| 0.7493 \| 4.0 \| 1360 \| 0.7254 \| 0.5 \| \| 0.7484 \| 5.0 \| 1700 \| 0.7484 \| 0.5 \| \| 0.7446 \| 6.0 \| 2040 \| 0.7319 \| 0.5 \| \| 0.7446 \| 7.0 \| 2380 \| 0.7327 \| 0.5 \| \| 0.7385 \| 8.0 \| 2720 \| 0.7423 \| 0.5 \| \| 0.7404 \| 9.0 \| 3060 \| 0.7351 \| 0.4875 \| \| 0.7404 \| 10.0 \| 3400 \| 0.7567 \| 0.5 \| \| 0.7576 \| 11.0 \| 3740 \| 0.7305 \| 0.5 \| \| 0.745 \| 12.0 \| 4080 \| 0.7374 \| 0.5 \| \| 0.745 \| 13.0 \| 4420 \| 0.7735 \| 0.5 \| \| 0.7346 \| 14.0 \| 4760 \| 0.7443 \| 0.5 \| \| 0.7407 \| 15.0 \| 5100 \| 0.7278 \| 0.5 \| \| 0.7407 \| 16.0 \| 5440 \| 0.7490 \| 0.5 \| \| 0.7367 \| 17.0 \| 5780 \| 0.7443 \| 0.5 \| \| 0.7377 \| 18.0 \| 6120 \| 0.7330 \| 0.5 \| \| 0.7377 \| 19.0 \| 6460 \| 0.7249 \| 0.5 \| \| 0.7325 \| 20.0 \| 6800 \| 0.7362 \| 0.5 \| \| 0.7357 \| 21.0 \| 7140 \| 0.7240 \| 0.5 \| \| 0.7357 \| 22.0 \| 7480 \| 0.7470 \| 0.5 \| \| 0.7356 \| 23.0 \| 7820 \| 0.7296 \| 0.5 \| \| 0.734 \| 24.0 \| 8160 \| 0.7255 \| 0.5 \| \| 0.7327 \| 25.0 \| 8500 \| 0.7249 \| 0.5 \| \| 0.7327 \| 26.0 \| 8840 \| 0.7262 \| 0.5 \| \| 0.7345 \| 27.0 \| 9180 \| 0.7743 \| 0.5 \| \| 0.735 \| 28.0 \| 9520 \| 0.7276 \| 0.5 \| \| 0.735 \| 29.0 \| 9860 \| 0.7271 \| 0.5 \| \| 0.7309 \| 30.0 \| 10200 \| 0.7239 \| 0.5 \| \| 0.7345 \| 31.0 \| 10540 \| 0.7241 \| 0.5 \| \| 0.7345 \| 32.0 \| 10880 \| 0.7599 \| 0.5 \| \| 0.733 \| 33.0 \| 11220 \| 0.7245 \| 0.5 \| \| 0.7325 \| 34.0 \| 11560 \| 0.7385 \| 0.5 \| \| 0.7325 \| 35.0 \| 11900 \| 0.7245 \| 0.5 \| \| 0.7318 \| 36.0 \| 12240 \| 0.7242 \| 0.5 \| \| 0.7308 \| 37.0 \| 12580 \| 0.7239 \| 0.5 \| \| 0.7308 \| 38.0 \| 12920 \| 0.7241 \| 0.5 \| \| 0.73 \| 39.0 \| 13260 \| 0.7317 \| 0.5 \| \| 0.7288 \| 40.0 \| 13600 \| 0.7258 \| 0.5 \| \| 0.7288 \| 41.0 \| 13940 \| 0.7241 \| 0.5 \| \| 0.7311 \| 42.0 \| 14280 \| 0.7241 \| 0.5 \| \| 0.7284 \| 43.0 \| 14620 \| 0.7344 \| 0.5 \| \| 0.7284 \| 44.0 \| 14960 \| 0.7297 \| 0.5 \| \| 0.73 \| 45.0 \| 15300 \| 0.7393 \| 0.5 \| \| 0.7269 \| 46.0 \| 15640 \| 0.7239 \| 0.5 \| \| 0.7269 \| 47.0 \| 15980 \| 0.7282 \| 0.5 \| \| 0.7284 \| 48.0 \| 16320 \| 0.7240 \| 0.5 \| \| 0.729 \| 49.0 \| 16660 \| 0.7343 \| 0.5 \| \| 0.7281 \| 50.0 \| 17000 \| 0.7240 \| 0.5 \| \| 0.7281 \| 51.0 \| 17340 \| 0.7245 \| 0.5 \| \| 0.7264 \| 52.0 \| 17680 \| 0.7291 \| 0.5 \| \| 0.7294 \| 53.0 \| 18020 \| 0.7255 \| 0.5 \| \| 0.7294 \| 54.0 \| 18360 \| 0.7251 \| 0.5 \| \| 0.7263 \| 55.0 \| 18700 \| 0.7256 \| 0.5 \| \| 0.7255 \| 56.0 \| 19040 \| 0.7294 \| 0.5 \| \| 0.7255 \| 57.0 \| 19380 \| 0.7242 \| 0.5 \| \| 0.7261 \| 58.0 \| 19720 \| 0.7243 \| 0.5 \| \| 0.7265 \| 59.0 \| 20060 \| 0.7315 \| 0.5 \| \| 0.7265 \| 60.0 \| 20400 \| 0.7239 \| 0.5 \| \| 0.7253 \| 61.0 \| 20740 \| 0.7246 \| 0.5 \| \| 0.7265 \| 62.0 \| 21080 \| 0.7248 \| 0.5 \| \| 0.7265 \| 63.0 \| 21420 \| 0.7244 \| 0.5 \| \| 0.7247 \| 64.0 \| 21760 \| 0.7250 \| 0.5 \| \| 0.7273 \| 65.0 \| 22100 \| 0.7240 \| 0.5 \| \| 0.7273 \| 66.0 \| 22440 \| 0.7251 \| 0.5 \| \| 0.7233 \| 67.0 \| 22780 \| 0.7239 \| 0.5 \| \| 0.7268 \| 68.0 \| 23120 \| 0.7239 \| 0.5 \| \| 0.7268 \| 69.0 \| 23460 \| 0.7251 \| 0.5 \| \| 0.7246 \| 70.0 \| 23800 \| 0.7241 \| 0.5 \| \| 0.7249 \| 71.0 \| 24140 \| 0.7242 \| 0.5 \| \| 0.7249 \| 72.0 \| 24480 \| 0.7259 \| 0.5 \| \| 0.7241 \| 73.0 \| 24820 \| 0.7247 \| 0.5 \| \| 0.7237 \| 74.0 \| 25160 \| 0.7257 \| 0.5 \| \| 0.7244 \| 75.0 \| 25500 \| 0.7245 \| 0.5 \| \| 0.7244 \| 76.0 \| 25840 \| 0.7251 \| 0.5 \| \| 0.7234 \| 77.0 \| 26180 \| 0.7240 \| 0.5 \| \| 0.7231 \| 78.0 \| 26520 \| 0.7243 \| 0.5 \| \| 0.7231 \| 79.0 \| 26860 \| 0.7243 \| 0.5 \| \| 0.7235 \| 80.0 \| 27200 \| 0.7242 \| 0.5 \| ### Framework versions - Transformers 4.26.1 - Pytorch 2.0.1+cu118 - Datasets 2.12.0 - Tokenizers 0.13.3
Pemaled/FraudGuard	Pemaled	2023-08-30T11:04:27Z	0	0	null	[ "license:creativeml-openrail-m", "region:us" ]	null	2023-08-30T11:04:27Z	--- license: creativeml-openrail-m ---
dhiruHF/llama2-docqa-v2-downloaded	dhiruHF	2023-08-30T11:02:44Z	12	0	peft	[ "peft", "autotrain", "text-generation", "region:us" ]	text-generation	2023-08-30T11:02:42Z	--- tags: - autotrain - text-generation widget: - text: "I love AutoTrain because " library_name: peft --- # Model Trained Using AutoTrain## Training procedure The following `bitsandbytes` quantization config was used during training: - load_in_8bit: False - load_in_4bit: True - llm_int8_threshold: 6.0 - llm_int8_skip_modules: None - llm_int8_enable_fp32_cpu_offload: False - llm_int8_has_fp16_weight: False - bnb_4bit_quant_type: nf4 - bnb_4bit_use_double_quant: True - bnb_4bit_compute_dtype: bfloat16 ### Framework versions - PEFT 0.4.0
aggie/edos-alpaca-option-llama-demo	aggie	2023-08-30T10:59:20Z	1	0	peft	[ "peft", "region:us" ]	null	2023-08-18T16:51:22Z	--- library_name: peft --- ## Training procedure ### Framework versions - PEFT 0.5.0.dev0
aggie/edos-alpaca-label-llama-demo	aggie	2023-08-30T10:58:07Z	0	0	peft	[ "peft", "region:us" ]	null	2023-08-30T10:48:38Z	--- library_name: peft --- ## Training procedure ### Framework versions - PEFT 0.5.0.dev0
aggie/edos-alpaca-label-llama	aggie	2023-08-30T10:57:38Z	2	0	peft	[ "peft", "region:us" ]	null	2023-08-30T10:48:11Z	--- library_name: peft --- ## Training procedure ### Framework versions - PEFT 0.5.0.dev0
ShekDass/donut-base-cord-hifi-100	ShekDass	2023-08-30T10:53:16Z	3	0	transformers	[ "transformers", "pytorch", "vision-encoder-decoder", "image-text-to-text", "generated_from_trainer", "dataset:imagefolder", "base_model:naver-clova-ix/donut-base-finetuned-cord-v2", "base_model:finetune:naver-clova-ix/donut-base-finetuned-cord-v2", "license:mit", "endpoints_compatible", "region:us" ]	image-text-to-text	2023-08-29T14:01:24Z	--- license: mit base_model: naver-clova-ix/donut-base-finetuned-cord-v2 tags: - generated_from_trainer datasets: - imagefolder model-index: - name: donut-base-cord-hifi-100 results: [] --- <!-- This model card has been generated automatically according to the information the Trainer had access to. You should probably proofread and complete it, then remove this comment. --> # donut-base-cord-hifi-100 This model is a fine-tuned version of [naver-clova-ix/donut-base-finetuned-cord-v2](https://huggingface.co/naver-clova-ix/donut-base-finetuned-cord-v2) on the imagefolder dataset. ## Model description More information needed ## Intended uses & limitations More information needed ## Training and evaluation data More information needed ## Training procedure ### Training hyperparameters The following hyperparameters were used during training: - learning_rate: 2e-05 - train_batch_size: 1 - eval_batch_size: 8 - seed: 42 - optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08 - lr_scheduler_type: linear - num_epochs: 20 ### Training results ### Framework versions - Transformers 4.32.1 - Pytorch 2.0.1+cu118 - Datasets 2.14.4 - Tokenizers 0.13.3
Snowad/French-Tortoise	Snowad	2023-08-30T10:42:57Z	0	28	null	[ "TTS", "text-to-speech", "fr", "license:apache-2.0", "region:us" ]	text-to-speech	2023-03-04T18:48:14Z	--- license: apache-2.0 language: - fr pipeline_tag: text-to-speech tags: - TTS - text-to-speech --- V2.5 Model : Fine tune of my V2 model on all CommonVoice dataset (517k sample) on 2.5k step (batch size 200), Voice cloning has improved a bit but is still not great. However, if you fine tune this model on your own personality dataset then you can get pretty good results. A good V3 model would be to fine tune for like 50k steps on this dataset and I think there would be a way to get good results but I won't try V2 Model : Tortoise base model Fine tuned on a custom multispeaker French dataset of 120k samples (SIWIS + Common Voice subset + M-AILABS) on 10k step with a RTX 3090 (~= 21 hours of training), with Text LR Weight at 1 Result : The model can speak French much better without an English accent but the voice clone hardly works V1 Model : Tortoise base model Fine tuned on a custom multispeaker French dataset of 24k samples (SIWIS + Common Voice subset) on 8850 step with a RTX 3090 (~= 19 hours of training) Inference : * You can use the model by downloading the "V2_9750_gpt.pth" model and use it in the tortoise-tts optimized forks (git.ecker.tech/mrq/ai-voice-cloning \| 152334H/tortoise-tts-fast) Fine tuning : * I used 152334H/DL-Art-School for training, if you want to resume training from my epoch, follow its documentation and download "V2_9750.state"
jankovicsandras/Llama-2-13b-chat-norwegian-Q5_K_M-GGUF	jankovicsandras	2023-08-30T10:42:29Z	4	0	null	[ "gguf", "no", "nb", "nn", "norsk", "norwegian", "llama2", "q5_k_m", "en", "license:other", "endpoints_compatible", "region:us" ]	null	2023-08-30T07:06:18Z	--- language: - no - nb - nn - en tags: - no - nb - nn - norsk - norwegian - llama2 - gguf - q5_k_m license: "other" --- # This is the Q5_K_M quantized GGUF format ( see https://github.com/ggerganov/llama.cpp ) version of this model: https://huggingface.co/RuterNorway/Llama-2-13b-chat-norwegian Description (copied and edited from RuterNorway/Llama-2-13b-chat-norwegian) : # Llama 2 13b Chat Norwegian Llama-2-13b-chat-norwegian is a variant of Meta´s Llama 2 13b Chat model, finetuned on a mix of norwegian datasets created in Ruter AI Lab the summer of 2023. The model is tuned to understand and generate text in Norwegian. It's trained for one epoch on norwegian-alpaca + 15000 samples of machine-translated data from OpenOrca. A small subset of custom-made instructional data is also included. For other versions of this model see: https://huggingface.co/RuterNorway/Llama-2-13b-chat-norwegian ## Data - Norwegian alpaca - 15k Norwegian OpenOrcra (to be released) - Small subset of custom made instructional data ## Limitations - This is an LLM, not a knowledge model. It can not be expected to have more information about Norway than the basemodel. - It will generally preform better on tasks that involves summarization, question answering and chat, than on tasks that requires more knowledge about Norway, specific domains, or tasks where the model can answer freely. - The data used for training is machine translated, and may contain grammatical errors and other errors. - The model is released as is, and would in most cases need prompt tuning to achieve optimal results. ## License Llama 2 is licensed under the LLAMA 2 Community License, Copyright © Meta Platforms, Inc. All Rights Reserved. See the original model card for more information. From norwegian-alpaca we also note that "the current version uses OpenAI's gpt-3.5-turbo; hence, this dataset cannot be used to create models that compete in any way against OpenAI." Disclaimer ## The model is available "as is". Ruter As takes no responsibility for further use. ## During testing, it seems that the safeguards implemented by Meta, still work as expected in this model. However, we want to point to the Ethical Considerations and Limitations from the origenal model card: Llama 2 is a new technology that carries risks with use. Testing conducted to date has been in English, and has not covered, nor could it cover all scenarios. For these reasons, as with all LLMs, Llama 2’s potential outputs cannot be predicted in advance, and the model may in some instances produce inaccurate, biased or other objectionable responses to user prompts. Therefore, before deploying any applications of Llama 2, developers should perform safety testing and tuning tailored to their specific applications of the model. Please see the Responsible Use Guide available at https://ai.meta.com/llama/responsible-use-guide/ ## Credits This model was made at Ruters AI Lab which is a part of Ruters Data & AI division.
1Poireau/Doully_Millet	1Poireau	2023-08-30T10:23:36Z	0	0	null	[ "license:openrail++", "region:us" ]	null	2023-08-30T10:07:32Z	--- license: openrail++ --- La voix de Doully Millet du Groland, créé par 1Poireau le 30/08/2023. Entrainé pendant 8H sur 31min d'audio que j'ai trouvé sur la chaine Youtube officielle du Groland, avec 300 epochs The voice of Doully Millet from Groland, created by 1Poireau on 30/08/2023. Trained for 8H on 31min audio that i found on Groland's official Youtube channel, with 300 epochs
vishnuhaasan/ppo-lunarlanderv2	vishnuhaasan	2023-08-30T10:20:21Z	0	0	stable-baselines3	[ "stable-baselines3", "LunarLander-v2", "deep-reinforcement-learning", "reinforcement-learning", "model-index", "region:us" ]	reinforcement-learning	2023-08-30T10:19:45Z	--- library_name: stable-baselines3 tags: - LunarLander-v2 - deep-reinforcement-learning - reinforcement-learning - stable-baselines3 model-index: - name: PPO results: - task: type: reinforcement-learning name: reinforcement-learning dataset: name: LunarLander-v2 type: LunarLander-v2 metrics: - type: mean_reward value: 206.33 +/- 94.63 name: mean_reward verified: false --- # PPO Agent playing LunarLander-v2 This is a trained model of a PPO agent playing LunarLander-v2 using the [stable-baselines3 library](https://github.com/DLR-RM/stable-baselines3). ## Usage (with Stable-baselines3) TODO: Add your code ```python from stable_baselines3 import ... from huggingface_sb3 import load_from_hub ... ```
hoang14/chatbot_29_8	hoang14	2023-08-30T10:10:02Z	0	0	null	[ "region:us" ]	null	2023-08-29T02:57:40Z	ATASET = "task-focus + sample from remain datasets" DATASET_FORMAT = 'input-output' PER_DEVICE_TRAIN_BATCH_SIZE = 2 GRADIENT_ACCUMULATION_STEPS = 4 LEARNING_RATE = 0.0003 LR_SCHEDULER_TYPE = 'cosine' WARMUP_RATIO = 0.03 LORA_R = 192 LORA_ALPHA = 64 LORA_DROPOUT = 0.1 TRAIN_ON_SOURCE = False SOURCE_MAX_LENGTH = 1024 TARGET_MAX_LENGTH = 1024 LOGGING_STEPS = 20 SAVE_STEPS = 100 SAVE_TOTAL_LIMIT = 4
eliept1/ppo-SnowballTarget	eliept1	2023-08-30T10:07:23Z	0	0	ml-agents	[ "ml-agents", "tensorboard", "onnx", "SnowballTarget", "deep-reinforcement-learning", "reinforcement-learning", "ML-Agents-SnowballTarget", "region:us" ]	reinforcement-learning	2023-08-30T10:07:19Z	--- library_name: ml-agents tags: - SnowballTarget - deep-reinforcement-learning - reinforcement-learning - ML-Agents-SnowballTarget --- # ppo Agent playing SnowballTarget This is a trained model of a ppo agent playing SnowballTarget using the [Unity ML-Agents Library](https://github.com/Unity-Technologies/ml-agents). ## Usage (with ML-Agents) The Documentation: https://unity-technologies.github.io/ml-agents/ML-Agents-Toolkit-Documentation/ We wrote a complete tutorial to learn to train your first agent using ML-Agents and publish it to the Hub: - A short tutorial where you teach Huggy the Dog 🐶 to fetch the stick and then play with him directly in your browser: https://huggingface.co/learn/deep-rl-course/unitbonus1/introduction - A longer tutorial to understand how works ML-Agents: https://huggingface.co/learn/deep-rl-course/unit5/introduction ### Resume the training ```bash mlagents-learn <your_configuration_file_path.yaml> --run-id=<run_id> --resume ``` ### Watch your Agent play You can watch your agent playing directly in your browser 1. If the environment is part of ML-Agents official environments, go to https://huggingface.co/unity 2. Step 1: Find your model_id: eliept1/ppo-SnowballTarget 3. Step 2: Select your .nn /.onnx file 4. Click on Watch the agent play 👀
julio030/ppo-Huggy	julio030	2023-08-30T10:02:14Z	3	0	ml-agents	[ "ml-agents", "tensorboard", "onnx", "Huggy", "deep-reinforcement-learning", "reinforcement-learning", "ML-Agents-Huggy", "region:us" ]	reinforcement-learning	2023-08-30T10:02:04Z	--- library_name: ml-agents tags: - Huggy - deep-reinforcement-learning - reinforcement-learning - ML-Agents-Huggy --- # ppo Agent playing Huggy This is a trained model of a ppo agent playing Huggy using the [Unity ML-Agents Library](https://github.com/Unity-Technologies/ml-agents). ## Usage (with ML-Agents) The Documentation: https://unity-technologies.github.io/ml-agents/ML-Agents-Toolkit-Documentation/ We wrote a complete tutorial to learn to train your first agent using ML-Agents and publish it to the Hub: - A short tutorial where you teach Huggy the Dog 🐶 to fetch the stick and then play with him directly in your browser: https://huggingface.co/learn/deep-rl-course/unitbonus1/introduction - A longer tutorial to understand how works ML-Agents: https://huggingface.co/learn/deep-rl-course/unit5/introduction ### Resume the training ```bash mlagents-learn <your_configuration_file_path.yaml> --run-id=<run_id> --resume ``` ### Watch your Agent play You can watch your agent playing directly in your browser 1. If the environment is part of ML-Agents official environments, go to https://huggingface.co/unity 2. Step 1: Find your model_id: julio030/ppo-Huggy 3. Step 2: Select your .nn /.onnx file 4. Click on Watch the agent play 👀
ibrahimciko/q-taxi-v3	ibrahimciko	2023-08-30T09:44:46Z	0	0	null	[ "Taxi-v3", "q-learning", "reinforcement-learning", "custom-implementation", "model-index", "region:us" ]	reinforcement-learning	2023-08-30T09:44:41Z	--- tags: - Taxi-v3 - q-learning - reinforcement-learning - custom-implementation model-index: - name: q-taxi-v3 results: - task: type: reinforcement-learning name: reinforcement-learning dataset: name: Taxi-v3 type: Taxi-v3 metrics: - type: mean_reward value: 7.56 +/- 2.71 name: mean_reward verified: false --- # Q-Learning Agent playing1 Taxi-v3 This is a trained model of a Q-Learning agent playing Taxi-v3 . ## Usage ```python model = load_from_hub(repo_id="ibrahimciko/q-taxi-v3", filename="q-learning.pkl") # Don't forget to check if you need to add additional attributes (is_slippery=False etc) env = gym.make(model["env_id"]) ```
vnktrmnb/MBERT_FT-TyDiQA_S311	vnktrmnb	2023-08-30T09:43:41Z	61	0	transformers	[ "transformers", "tf", "tensorboard", "bert", "question-answering", "generated_from_keras_callback", "base_model:google-bert/bert-base-multilingual-cased", "base_model:finetune:google-bert/bert-base-multilingual-cased", "license:apache-2.0", "endpoints_compatible", "region:us" ]	question-answering	2023-08-30T08:49:08Z	--- license: apache-2.0 base_model: bert-base-multilingual-cased tags: - generated_from_keras_callback model-index: - name: vnktrmnb/MBERT_FT-TyDiQA_S311 results: [] --- <!-- This model card has been generated automatically according to the information Keras had access to. You should probably proofread and complete it, then remove this comment. --> # vnktrmnb/MBERT_FT-TyDiQA_S311 This model is a fine-tuned version of [bert-base-multilingual-cased](https://huggingface.co/bert-base-multilingual-cased) on an unknown dataset. It achieves the following results on the evaluation set: - Train Loss: 0.6000 - Train End Logits Accuracy: 0.8451 - Train Start Logits Accuracy: 0.8711 - Validation Loss: 0.4689 - Validation End Logits Accuracy: 0.8686 - Validation Start Logits Accuracy: 0.9137 - Epoch: 2 ## Model description More information needed ## Intended uses & limitations More information needed ## Training and evaluation data More information needed ## Training procedure ### Training hyperparameters The following hyperparameters were used during training: - optimizer: {'name': 'Adam', 'weight_decay': None, 'clipnorm': None, 'global_clipnorm': None, 'clipvalue': None, 'use_ema': False, 'ema_momentum': 0.99, 'ema_overwrite_frequency': None, 'jit_compile': True, 'is_legacy_optimizer': False, 'learning_rate': {'class_name': 'PolynomialDecay', 'config': {'initial_learning_rate': 2e-05, 'decay_steps': 2412, 'end_learning_rate': 0.0, 'power': 1.0, 'cycle': False, 'name': None}}, 'beta_1': 0.9, 'beta_2': 0.999, 'epsilon': 1e-08, 'amsgrad': False} - training_precision: float32 ### Training results \| Train Loss \| Train End Logits Accuracy \| Train Start Logits Accuracy \| Validation Loss \| Validation End Logits Accuracy \| Validation Start Logits Accuracy \| Epoch \| \|:----------:\|:-------------------------:\|:---------------------------:\|:---------------:\|:------------------------------:\|:--------------------------------:\|:-----:\| \| 1.4308 \| 0.6683 \| 0.6988 \| 0.4811 \| 0.8647 \| 0.9072 \| 0 \| \| 0.8301 \| 0.7904 \| 0.8263 \| 0.4455 \| 0.8698 \| 0.9111 \| 1 \| \| 0.6000 \| 0.8451 \| 0.8711 \| 0.4689 \| 0.8686 \| 0.9137 \| 2 \| ### Framework versions - Transformers 4.32.1 - TensorFlow 2.12.0 - Datasets 2.14.4 - Tokenizers 0.13.3
ThuyNT03/xlm-roberta-base-Final_VietNam-aug_replace_synonym	ThuyNT03	2023-08-30T09:39:50Z	104	0	transformers	[ "transformers", "pytorch", "xlm-roberta", "text-classification", "generated_from_trainer", "base_model:FacebookAI/xlm-roberta-base", "base_model:finetune:FacebookAI/xlm-roberta-base", "license:mit", "autotrain_compatible", "endpoints_compatible", "region:us" ]	text-classification	2023-08-30T09:32:37Z	--- license: mit base_model: xlm-roberta-base tags: - generated_from_trainer metrics: - accuracy - f1 model-index: - name: xlm-roberta-base-Final_VietNam-aug_replace_synonym results: [] --- <!-- This model card has been generated automatically according to the information the Trainer had access to. You should probably proofread and complete it, then remove this comment. --> # xlm-roberta-base-Final_VietNam-aug_replace_synonym This model is a fine-tuned version of [xlm-roberta-base](https://huggingface.co/xlm-roberta-base) on the None dataset. It achieves the following results on the evaluation set: - Loss: 1.1196 - Accuracy: 0.74 - F1: 0.7476 ## Model description More information needed ## Intended uses & limitations More information needed ## Training and evaluation data More information needed ## Training procedure ### Training hyperparameters The following hyperparameters were used during training: - learning_rate: 2e-05 - train_batch_size: 16 - eval_batch_size: 16 - seed: 42 - optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08 - lr_scheduler_type: linear - num_epochs: 8 ### Training results \| Training Loss \| Epoch \| Step \| Validation Loss \| Accuracy \| F1 \| \|:-------------:\|:-----:\|:----:\|:---------------:\|:--------:\|:------:\| \| 1.0236 \| 1.0 \| 86 \| 0.8388 \| 0.66 \| 0.6569 \| \| 0.698 \| 2.0 \| 172 \| 0.7581 \| 0.71 \| 0.7064 \| \| 0.5136 \| 3.0 \| 258 \| 0.7604 \| 0.74 \| 0.7348 \| \| 0.3853 \| 4.0 \| 344 \| 0.8044 \| 0.75 \| 0.7563 \| \| 0.2299 \| 5.0 \| 430 \| 0.7855 \| 0.74 \| 0.7484 \| \| 0.2054 \| 6.0 \| 516 \| 0.9726 \| 0.74 \| 0.7459 \| \| 0.1438 \| 7.0 \| 602 \| 1.1100 \| 0.73 \| 0.7386 \| \| 0.108 \| 8.0 \| 688 \| 1.1196 \| 0.74 \| 0.7476 \| ### Framework versions - Transformers 4.32.1 - Pytorch 2.0.0 - Datasets 2.14.4 - Tokenizers 0.13.3
AdanLee/poca-SoccerTwos	AdanLee	2023-08-30T09:27:16Z	1	0	ml-agents	[ "ml-agents", "tensorboard", "onnx", "SoccerTwos", "deep-reinforcement-learning", "reinforcement-learning", "ML-Agents-SoccerTwos", "region:us" ]	reinforcement-learning	2023-08-30T09:26:09Z	--- library_name: ml-agents tags: - SoccerTwos - deep-reinforcement-learning - reinforcement-learning - ML-Agents-SoccerTwos --- # poca Agent playing SoccerTwos This is a trained model of a poca agent playing SoccerTwos using the [Unity ML-Agents Library](https://github.com/Unity-Technologies/ml-agents). ## Usage (with ML-Agents) The Documentation: https://unity-technologies.github.io/ml-agents/ML-Agents-Toolkit-Documentation/ We wrote a complete tutorial to learn to train your first agent using ML-Agents and publish it to the Hub: - A short tutorial where you teach Huggy the Dog 🐶 to fetch the stick and then play with him directly in your browser: https://huggingface.co/learn/deep-rl-course/unitbonus1/introduction - A longer tutorial to understand how works ML-Agents: https://huggingface.co/learn/deep-rl-course/unit5/introduction ### Resume the training ```bash mlagents-learn <your_configuration_file_path.yaml> --run-id=<run_id> --resume ``` ### Watch your Agent play You can watch your agent playing directly in your browser 1. If the environment is part of ML-Agents official environments, go to https://huggingface.co/unity 2. Step 1: Find your model_id: AdanLee/poca-SoccerTwos 3. Step 2: Select your .nn /.onnx file 4. Click on Watch the agent play 👀
ThuyNT03/xlm-roberta-base-Final_VietNam-aug_insert_tfidf	ThuyNT03	2023-08-30T09:25:39Z	124	0	transformers	[ "transformers", "pytorch", "xlm-roberta", "text-classification", "generated_from_trainer", "base_model:FacebookAI/xlm-roberta-base", "base_model:finetune:FacebookAI/xlm-roberta-base", "license:mit", "autotrain_compatible", "endpoints_compatible", "region:us" ]	text-classification	2023-08-30T08:45:51Z	--- license: mit base_model: xlm-roberta-base tags: - generated_from_trainer metrics: - accuracy - f1 model-index: - name: xlm-roberta-base-Final_VietNam-aug_insert_tfidf results: [] --- <!-- This model card has been generated automatically according to the information the Trainer had access to. You should probably proofread and complete it, then remove this comment. --> # xlm-roberta-base-Final_VietNam-aug_insert_tfidf This model is a fine-tuned version of [xlm-roberta-base](https://huggingface.co/xlm-roberta-base) on the None dataset. It achieves the following results on the evaluation set: - Loss: 1.2753 - Accuracy: 0.71 - F1: 0.7161 ## Model description More information needed ## Intended uses & limitations More information needed ## Training and evaluation data More information needed ## Training procedure ### Training hyperparameters The following hyperparameters were used during training: - learning_rate: 2e-05 - train_batch_size: 16 - eval_batch_size: 16 - seed: 42 - optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08 - lr_scheduler_type: linear - num_epochs: 8 ### Training results \| Training Loss \| Epoch \| Step \| Validation Loss \| Accuracy \| F1 \| \|:-------------:\|:-----:\|:----:\|:---------------:\|:--------:\|:------:\| \| 1.0036 \| 1.0 \| 87 \| 0.7585 \| 0.63 \| 0.6038 \| \| 0.6784 \| 2.0 \| 174 \| 0.6901 \| 0.7 \| 0.6939 \| \| 0.5132 \| 3.0 \| 261 \| 0.6510 \| 0.76 \| 0.7658 \| \| 0.3868 \| 4.0 \| 348 \| 0.7266 \| 0.74 \| 0.7436 \| \| 0.2694 \| 5.0 \| 435 \| 0.8702 \| 0.72 \| 0.7264 \| \| 0.1805 \| 6.0 \| 522 \| 1.1744 \| 0.72 \| 0.7207 \| \| 0.1813 \| 7.0 \| 609 \| 1.2328 \| 0.72 \| 0.7256 \| \| 0.1258 \| 8.0 \| 696 \| 1.2753 \| 0.71 \| 0.7161 \| ### Framework versions - Transformers 4.32.1 - Pytorch 2.0.0 - Datasets 2.14.4 - Tokenizers 0.13.3
aviroes/elderly_whisper-medium-LoRA	aviroes	2023-08-30T09:11:57Z	0	0	null	[ "generated_from_trainer", "base_model:openai/whisper-medium", "base_model:finetune:openai/whisper-medium", "license:apache-2.0", "region:us" ]	null	2023-08-30T07:18:13Z	--- license: apache-2.0 base_model: openai/whisper-medium tags: - generated_from_trainer model-index: - name: elderly_whisper-medium-LoRA results: [] --- <!-- This model card has been generated automatically according to the information the Trainer had access to. You should probably proofread and complete it, then remove this comment. --> # elderly_whisper-medium-LoRA This model is a fine-tuned version of [openai/whisper-medium](https://huggingface.co/openai/whisper-medium) on the None dataset. ## Model description More information needed ## Intended uses & limitations More information needed ## Training and evaluation data More information needed ## Training procedure ### Training hyperparameters The following hyperparameters were used during training: - learning_rate: 0.001 - train_batch_size: 8 - eval_batch_size: 8 - seed: 42 - optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08 - lr_scheduler_type: linear - lr_scheduler_warmup_steps: 200 - num_epochs: 3 ### Framework versions - Transformers 4.33.0.dev0 - Pytorch 2.0.1+cu118 - Datasets 2.14.4 - Tokenizers 0.13.3
ldhldh/12.8b_lora	ldhldh	2023-08-30T09:10:10Z	0	0	peft	[ "peft", "region:us" ]	null	2023-08-28T10:05:47Z	--- library_name: peft --- ## Training procedure The following `bitsandbytes` quantization config was used during training: - quant_method: bitsandbytes - load_in_8bit: False - load_in_4bit: True - llm_int8_threshold: 6.0 - llm_int8_skip_modules: None - llm_int8_enable_fp32_cpu_offload: False - llm_int8_has_fp16_weight: False - bnb_4bit_quant_type: nf4 - bnb_4bit_use_double_quant: False - bnb_4bit_compute_dtype: bfloat16 ### Framework versions - PEFT 0.6.0.dev0
AbdelKarim95/ppo-Huggy	AbdelKarim95	2023-08-30T08:39:48Z	0	0	ml-agents	[ "ml-agents", "tensorboard", "onnx", "Huggy", "deep-reinforcement-learning", "reinforcement-learning", "ML-Agents-Huggy", "region:us" ]	reinforcement-learning	2023-08-30T07:53:50Z	--- library_name: ml-agents tags: - Huggy - deep-reinforcement-learning - reinforcement-learning - ML-Agents-Huggy --- # ppo Agent playing Huggy This is a trained model of a ppo agent playing Huggy using the [Unity ML-Agents Library](https://github.com/Unity-Technologies/ml-agents). ## Usage (with ML-Agents) The Documentation: https://unity-technologies.github.io/ml-agents/ML-Agents-Toolkit-Documentation/ We wrote a complete tutorial to learn to train your first agent using ML-Agents and publish it to the Hub: - A short tutorial where you teach Huggy the Dog 🐶 to fetch the stick and then play with him directly in your browser: https://huggingface.co/learn/deep-rl-course/unitbonus1/introduction - A longer tutorial to understand how works ML-Agents: https://huggingface.co/learn/deep-rl-course/unit5/introduction ### Resume the training ```bash mlagents-learn <your_configuration_file_path.yaml> --run-id=<run_id> --resume ``` ### Watch your Agent play You can watch your agent playing directly in your browser 1. If the environment is part of ML-Agents official environments, go to https://huggingface.co/unity 2. Step 1: Find your model_id: AbdelKarim95/ppo-Huggy 3. Step 2: Select your .nn /.onnx file 4. Click on Watch the agent play 👀
jasanfa/distilbert-base-uncased-finetuned-emotion	jasanfa	2023-08-30T08:37:10Z	106	0	transformers	[ "transformers", "pytorch", "distilbert", "text-classification", "generated_from_trainer", "base_model:distilbert/distilbert-base-uncased", "base_model:finetune:distilbert/distilbert-base-uncased", "license:apache-2.0", "autotrain_compatible", "endpoints_compatible", "region:us" ]	text-classification	2023-08-29T14:43:00Z	--- license: apache-2.0 base_model: distilbert-base-uncased tags: - generated_from_trainer metrics: - accuracy - f1 model-index: - name: distilbert-base-uncased-finetuned-emotion results: [] --- <!-- This model card has been generated automatically according to the information the Trainer had access to. You should probably proofread and complete it, then remove this comment. --> # distilbert-base-uncased-finetuned-emotion This model is a fine-tuned version of [distilbert-base-uncased](https://huggingface.co/distilbert-base-uncased) on an unknown dataset. It achieves the following results on the evaluation set: - Loss: 0.2165 - Accuracy: 0.9195 - F1: 0.9197 ## Model description More information needed ## Intended uses & limitations More information needed ## Training and evaluation data More information needed ## Training procedure ### Training hyperparameters The following hyperparameters were used during training: - learning_rate: 2e-05 - train_batch_size: 64 - eval_batch_size: 64 - seed: 42 - optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08 - lr_scheduler_type: linear - num_epochs: 2 ### Training results \| Training Loss \| Epoch \| Step \| Validation Loss \| Accuracy \| F1 \| \|:-------------:\|:-----:\|:----:\|:---------------:\|:--------:\|:------:\| \| 0.8388 \| 1.0 \| 250 \| 0.3212 \| 0.906 \| 0.9049 \| \| 0.251 \| 2.0 \| 500 \| 0.2165 \| 0.9195 \| 0.9197 \| ### Framework versions - Transformers 4.32.1 - Pytorch 2.0.1+cu118 - Tokenizers 0.13.3
yuanxing/Inuyasha	yuanxing	2023-08-30T08:36:02Z	0	0	null	[ "region:us" ]	null	2023-08-30T08:26:57Z	Inuyasha_犬夜叉 ![image/png](https://cdn-uploads.huggingface.co/production/uploads/64326351b0dee98d1779e08b/4kM2o96wf3D337MtdAVG0.png) ![image/png](https://cdn-uploads.huggingface.co/production/uploads/64326351b0dee98d1779e08b/C-u9z8AMIZKwaEMRIJQWp.png)
rasha-salim/autotrain-dreambooth-512	rasha-salim	2023-08-30T08:34:29Z	2	1	diffusers	[ "diffusers", "text-to-image", "autotrain", "base_model:stabilityai/stable-diffusion-xl-base-1.0", "base_model:finetune:stabilityai/stable-diffusion-xl-base-1.0", "region:us" ]	text-to-image	2023-08-30T05:34:35Z	--- base_model: stabilityai/stable-diffusion-xl-base-1.0 instance_prompt: photo of a sks adam tags: - text-to-image - diffusers - autotrain inference: true --- # DreamBooth trained by AutoTrain Text encoder was not trained.
sunkai4u/sd-class-butterflies-32	sunkai4u	2023-08-30T08:26:49Z	46	0	diffusers	[ "diffusers", "safetensors", "pytorch", "unconditional-image-generation", "diffusion-models-class", "license:mit", "diffusers:DDPMPipeline", "region:us" ]	unconditional-image-generation	2023-08-30T08:25:57Z	--- license: mit tags: - pytorch - diffusers - unconditional-image-generation - diffusion-models-class --- # Model Card for Unit 1 of the [Diffusion Models Class 🧨](https://github.com/huggingface/diffusion-models-class) This model is a diffusion model for unconditional image generation of cute 🦋. ## Usage ```python from diffusers import DDPMPipeline pipeline = DDPMPipeline.from_pretrained('sunkai4u/sd-class-butterflies-32') image = pipeline().images[0] image ```
RyyyT/ppo-Huggy	RyyyT	2023-08-30T08:22:39Z	0	0	ml-agents	[ "ml-agents", "tensorboard", "onnx", "Huggy", "deep-reinforcement-learning", "reinforcement-learning", "ML-Agents-Huggy", "region:us" ]	reinforcement-learning	2023-08-30T08:22:27Z	--- library_name: ml-agents tags: - Huggy - deep-reinforcement-learning - reinforcement-learning - ML-Agents-Huggy --- # ppo Agent playing Huggy This is a trained model of a ppo agent playing Huggy using the [Unity ML-Agents Library](https://github.com/Unity-Technologies/ml-agents). ## Usage (with ML-Agents) The Documentation: https://unity-technologies.github.io/ml-agents/ML-Agents-Toolkit-Documentation/ We wrote a complete tutorial to learn to train your first agent using ML-Agents and publish it to the Hub: - A short tutorial where you teach Huggy the Dog 🐶 to fetch the stick and then play with him directly in your browser: https://huggingface.co/learn/deep-rl-course/unitbonus1/introduction - A longer tutorial to understand how works ML-Agents: https://huggingface.co/learn/deep-rl-course/unit5/introduction ### Resume the training ```bash mlagents-learn <your_configuration_file_path.yaml> --run-id=<run_id> --resume ``` ### Watch your Agent play You can watch your agent playing directly in your browser 1. If the environment is part of ML-Agents official environments, go to https://huggingface.co/unity 2. Step 1: Find your model_id: RyyyT/ppo-Huggy 3. Step 2: Select your .nn /.onnx file 4. Click on Watch the agent play 👀
r0llingclouds/distilbert-base-uncased-finetuned-clinc	r0llingclouds	2023-08-30T08:21:12Z	105	0	transformers	[ "transformers", "pytorch", "tensorboard", "distilbert", "text-classification", "generated_from_trainer", "dataset:clinc_oos", "license:apache-2.0", "model-index", "autotrain_compatible", "endpoints_compatible", "region:us" ]	text-classification	2023-08-30T08:17:41Z	--- license: apache-2.0 tags: - generated_from_trainer datasets: - clinc_oos metrics: - accuracy model-index: - name: distilbert-base-uncased-finetuned-clinc results: - task: name: Text Classification type: text-classification dataset: name: clinc_oos type: clinc_oos args: plus metrics: - name: Accuracy type: accuracy value: 0.9161290322580645 --- <!-- This model card has been generated automatically according to the information the Trainer had access to. You should probably proofread and complete it, then remove this comment. --> # distilbert-base-uncased-finetuned-clinc This model is a fine-tuned version of [distilbert-base-uncased](https://huggingface.co/distilbert-base-uncased) on the clinc_oos dataset. It achieves the following results on the evaluation set: - Loss: 0.7754 - Accuracy: 0.9161 ## Model description More information needed ## Intended uses & limitations More information needed ## Training and evaluation data More information needed ## Training procedure ### Training hyperparameters The following hyperparameters were used during training: - learning_rate: 2e-05 - train_batch_size: 48 - eval_batch_size: 48 - seed: 42 - optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08 - lr_scheduler_type: linear - num_epochs: 5 ### Training results \| Training Loss \| Epoch \| Step \| Validation Loss \| Accuracy \| \|:-------------:\|:-----:\|:----:\|:---------------:\|:--------:\| \| 4.2893 \| 1.0 \| 318 \| 3.2831 \| 0.7397 \| \| 2.6289 \| 2.0 \| 636 \| 1.8731 \| 0.8345 \| \| 1.5481 \| 3.0 \| 954 \| 1.1580 \| 0.89 \| \| 1.0137 \| 4.0 \| 1272 \| 0.8584 \| 0.9077 \| \| 0.7969 \| 5.0 \| 1590 \| 0.7754 \| 0.9161 \| ### Framework versions - Transformers 4.16.2 - Pytorch 2.0.1+cu118 - Datasets 1.16.1 - Tokenizers 0.13.3
nishant-glance/model-sd-1-4-priorp-unet-2000-lr2e-ab	nishant-glance	2023-08-30T08:09:13Z	0	0	diffusers	[ "diffusers", "tensorboard", "safetensors", "stable-diffusion", "stable-diffusion-diffusers", "text-to-image", "dreambooth", "base_model:CompVis/stable-diffusion-v1-4", "base_model:finetune:CompVis/stable-diffusion-v1-4", "license:creativeml-openrail-m", "autotrain_compatible", "endpoints_compatible", "diffusers:StableDiffusionPipeline", "region:us" ]	text-to-image	2023-08-29T07:24:54Z	--- license: creativeml-openrail-m base_model: CompVis/stable-diffusion-v1-4 instance_prompt: a photo of yqg person tags: - stable-diffusion - stable-diffusion-diffusers - text-to-image - diffusers - dreambooth inference: true --- # DreamBooth - nishant-glance/model-sd-1-4-priorp-unet-2000-lr2e-ab This is a dreambooth model derived from CompVis/stable-diffusion-v1-4. The weights were trained on a photo of yqg person using [DreamBooth](https://dreambooth.github.io/). You can find some example images in the following. DreamBooth for the text encoder was enabled: True.
IA-hf/sd-pokemon-model-lora-sdxl	IA-hf	2023-08-30T08:04:08Z	1	2	diffusers	[ "diffusers", "stable-diffusion-xl", "stable-diffusion-xl-diffusers", "text-to-image", "lora", "base_model:stabilityai/stable-diffusion-xl-base-1.0", "base_model:adapter:stabilityai/stable-diffusion-xl-base-1.0", "license:creativeml-openrail-m", "region:us" ]	text-to-image	2023-08-30T05:17:25Z	--- license: creativeml-openrail-m base_model: stabilityai/stable-diffusion-xl-base-1.0 dataset: lambdalabs/pokemon-blip-captions tags: - stable-diffusion-xl - stable-diffusion-xl-diffusers - text-to-image - diffusers - lora inference: true --- # LoRA text2image fine-tuning - IA-hf/sd-pokemon-model-lora-sdxl These are LoRA adaption weights for stabilityai/stable-diffusion-xl-base-1.0. The weights were fine-tuned on the lambdalabs/pokemon-blip-captions dataset. You can find some example images in the following. ![img_0](./image_0.png) ![img_1](./image_1.png) ![img_2](./image_2.png) ![img_3](./image_3.png) LoRA for the text encoder was enabled: False. Special VAE used for training: madebyollin/sdxl-vae-fp16-fix.
raruidol/SchemeClassifier-ENG	raruidol	2023-08-30T08:02:50Z	114	0	transformers	[ "transformers", "pytorch", "safetensors", "roberta", "text-classification", "license:cc-by-nc-sa-4.0", "autotrain_compatible", "endpoints_compatible", "region:us" ]	text-classification	2023-08-24T18:37:43Z	--- license: cc-by-nc-sa-4.0 ---
TFMC/ELYZA-japanese-Llama-2-7b-instruct-GPTQ-4bit-64g	TFMC	2023-08-30T07:53:28Z	96	4	transformers	[ "transformers", "safetensors", "llama", "text-generation", "license:llama2", "autotrain_compatible", "text-generation-inference", "endpoints_compatible", "region:us" ]	text-generation	2023-08-30T01:30:19Z	--- license: llama2 --- # ELYZA-japanese-Llama-2-7b-instruct-GPTQ-4bit-64g GPTQ model for ["elyza/ELYZA-japanese-Llama-2-7b-instruct"](https://huggingface.co/elyza/ELYZA-japanese-Llama-2-7b-instruct) : 4bits, gr64, desc_act=True
Idriska/my_awesome_model	Idriska	2023-08-30T07:46:17Z	105	0	transformers	[ "transformers", "pytorch", "distilbert", "text-classification", "generated_from_trainer", "dataset:imdb", "base_model:distilbert/distilbert-base-uncased", "base_model:finetune:distilbert/distilbert-base-uncased", "license:apache-2.0", "model-index", "autotrain_compatible", "endpoints_compatible", "region:us" ]	text-classification	2023-08-29T14:48:43Z	--- license: apache-2.0 base_model: distilbert-base-uncased tags: - generated_from_trainer datasets: - imdb metrics: - accuracy model-index: - name: my_awesome_model results: - task: name: Text Classification type: text-classification dataset: name: imdb type: imdb config: plain_text split: test args: plain_text metrics: - name: Accuracy type: accuracy value: 0.93348 --- <!-- This model card has been generated automatically according to the information the Trainer had access to. You should probably proofread and complete it, then remove this comment. --> # my_awesome_model This model is a fine-tuned version of [distilbert-base-uncased](https://huggingface.co/distilbert-base-uncased) on the imdb dataset. It achieves the following results on the evaluation set: - Loss: 0.2250 - Accuracy: 0.9335 ## Model description More information needed ## Intended uses & limitations More information needed ## Training and evaluation data More information needed ## Training procedure ### Training hyperparameters The following hyperparameters were used during training: - learning_rate: 2e-05 - train_batch_size: 16 - eval_batch_size: 16 - seed: 42 - optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08 - lr_scheduler_type: linear - num_epochs: 2 ### Training results \| Training Loss \| Epoch \| Step \| Validation Loss \| Accuracy \| \|:-------------:\|:-----:\|:----:\|:---------------:\|:--------:\| \| 0.2249 \| 1.0 \| 1563 \| 0.2371 \| 0.9148 \| \| 0.1521 \| 2.0 \| 3126 \| 0.2250 \| 0.9335 \| ### Framework versions - Transformers 4.32.1 - Pytorch 2.0.1+cu118 - Datasets 2.14.4 - Tokenizers 0.13.3
nikinetrahutama/afx-ai-llama-chat-model-21	nikinetrahutama	2023-08-30T07:42:37Z	0	0	peft	[ "peft", "region:us" ]	null	2023-08-30T07:42:29Z	--- library_name: peft --- ## Training procedure The following `bitsandbytes` quantization config was used during training: - quant_method: bitsandbytes - load_in_8bit: False - load_in_4bit: True - llm_int8_threshold: 6.0 - llm_int8_skip_modules: None - llm_int8_enable_fp32_cpu_offload: False - llm_int8_has_fp16_weight: False - bnb_4bit_quant_type: nf4 - bnb_4bit_use_double_quant: True - bnb_4bit_compute_dtype: bfloat16 ### Framework versions - PEFT 0.6.0.dev0
Vineetttt/layoutlmv3-base-finetuned-FUNSD	Vineetttt	2023-08-30T07:29:45Z	126	0	transformers	[ "transformers", "pytorch", "layoutlmv3", "token-classification", "generated_from_trainer", "dataset:funsd-layoutlmv3", "base_model:microsoft/layoutlmv3-base", "base_model:finetune:microsoft/layoutlmv3-base", "license:cc-by-nc-sa-4.0", "model-index", "autotrain_compatible", "endpoints_compatible", "region:us" ]	token-classification	2023-08-30T07:15:58Z	--- license: cc-by-nc-sa-4.0 base_model: microsoft/layoutlmv3-base tags: - generated_from_trainer datasets: - funsd-layoutlmv3 metrics: - precision - recall - f1 - accuracy model-index: - name: test results: - task: name: Token Classification type: token-classification dataset: name: funsd-layoutlmv3 type: funsd-layoutlmv3 config: funsd split: test args: funsd metrics: - name: Precision type: precision value: 0.9002457002457003 - name: Recall type: recall value: 0.9100844510680576 - name: F1 type: f1 value: 0.9051383399209486 - name: Accuracy type: accuracy value: 0.8547486033519553 --- <!-- This model card has been generated automatically according to the information the Trainer had access to. You should probably proofread and complete it, then remove this comment. --> # test This model is a fine-tuned version of [microsoft/layoutlmv3-base](https://huggingface.co/microsoft/layoutlmv3-base) on the funsd-layoutlmv3 dataset. It achieves the following results on the evaluation set: - Loss: 0.6194 - Precision: 0.9002 - Recall: 0.9101 - F1: 0.9051 - Accuracy: 0.8547 ## Model description More information needed ## Intended uses & limitations More information needed ## Training and evaluation data More information needed ## Training procedure ### Training hyperparameters The following hyperparameters were used during training: - learning_rate: 1e-05 - train_batch_size: 2 - eval_batch_size: 2 - seed: 42 - optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08 - lr_scheduler_type: linear - training_steps: 1000 ### Training results \| Training Loss \| Epoch \| Step \| Validation Loss \| Precision \| Recall \| F1 \| Accuracy \| \|:-------------:\|:-----:\|:----:\|:---------------:\|:---------:\|:------:\|:------:\|:--------:\| \| No log \| 1.33 \| 100 \| 0.6953 \| 0.7761 \| 0.8058 \| 0.7906 \| 0.7680 \| \| No log \| 2.67 \| 200 \| 0.5117 \| 0.8250 \| 0.8808 \| 0.8520 \| 0.8290 \| \| No log \| 4.0 \| 300 \| 0.5177 \| 0.8397 \| 0.8897 \| 0.8640 \| 0.8337 \| \| No log \| 5.33 \| 400 \| 0.5165 \| 0.8642 \| 0.9106 \| 0.8868 \| 0.8509 \| \| 0.5653 \| 6.67 \| 500 \| 0.5378 \| 0.8735 \| 0.9091 \| 0.8909 \| 0.8458 \| \| 0.5653 \| 8.0 \| 600 \| 0.5698 \| 0.8733 \| 0.9111 \| 0.8918 \| 0.8482 \| \| 0.5653 \| 9.33 \| 700 \| 0.5773 \| 0.8934 \| 0.9076 \| 0.9004 \| 0.8557 \| \| 0.5653 \| 10.67 \| 800 \| 0.6073 \| 0.8905 \| 0.9006 \| 0.8955 \| 0.8520 \| \| 0.5653 \| 12.0 \| 900 \| 0.6090 \| 0.8940 \| 0.9091 \| 0.9015 \| 0.8513 \| \| 0.1357 \| 13.33 \| 1000 \| 0.6194 \| 0.9002 \| 0.9101 \| 0.9051 \| 0.8547 \| ### Framework versions - Transformers 4.33.0.dev0 - Pytorch 2.0.1+cu118 - Datasets 2.14.4 - Tokenizers 0.13.3
morpheuslord/secllama	morpheuslord	2023-08-30T07:17:51Z	1	2	null	[ "tensorboard", "generated_from_trainer", "region:us" ]	null	2023-07-30T12:22:06Z	--- tags: - generated_from_trainer model-index: - name: secllama results: [] --- <!-- This model card has been generated automatically according to the information the Trainer had access to. You should probably proofread and complete it, then remove this comment. --> # secllama This model is a fine-tuned version of [meta-llama/Llama-2-7b-hf](https://huggingface.co/meta-llama/Llama-2-7b-hf) on an unknown dataset. ## Model description More information needed ## Intended uses & limitations More information needed ## Training and evaluation data More information needed ## Training procedure ### Training hyperparameters The following hyperparameters were used during training: - learning_rate: 0.0002 - train_batch_size: 1 - eval_batch_size: 8 - seed: 42 - gradient_accumulation_steps: 16 - total_train_batch_size: 16 - optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08 - lr_scheduler_type: linear - training_steps: 1 ### Training results ### Framework versions - Transformers 4.30.2 - Pytorch 2.0.1+cu118 - Datasets 2.14.4 - Tokenizers 0.13.3
prashanth07/falcon7b_updated_samples_D	prashanth07	2023-08-30T07:14:51Z	0	0	peft	[ "peft", "region:us" ]	null	2023-08-30T07:14:33Z	--- library_name: peft --- ## Training procedure The following `bitsandbytes` quantization config was used during training: - quant_method: bitsandbytes - load_in_8bit: False - load_in_4bit: True - llm_int8_threshold: 6.0 - llm_int8_skip_modules: None - llm_int8_enable_fp32_cpu_offload: False - llm_int8_has_fp16_weight: False - bnb_4bit_quant_type: nf4 - bnb_4bit_use_double_quant: False - bnb_4bit_compute_dtype: float16 ### Framework versions - PEFT 0.6.0.dev0
Raychani1/slovakbert-ner-v2	Raychani1	2023-08-30T06:59:38Z	111	0	transformers	[ "transformers", "pytorch", "roberta", "token-classification", "generated_from_trainer", "sk", "dataset:NBS_sentence", "license:mit", "model-index", "autotrain_compatible", "region:us" ]	token-classification	2023-08-30T05:42:08Z	--- license: mit language: - sk tags: - generated_from_trainer datasets: - NBS_sentence metrics: - precision - recall - f1 - accuracy inference: false model-index: - name: slovakbert-ner-v2 results: - task: name: Token Classification type: token-classification metrics: - name: Precision type: precision value: 0.9715 - name: Recall type: recall value: 0.9433 - name: F1 type: f1 value: 0.9547 - name: Accuracy type: accuracy value: 0.9897 --- # SlovakBERT based Named Entity Recognition Deep Learning model developed for Named Entity Recognition (NER) in Slovak. The [Gerulata/SlovakBERT](https://huggingface.co/gerulata/slovakbert) based model is fine-tuned on webscraped Slovak news articles. The finished model supports the following IOB tagged entity categories: PERSON, ORGANIZATION, LOCATION, DATE, TIME, MONEY and PERCENTAGE ### Related Work [![Thesis][Thesis]][Thesis-url] ## Model usage ### Simple Named Entity Recognition (NER) ```python from transformers import pipeline ner_pipeline = pipeline(task='ner', model='Raychani1/slovakbert-ner-v2') input_sentence = 'Hoci podľa ostatných údajov NBS pre Bratislavský kraj je aktuálna priemerná cena nehnuteľností na úrovni 2 072 eur za štvorcový meter, ceny bytov v hlavnom meste sú podstatne vyššie.' classifications = ner_pipeline(input_sentence) ``` ### Named Entity Recognition (NER) with Visualization For a Visualization Example please refer to the following [Gist](https://gist.github.com/Raychani1/7d4455491f0aa681ed8ea99d8b1d8279). ### Model Prediction Output Example ![prediction_output](https://github.com/Raychani1/Text_Parsing_Methods_Using_NLP/assets/45550552/723ab7f1-4efb-4d03-87d6-b9ac1e40990f) ## Model Training ### Training Hyperparameters \| Parameter \| Value \| \|:---------------------------:\|:---------:\| \| per_device_train_batch_size \| 4 \| \| per_device_eval_batch_size \| 4 \| \| learning_rate \| 5e-05 \| \| adam_beta1 \| 0.9 \| \| adam_beta1 \| 0.999 \| \| adam_epsilon \| 1e-08 \| \| num_train_epochs \| 15 \| \| lr_scheduler_type \| linear \| \| seed \| 42 \| ### Training results Best model results are reached in the 8th training epoch. \| Training Loss \| Epoch \| Step \| Validation Loss \| Precision \| Recall \| F1 \| Accuracy \| \|:-------------:\|:-----:\|:----:\|:---------------:\|:---------:\|:------:\|:------:\|:--------:\| \| 0.6721 \| 1.0 \| 70 \| 0.2214 \| 0.6972 \| 0.7308 \| 0.7136 \| 0.9324 \| \| 0.1849 \| 2.0 \| 140 \| 0.1697 \| 0.8056 \| 0.8365 \| 0.8208 \| 0.952 \| \| 0.0968 \| 3.0 \| 210 \| 0.1213 \| 0.882 \| 0.8622 \| 0.872 \| 0.9728 \| \| 0.0468 \| 4.0 \| 280 \| 0.1107 \| 0.8372 \| 0.907 \| 0.8708 \| 0.9684 \| \| 0.0415 \| 5.0 \| 350 \| 0.1644 \| 0.8059 \| 0.8782 \| 0.8405 \| 0.9615 \| \| 0.0233 \| 6.0 \| 420 \| 0.1255 \| 0.8576 \| 0.8878 \| 0.8724 \| 0.9716 \| \| 0.0198 \| 7.0 \| 490 \| 0.1383 \| 0.8545 \| 0.8846 \| 0.8693 \| 0.9703 \| \| 0.0133 \| 8.0 \| 560 \| 0.1241 \| 0.884 \| 0.9038 \| 0.8938 \| 0.9735 \| ## Model Evaluation ### Evaluation Dataset Distribution \| NER Tag \| Number of Tokens \| \|:-----------------:\|:--------------------:\| \| 0 \| 6568 \| \| B-Person \| 96 \| \| I-Person \| 83 \| \| B-Organizaton \| 583 \| \| I-Organizaton \| 585 \| \| B-Location \| 59 \| \| I-Location \| 15 \| \| B-Date \| 113 \| \| I-Date \| 87 \| \| Time \| 5 \| \| B-Money \| 44 \| \| I-Money \| 74 \| \| B-Percentage \| 57 \| \| I-Percentage \| 54 \| ### Evaluation Confusion Matrix ![image](https://github.com/Raychani1/Text_Parsing_Methods_Using_NLP/assets/45550552/e6d1a1c6-e02f-4de9-9684-5882a405d31f) ### Evaluation Model Metrics \| Precision \| Macro-Precision \| Recall \| Macro-Recall \| F1 \| Macro-F1 \| Accuracy \| \|:-------------:\|:-------------------:\|:----------:\|:----------------:\|:------:\|:------------:\|:------------:\| \| 0.9897 \| 0.9715 \| 0.9897 \| 0.9433 \| 0.9895 \| 0.9547 \| 0.9897 \| ## Framework Versions - Transformers 4.26.1 - PyTorch 1.13.1 - Tokenizers 0.13.2 <!-- Variables --> [Thesis]: https://img.shields.io/badge/%F0%9F%93%9C-Masters%20Thesis-blue?style=for-the-badge [Thesis-url]: https://opac.crzp.sk/?fn=detailBiblioForm&sid=C0DEB8E07572332BA2230915805F
susmitabhatt/my_awesome_mind_model	susmitabhatt	2023-08-30T06:49:32Z	14	0	transformers	[ "transformers", "pytorch", "tensorboard", "wav2vec2", "audio-classification", "generated_from_trainer", "dataset:marsyas/gtzan", "license:apache-2.0", "endpoints_compatible", "region:us" ]	audio-classification	2023-08-28T04:50:07Z	--- license: apache-2.0 tags: - generated_from_trainer datasets: - marsyas/gtzan metrics: - accuracy model-index: - name: my_awesome_mind_model results: [] --- <!-- This model card has been generated automatically according to the information the Trainer had access to. You should probably proofread and complete it, then remove this comment. --> # my_awesome_mind_model This model is a fine-tuned version of [facebook/wav2vec2-base](https://huggingface.co/facebook/wav2vec2-base) on the GTZAN dataset. It achieves the following results on the evaluation set: - Loss: 1.7894 - Accuracy: 0.46 ## Model description More information needed ## Intended uses & limitations More information needed ## Training and evaluation data More information needed ## Training procedure ### Training hyperparameters The following hyperparameters were used during training: - learning_rate: 7e-05 - train_batch_size: 32 - eval_batch_size: 32 - seed: 42 - gradient_accumulation_steps: 4 - total_train_batch_size: 128 - optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08 - lr_scheduler_type: linear - lr_scheduler_warmup_ratio: 0.1 - num_epochs: 10 ### Training results \| Training Loss \| Epoch \| Step \| Validation Loss \| Accuracy \| \|:-------------:\|:-----:\|:----:\|:---------------:\|:--------:\| \| No log \| 0.96 \| 6 \| 2.2804 \| 0.22 \| \| 2.2742 \| 1.92 \| 12 \| 2.1887 \| 0.25 \| \| 2.2742 \| 2.88 \| 18 \| 2.1220 \| 0.33 \| \| 2.1487 \| 4.0 \| 25 \| 2.0666 \| 0.355 \| \| 2.0354 \| 4.96 \| 31 \| 1.9963 \| 0.325 \| \| 2.0354 \| 5.92 \| 37 \| 1.9090 \| 0.375 \| \| 1.9246 \| 6.88 \| 43 \| 1.8679 \| 0.405 \| \| 1.8416 \| 8.0 \| 50 \| 1.8185 \| 0.44 \| \| 1.8416 \| 8.96 \| 56 \| 1.7970 \| 0.43 \| \| 1.7858 \| 9.6 \| 60 \| 1.7894 \| 0.46 \| ### Framework versions - Transformers 4.30.2 - Pytorch 2.0.0 - Datasets 2.1.0 - Tokenizers 0.13.3
dkqjrm/20230830102643	dkqjrm	2023-08-30T06:29:45Z	105	0	transformers	[ "transformers", "pytorch", "tensorboard", "bert", "text-classification", "generated_from_trainer", "dataset:super_glue", "license:apache-2.0", "autotrain_compatible", "endpoints_compatible", "region:us" ]	text-classification	2023-08-30T01:27:07Z	--- license: apache-2.0 tags: - generated_from_trainer datasets: - super_glue metrics: - accuracy model-index: - name: '20230830102643' results: [] --- <!-- This model card has been generated automatically according to the information the Trainer had access to. You should probably proofread and complete it, then remove this comment. --> # 20230830102643 This model is a fine-tuned version of [bert-large-cased](https://huggingface.co/bert-large-cased) on the super_glue dataset. It achieves the following results on the evaluation set: - Loss: 0.6732 - Accuracy: 0.5 ## Model description More information needed ## Intended uses & limitations More information needed ## Training and evaluation data More information needed ## Training procedure ### Training hyperparameters The following hyperparameters were used during training: - learning_rate: 0.0007 - train_batch_size: 16 - eval_batch_size: 8 - seed: 11 - optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08 - lr_scheduler_type: linear - num_epochs: 80.0 ### Training results \| Training Loss \| Epoch \| Step \| Validation Loss \| Accuracy \| \|:-------------:\|:-----:\|:-----:\|:---------------:\|:--------:\| \| No log \| 1.0 \| 340 \| 0.6910 \| 0.5 \| \| 0.7049 \| 2.0 \| 680 \| 0.6698 \| 0.5078 \| \| 0.6976 \| 3.0 \| 1020 \| 0.6913 \| 0.5 \| \| 0.6976 \| 4.0 \| 1360 \| 0.7490 \| 0.5 \| \| 0.6898 \| 5.0 \| 1700 \| 0.8592 \| 0.5078 \| \| 0.6423 \| 6.0 \| 2040 \| 0.7080 \| 0.5987 \| \| 0.6423 \| 7.0 \| 2380 \| 0.6940 \| 0.5 \| \| 0.6463 \| 8.0 \| 2720 \| 0.6703 \| 0.5 \| \| 0.7054 \| 9.0 \| 3060 \| 0.6741 \| 0.5047 \| \| 0.7054 \| 10.0 \| 3400 \| 0.6784 \| 0.5 \| \| 0.7011 \| 11.0 \| 3740 \| 0.6707 \| 0.5 \| \| 0.6895 \| 12.0 \| 4080 \| 0.6941 \| 0.5 \| \| 0.6895 \| 13.0 \| 4420 \| 0.7550 \| 0.5 \| \| 0.6898 \| 14.0 \| 4760 \| 0.7095 \| 0.5 \| \| 0.6916 \| 15.0 \| 5100 \| 0.6711 \| 0.5 \| \| 0.6916 \| 16.0 \| 5440 \| 0.7051 \| 0.5 \| \| 0.6891 \| 17.0 \| 5780 \| 0.6831 \| 0.5 \| \| 0.6875 \| 18.0 \| 6120 \| 0.6733 \| 0.5172 \| \| 0.6875 \| 19.0 \| 6460 \| 0.6707 \| 0.5 \| \| 0.6854 \| 20.0 \| 6800 \| 0.6851 \| 0.5 \| \| 0.686 \| 21.0 \| 7140 \| 0.6704 \| 0.5 \| \| 0.686 \| 22.0 \| 7480 \| 0.6938 \| 0.5 \| \| 0.6874 \| 23.0 \| 7820 \| 0.6848 \| 0.5 \| \| 0.6848 \| 24.0 \| 8160 \| 0.6710 \| 0.5 \| \| 0.6811 \| 25.0 \| 8500 \| 0.6783 \| 0.4953 \| \| 0.6811 \| 26.0 \| 8840 \| 0.6886 \| 0.5 \| \| 0.6837 \| 27.0 \| 9180 \| 0.7146 \| 0.5 \| \| 0.6851 \| 28.0 \| 9520 \| 0.6703 \| 0.5 \| \| 0.6851 \| 29.0 \| 9860 \| 0.6884 \| 0.5 \| \| 0.6813 \| 30.0 \| 10200 \| 0.6704 \| 0.5 \| \| 0.6826 \| 31.0 \| 10540 \| 0.6704 \| 0.5 \| \| 0.6826 \| 32.0 \| 10880 \| 0.6970 \| 0.5 \| \| 0.6813 \| 33.0 \| 11220 \| 0.6707 \| 0.5 \| \| 0.6786 \| 34.0 \| 11560 \| 0.6840 \| 0.5 \| \| 0.6786 \| 35.0 \| 11900 \| 0.6722 \| 0.5 \| \| 0.6821 \| 36.0 \| 12240 \| 0.6706 \| 0.5 \| \| 0.6799 \| 37.0 \| 12580 \| 0.6707 \| 0.5 \| \| 0.6799 \| 38.0 \| 12920 \| 0.6824 \| 0.4953 \| \| 0.6803 \| 39.0 \| 13260 \| 0.6995 \| 0.5 \| \| 0.6775 \| 40.0 \| 13600 \| 0.6728 \| 0.5 \| \| 0.6775 \| 41.0 \| 13940 \| 0.6711 \| 0.4984 \| \| 0.679 \| 42.0 \| 14280 \| 0.6743 \| 0.5 \| \| 0.6775 \| 43.0 \| 14620 \| 0.6742 \| 0.5016 \| \| 0.6775 \| 44.0 \| 14960 \| 0.6724 \| 0.5016 \| \| 0.6766 \| 45.0 \| 15300 \| 0.6884 \| 0.5 \| \| 0.6755 \| 46.0 \| 15640 \| 0.6709 \| 0.5 \| \| 0.6755 \| 47.0 \| 15980 \| 0.6811 \| 0.5 \| \| 0.6785 \| 48.0 \| 16320 \| 0.6709 \| 0.5 \| \| 0.6765 \| 49.0 \| 16660 \| 0.6813 \| 0.5 \| \| 0.6761 \| 50.0 \| 17000 \| 0.6724 \| 0.5 \| \| 0.6761 \| 51.0 \| 17340 \| 0.6713 \| 0.5 \| \| 0.6764 \| 52.0 \| 17680 \| 0.6715 \| 0.5016 \| \| 0.6774 \| 53.0 \| 18020 \| 0.6730 \| 0.5 \| \| 0.6774 \| 54.0 \| 18360 \| 0.6730 \| 0.5 \| \| 0.673 \| 55.0 \| 18700 \| 0.6716 \| 0.5016 \| \| 0.6718 \| 56.0 \| 19040 \| 0.6714 \| 0.5 \| \| 0.6718 \| 57.0 \| 19380 \| 0.6714 \| 0.4984 \| \| 0.6745 \| 58.0 \| 19720 \| 0.6715 \| 0.5016 \| \| 0.6735 \| 59.0 \| 20060 \| 0.6863 \| 0.5 \| \| 0.6735 \| 60.0 \| 20400 \| 0.6710 \| 0.4984 \| \| 0.6725 \| 61.0 \| 20740 \| 0.6718 \| 0.5063 \| \| 0.6734 \| 62.0 \| 21080 \| 0.6717 \| 0.5 \| \| 0.6734 \| 63.0 \| 21420 \| 0.6714 \| 0.4984 \| \| 0.6725 \| 64.0 \| 21760 \| 0.6749 \| 0.5 \| \| 0.6719 \| 65.0 \| 22100 \| 0.6723 \| 0.5 \| \| 0.6719 \| 66.0 \| 22440 \| 0.6776 \| 0.5 \| \| 0.6714 \| 67.0 \| 22780 \| 0.6716 \| 0.5 \| \| 0.6724 \| 68.0 \| 23120 \| 0.6723 \| 0.5 \| \| 0.6724 \| 69.0 \| 23460 \| 0.6717 \| 0.5 \| \| 0.6705 \| 70.0 \| 23800 \| 0.6712 \| 0.4984 \| \| 0.6722 \| 71.0 \| 24140 \| 0.6725 \| 0.5 \| \| 0.6722 \| 72.0 \| 24480 \| 0.6775 \| 0.5 \| \| 0.6715 \| 73.0 \| 24820 \| 0.6744 \| 0.5 \| \| 0.6708 \| 74.0 \| 25160 \| 0.6762 \| 0.5 \| \| 0.6705 \| 75.0 \| 25500 \| 0.6737 \| 0.5 \| \| 0.6705 \| 76.0 \| 25840 \| 0.6745 \| 0.5 \| \| 0.6698 \| 77.0 \| 26180 \| 0.6717 \| 0.5 \| \| 0.6691 \| 78.0 \| 26520 \| 0.6735 \| 0.5 \| \| 0.6691 \| 79.0 \| 26860 \| 0.6732 \| 0.5 \| \| 0.6693 \| 80.0 \| 27200 \| 0.6732 \| 0.5 \| ### Framework versions - Transformers 4.26.1 - Pytorch 2.0.1+cu118 - Datasets 2.12.0 - Tokenizers 0.13.3
dkqjrm/20230830102630	dkqjrm	2023-08-30T06:17:08Z	114	0	transformers	[ "transformers", "pytorch", "tensorboard", "bert", "text-classification", "generated_from_trainer", "dataset:super_glue", "license:apache-2.0", "autotrain_compatible", "endpoints_compatible", "region:us" ]	text-classification	2023-08-30T01:26:53Z	--- license: apache-2.0 tags: - generated_from_trainer datasets: - super_glue metrics: - accuracy model-index: - name: '20230830102630' results: [] --- <!-- This model card has been generated automatically according to the information the Trainer had access to. You should probably proofread and complete it, then remove this comment. --> # 20230830102630 This model is a fine-tuned version of [bert-large-cased](https://huggingface.co/bert-large-cased) on the super_glue dataset. It achieves the following results on the evaluation set: - Loss: 0.6751 - Accuracy: 0.4984 ## Model description More information needed ## Intended uses & limitations More information needed ## Training and evaluation data More information needed ## Training procedure ### Training hyperparameters The following hyperparameters were used during training: - learning_rate: 0.0005 - train_batch_size: 16 - eval_batch_size: 8 - seed: 11 - optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08 - lr_scheduler_type: linear - num_epochs: 80.0 ### Training results \| Training Loss \| Epoch \| Step \| Validation Loss \| Accuracy \| \|:-------------:\|:-----:\|:-----:\|:---------------:\|:--------:\| \| No log \| 1.0 \| 340 \| 0.6962 \| 0.5 \| \| 0.7003 \| 2.0 \| 680 \| 0.6657 \| 0.5345 \| \| 0.6959 \| 3.0 \| 1020 \| 0.6703 \| 0.5 \| \| 0.6959 \| 4.0 \| 1360 \| 0.6845 \| 0.5 \| \| 0.6978 \| 5.0 \| 1700 \| 0.6767 \| 0.5 \| \| 0.6854 \| 6.0 \| 2040 \| 0.6876 \| 0.5 \| \| 0.6854 \| 7.0 \| 2380 \| 0.6705 \| 0.5 \| \| 0.6851 \| 8.0 \| 2720 \| 0.6806 \| 0.5 \| \| 0.6845 \| 9.0 \| 3060 \| 0.6881 \| 0.5 \| \| 0.6845 \| 10.0 \| 3400 \| 0.6737 \| 0.5 \| \| 0.6835 \| 11.0 \| 3740 \| 0.6734 \| 0.5 \| \| 0.6821 \| 12.0 \| 4080 \| 0.7058 \| 0.5 \| \| 0.6821 \| 13.0 \| 4420 \| 0.7057 \| 0.5 \| \| 0.682 \| 14.0 \| 4760 \| 0.7057 \| 0.5 \| \| 0.6827 \| 15.0 \| 5100 \| 0.6771 \| 0.5 \| \| 0.6827 \| 16.0 \| 5440 \| 0.6848 \| 0.5 \| \| 0.6803 \| 17.0 \| 5780 \| 0.7044 \| 0.5 \| \| 0.6821 \| 18.0 \| 6120 \| 0.6720 \| 0.4984 \| \| 0.6821 \| 19.0 \| 6460 \| 0.6716 \| 0.5 \| \| 0.6784 \| 20.0 \| 6800 \| 0.6855 \| 0.5 \| \| 0.6821 \| 21.0 \| 7140 \| 0.6705 \| 0.5 \| \| 0.6821 \| 22.0 \| 7480 \| 0.6753 \| 0.5 \| \| 0.6888 \| 23.0 \| 7820 \| 0.6745 \| 0.4953 \| \| 0.6821 \| 24.0 \| 8160 \| 0.6716 \| 0.5 \| \| 0.682 \| 25.0 \| 8500 \| 0.6702 \| 0.5 \| \| 0.682 \| 26.0 \| 8840 \| 0.6791 \| 0.5 \| \| 0.6829 \| 27.0 \| 9180 \| 0.6771 \| 0.5 \| \| 0.6807 \| 28.0 \| 9520 \| 0.6719 \| 0.5 \| \| 0.6807 \| 29.0 \| 9860 \| 0.6739 \| 0.5 \| \| 0.6783 \| 30.0 \| 10200 \| 0.6716 \| 0.5 \| \| 0.6789 \| 31.0 \| 10540 \| 0.6706 \| 0.5 \| \| 0.6789 \| 32.0 \| 10880 \| 0.7163 \| 0.5 \| \| 0.6798 \| 33.0 \| 11220 \| 0.6703 \| 0.5 \| \| 0.6785 \| 34.0 \| 11560 \| 0.6822 \| 0.5 \| \| 0.6785 \| 35.0 \| 11900 \| 0.6715 \| 0.5 \| \| 0.6783 \| 36.0 \| 12240 \| 0.6720 \| 0.5 \| \| 0.6781 \| 37.0 \| 12580 \| 0.6733 \| 0.5 \| \| 0.6781 \| 38.0 \| 12920 \| 0.6707 \| 0.5 \| \| 0.6798 \| 39.0 \| 13260 \| 0.6950 \| 0.5 \| \| 0.6755 \| 40.0 \| 13600 \| 0.6705 \| 0.5 \| \| 0.6755 \| 41.0 \| 13940 \| 0.6715 \| 0.5 \| \| 0.6776 \| 42.0 \| 14280 \| 0.6704 \| 0.5 \| \| 0.6772 \| 43.0 \| 14620 \| 0.6789 \| 0.5 \| \| 0.6772 \| 44.0 \| 14960 \| 0.6707 \| 0.5 \| \| 0.6755 \| 45.0 \| 15300 \| 0.6925 \| 0.5 \| \| 0.6748 \| 46.0 \| 15640 \| 0.6727 \| 0.5 \| \| 0.6748 \| 47.0 \| 15980 \| 0.6801 \| 0.5 \| \| 0.6754 \| 48.0 \| 16320 \| 0.6714 \| 0.5 \| \| 0.6762 \| 49.0 \| 16660 \| 0.6882 \| 0.5 \| \| 0.6753 \| 50.0 \| 17000 \| 0.6710 \| 0.5 \| \| 0.6753 \| 51.0 \| 17340 \| 0.6707 \| 0.5 \| \| 0.6734 \| 52.0 \| 17680 \| 0.6726 \| 0.5063 \| \| 0.678 \| 53.0 \| 18020 \| 0.6727 \| 0.5 \| \| 0.678 \| 54.0 \| 18360 \| 0.6751 \| 0.5 \| \| 0.6719 \| 55.0 \| 18700 \| 0.6712 \| 0.5 \| \| 0.6726 \| 56.0 \| 19040 \| 0.6721 \| 0.5 \| \| 0.6726 \| 57.0 \| 19380 \| 0.6715 \| 0.5 \| \| 0.6732 \| 58.0 \| 19720 \| 0.6717 \| 0.5016 \| \| 0.6736 \| 59.0 \| 20060 \| 0.6819 \| 0.5 \| \| 0.6736 \| 60.0 \| 20400 \| 0.6728 \| 0.5141 \| \| 0.6732 \| 61.0 \| 20740 \| 0.6716 \| 0.5016 \| \| 0.6727 \| 62.0 \| 21080 \| 0.6747 \| 0.5 \| \| 0.6727 \| 63.0 \| 21420 \| 0.6715 \| 0.4984 \| \| 0.6726 \| 64.0 \| 21760 \| 0.6737 \| 0.5 \| \| 0.6721 \| 65.0 \| 22100 \| 0.6724 \| 0.5 \| \| 0.6721 \| 66.0 \| 22440 \| 0.6744 \| 0.5 \| \| 0.6711 \| 67.0 \| 22780 \| 0.6720 \| 0.5 \| \| 0.6725 \| 68.0 \| 23120 \| 0.6722 \| 0.4984 \| \| 0.6725 \| 69.0 \| 23460 \| 0.6722 \| 0.4984 \| \| 0.6713 \| 70.0 \| 23800 \| 0.6722 \| 0.4984 \| \| 0.6708 \| 71.0 \| 24140 \| 0.6743 \| 0.5 \| \| 0.6708 \| 72.0 \| 24480 \| 0.6794 \| 0.5 \| \| 0.6703 \| 73.0 \| 24820 \| 0.6756 \| 0.5 \| \| 0.6702 \| 74.0 \| 25160 \| 0.6760 \| 0.5 \| \| 0.6688 \| 75.0 \| 25500 \| 0.6741 \| 0.4984 \| \| 0.6688 \| 76.0 \| 25840 \| 0.6753 \| 0.5 \| \| 0.67 \| 77.0 \| 26180 \| 0.6730 \| 0.4984 \| \| 0.6688 \| 78.0 \| 26520 \| 0.6751 \| 0.4984 \| \| 0.6688 \| 79.0 \| 26860 \| 0.6750 \| 0.4984 \| \| 0.6685 \| 80.0 \| 27200 \| 0.6751 \| 0.4984 \| ### Framework versions - Transformers 4.26.1 - Pytorch 2.0.1+cu118 - Datasets 2.12.0 - Tokenizers 0.13.3
fuguizhen/dqn-SpaceInvadersNoFrameskip-v4	fuguizhen	2023-08-30T06:16:35Z	0	0	stable-baselines3	[ "stable-baselines3", "SpaceInvadersNoFrameskip-v4", "deep-reinforcement-learning", "reinforcement-learning", "model-index", "region:us" ]	reinforcement-learning	2023-08-30T06:16:12Z	--- library_name: stable-baselines3 tags: - SpaceInvadersNoFrameskip-v4 - deep-reinforcement-learning - reinforcement-learning - stable-baselines3 model-index: - name: DQN results: - task: type: reinforcement-learning name: reinforcement-learning dataset: name: SpaceInvadersNoFrameskip-v4 type: SpaceInvadersNoFrameskip-v4 metrics: - type: mean_reward value: 6.50 +/- 10.74 name: mean_reward verified: false --- # DQN Agent playing SpaceInvadersNoFrameskip-v4 This is a trained model of a DQN agent playing SpaceInvadersNoFrameskip-v4 using the [stable-baselines3 library](https://github.com/DLR-RM/stable-baselines3) and the [RL Zoo](https://github.com/DLR-RM/rl-baselines3-zoo). The RL Zoo is a training framework for Stable Baselines3 reinforcement learning agents, with hyperparameter optimization and pre-trained agents included. ## Usage (with SB3 RL Zoo) RL Zoo: https://github.com/DLR-RM/rl-baselines3-zoo<br/> SB3: https://github.com/DLR-RM/stable-baselines3<br/> SB3 Contrib: https://github.com/Stable-Baselines-Team/stable-baselines3-contrib Install the RL Zoo (with SB3 and SB3-Contrib): ```bash pip install rl_zoo3 ``` ``` # Download model and save it into the logs/ folder python -m rl_zoo3.load_from_hub --algo dqn --env SpaceInvadersNoFrameskip-v4 -orga fuguizhen -f logs/ python -m rl_zoo3.enjoy --algo dqn --env SpaceInvadersNoFrameskip-v4 -f logs/ ``` If you installed the RL Zoo3 via pip (`pip install rl_zoo3`), from anywhere you can do: ``` python -m rl_zoo3.load_from_hub --algo dqn --env SpaceInvadersNoFrameskip-v4 -orga fuguizhen -f logs/ python -m rl_zoo3.enjoy --algo dqn --env SpaceInvadersNoFrameskip-v4 -f logs/ ``` ## Training (with the RL Zoo) ``` python -m rl_zoo3.train --algo dqn --env SpaceInvadersNoFrameskip-v4 -f logs/ # Upload the model and generate video (when possible) python -m rl_zoo3.push_to_hub --algo dqn --env SpaceInvadersNoFrameskip-v4 -f logs/ -orga fuguizhen ``` ## Hyperparameters ```python OrderedDict([('batch_size', 32), ('buffer_size', 100000), ('env_wrapper', ['stable_baselines3.common.atari_wrappers.AtariWrapper']), ('exploration_final_eps', 0.01), ('exploration_fraction', 0.1), ('frame_stack', 4), ('gradient_steps', 1), ('learning_rate', 0.0001), ('learning_starts', 100000), ('n_timesteps', 100000.0), ('optimize_memory_usage', False), ('policy', 'CnnPolicy'), ('target_update_interval', 1000), ('train_freq', 4), ('normalize', False)]) ``` # Environment Arguments ```python {'render_mode': 'rgb_array'} ```
hellomyoh/nmt-s12000-kullm-polyglot-5.8b-v1	hellomyoh	2023-08-30T06:15:48Z	8	0	transformers	[ "transformers", "pytorch", "gpt_neox", "text-generation", "autotrain_compatible", "text-generation-inference", "endpoints_compatible", "region:us" ]	text-generation	2023-08-30T03:41:44Z	\|no\|english\|korean\| --\|--\|--\| \|1 \| Do you know who i am?\| '제가 누군지 아시겠어요?'\| \|2 \|Tired of scrolling through the same posts? When you create an account you’ll always come back to where you left off. With an account you can also be notified of new replies, save bookmarks, and use likes to thank others. We can all work together to make this community great.. \| 똑같은 글을 스크롤하는 것에 지쳤나요? 계정을 만들면 항상 처음으로 돌아오게 됩니다. 계정을 만들면 새로운 댓글을 확인할 수 있고 북마크를 저장할 수 있으며 다른 사람들에게 감사할 수 있습니다. 우리는 함께 이 커뮤니티를 멋지게 만들어 나갈 수 있습니다. \| \|3 \| As technology continues to advance, new vulnerabilities emerge and the importance of security becomes increasingly crucial. In this regard, Cho Hong Ki, the Information Security Specialist at the 2bytes, shared valuable knowledge on the significance of game security and the solutions it offers.\|기술이 계속 발전함에 따라 새로운 취약점이 발생하고 보안의 중요성이 더욱 커지고 있습니다. 이와 관련해 2bytes의 정보보안 전문가인 조홍기 씨는 게임 보안의 중요성과 그 해결책에 대한 귀중한 지식을 공유했습니다.\| \|4 \|They are <i>gifts</i> to my Queen from the goddess Tiamat herself. They reside in the great city of Tu'narath, awaiting the privilege of battle \|</i> 신들의 여왕인 티아맛의 선물입니다. 이들은 투나랏의 거대한 도시에 살고 있으며, 전투의 영광을 누리고 있습니다.\| \|5\|I showed my masterpiece to the grown-ups, and asked them whether the drawing frightened them. But they answered: 'Frighten? Why should any one be frightened by a hat?' \| "나는 어른들에게 내 작품을 보여주었고, 그들에게 모자가 무섭지 않으냐고 물었다. 하지만 그들은 대답했다. '무섭다고? 왜 누군가가 무서워하지? 모자가 왜 무서워?"\| \|6 \|New York. CNN. The Federal Aviation Administration has certified for testing a vehicle that a California startup describes as a flying car — the first fully electric vehicle that can both fly and travel on roads to receive US government approval. Alef Automotive said that its vehicle/aircraft, dubbed the “Model A,” is the first flying vehicle that is drivable on public roads and able to park like a normal car. It also has vertical takeoff and landing capabilities. It apparently will be able to carry one or two occupants and will have a road-range of 200 miles and a flying range of 110 miles. The company expects to sell the vehicle for $300,000 each with the first delivery by projected for the end of 2025. The FAA confirmed that it has issued the company a special airworthiness certificate, allowing for limited purposes that include exhibition, research and development. Numerous companies are working on all-electric VTOLs, which stands for vehicle takeoff and landing aircraft. The FAA said that Alef is “not the first aircraft of its kind” to get a special airworthiness certificate. However, Alef noted that its vehicle is different because of its ability to function both on roads and in the air, to appear like a normal car and to park in a normal parking space. \| Alef Automotive는 캘리포니아의 한 신생 기업이 묘사한 것처럼 비행할 수 있는 최초의 완전 전기 자동차를 테스트하기 위해 뉴욕의 CNN과 연방 항공국에 의해 인증을 받았다고 말했습니다. Alef Automotive는 이 차량을 "모델 A"라고 부르며 도로에서도 날 수 있고 일반 자동차처럼 주차할 수 있는 최초의 비행 가능한 차량이라고 말했습니다. 수직 이착륙과 도로 주행이 가능하며 한 두 명 정도의 승객을 태울 수 있고 주행 거리는 200마일과 110마일의 비행이 가능할 것으로 예상하고 있습니다. FAA는 이 회사가 2025년 말까지 첫 번째 배송을 완료할 것으로 예상되는 가운데 이 차량을 3억 달러에 판매할 수 있을 것으로 기대하고 있다고 밝혔습니다. FAA는 이 회사에 특별한 인증을 발급하여 전시, 연구 및 개발을 위한 제한된 목적에 한해서만 사용할 수 있도록 했습니다. 많은 회사들이 모든 전기식 비행체를 개발하고 있으며 이는 비행기에 착륙하고 이륙하는 차량을 의미합니다. FAA는 Alef가 도로에서나 공중에서나 정상적인 차량처럼 보이고 일반적인 주차 공간에 주차할 수 있는 차량이라는 점에서 다른 차량과 다르다고 말했습니다.\|
trieudemo11/llama_7b_attrb_cate_batch8_len320_gpu8_10	trieudemo11	2023-08-30T06:13:44Z	1	0	peft	[ "peft", "region:us" ]	null	2023-08-30T06:13:28Z	--- library_name: peft --- ## Training procedure ### Framework versions - PEFT 0.6.0.dev0 - PEFT 0.6.0.dev0 - PEFT 0.6.0.dev0 - PEFT 0.6.0.dev0 - PEFT 0.6.0.dev0 - PEFT 0.6.0.dev0 - PEFT 0.6.0.dev0 - PEFT 0.6.0.dev0 - PEFT 0.6.0.dev0 - PEFT 0.6.0.dev0 - PEFT 0.6.0.dev0 - PEFT 0.6.0.dev0 - PEFT 0.6.0.dev0 - PEFT 0.6.0.dev0 - PEFT 0.6.0.dev0 - PEFT 0.6.0.dev0 - PEFT 0.6.0.dev0 - PEFT 0.6.0.dev0 - PEFT 0.6.0.dev0 - PEFT 0.6.0.dev0 - PEFT 0.6.0.dev0 - PEFT 0.6.0.dev0 - PEFT 0.6.0.dev0 - PEFT 0.6.0.dev0 - PEFT 0.6.0.dev0 - PEFT 0.6.0.dev0 - PEFT 0.6.0.dev0 - PEFT 0.6.0.dev0 - PEFT 0.6.0.dev0 - PEFT 0.6.0.dev0 - PEFT 0.6.0.dev0 - PEFT 0.6.0.dev0 - PEFT 0.6.0.dev0 - PEFT 0.6.0.dev0 - PEFT 0.6.0.dev0 - PEFT 0.6.0.dev0 - PEFT 0.6.0.dev0 - PEFT 0.6.0.dev0 - PEFT 0.6.0.dev0 - PEFT 0.6.0.dev0 - PEFT 0.6.0.dev0 - PEFT 0.6.0.dev0 - PEFT 0.6.0.dev0 - PEFT 0.6.0.dev0 - PEFT 0.6.0.dev0 - PEFT 0.6.0.dev0 - PEFT 0.6.0.dev0 - PEFT 0.6.0.dev0 - PEFT 0.6.0.dev0 - PEFT 0.6.0.dev0 - PEFT 0.6.0.dev0 - PEFT 0.6.0.dev0 - PEFT 0.6.0.dev0 - PEFT 0.6.0.dev0 - PEFT 0.6.0.dev0 - PEFT 0.6.0.dev0 - PEFT 0.6.0.dev0 - PEFT 0.6.0.dev0 - PEFT 0.6.0.dev0 - PEFT 0.6.0.dev0 - PEFT 0.6.0.dev0 - PEFT 0.6.0.dev0 - PEFT 0.6.0.dev0 - PEFT 0.6.0.dev0 - PEFT 0.6.0.dev0 - PEFT 0.6.0.dev0 - PEFT 0.6.0.dev0 - PEFT 0.6.0.dev0 - PEFT 0.6.0.dev0
LarryAIDraw/yukina_kiritani_imocho	LarryAIDraw	2023-08-30T05:56:58Z	0	0	null	[ "license:creativeml-openrail-m", "region:us" ]	null	2023-08-30T05:41:01Z	--- license: creativeml-openrail-m --- https://civitai.com/models/137304/yukina-kiritani-recently-my-sister-is-unusual
LarryAIDraw/Katori-10	LarryAIDraw	2023-08-30T05:54:54Z	0	0	null	[ "license:creativeml-openrail-m", "region:us" ]	null	2023-08-30T05:42:27Z	--- license: creativeml-openrail-m --- https://civitai.com/models/135972/katori-batsuunsai-bleach
LarryAIDraw/August	LarryAIDraw	2023-08-30T05:53:12Z	0	0	null	[ "license:creativeml-openrail-m", "region:us" ]	null	2023-08-30T05:36:19Z	--- license: creativeml-openrail-m --- https://civitai.com/models/137145/augustvonparseval-azur-lane
LarryAIDraw/Modernia-FirstAffection-LoRA-v1	LarryAIDraw	2023-08-30T05:50:06Z	0	0	null	[ "license:creativeml-openrail-m", "region:us" ]	null	2023-08-30T05:34:44Z	--- license: creativeml-openrail-m --- https://civitai.com/models/136485/modernia-first-affection-nikke-goddess-of-victory
LarryAIDraw/InoueOrihime-2k	LarryAIDraw	2023-08-30T05:49:25Z	0	0	null	[ "license:creativeml-openrail-m", "region:us" ]	null	2023-08-30T05:33:36Z	--- license: creativeml-openrail-m --- https://civitai.com/models/136699/inoue-orihime-from-bleach
LarryAIDraw/djeeta_20230722	LarryAIDraw	2023-08-30T05:40:07Z	0	0	null	[ "license:creativeml-openrail-m", "region:us" ]	null	2023-08-30T05:40:07Z	--- license: creativeml-openrail-m ---
BM-K/NewsKoT5-small	BM-K	2023-08-30T05:37:14Z	109	0	transformers	[ "transformers", "pytorch", "safetensors", "t5", "text2text-generation", "ko", "autotrain_compatible", "text-generation-inference", "endpoints_compatible", "region:us" ]	text2text-generation	2023-06-01T07:36:27Z	--- language: - ko --- # NewsKoT5 The training data for this T5 model consists of Korean news articles (29GB). However, the performance has not been fine-tuned through the use of small batches and a limited number of training steps, so it may not be fully optimized. ## Quick tour ```python from transformers import AutoTokenizer, T5ForConditionalGeneration tokenizer = AutoTokenizer.from_pretrained("BM-K/NewsKoT5-small") model = T5ForConditionalGeneration.from_pretrained("BM-K/NewsKoT5-small") input_ids = tokenizer("한국형발사체 누리호가 실용급 <extra_id_0> 발사체로서 ‘데뷔’를 성공적으로 <extra_id_1>", return_tensors="pt").input_ids labels = tokenizer("<extra_id_0> 위성 <extra_id_1> 마쳤다 <extra_id_2>", return_tensors="pt").input_ids outputs = model(input_ids=input_ids, labels=labels) ``` ## News Summarization Performance (F1-score) After restoring the model's tokenized output to the original text, Rouge performance was evaluated by comparing it to the reference and hypothesis tokenized using [mecab](https://konlpy.org/ko/v0.4.0/). - Dacon 한국어 문서 생성요약 AI 경진대회 [Dataset](https://dacon.io/competitions/official/235673/overview/description) - Training: 29,432 - Validation: 7,358 - Test: 9,182 \| \| #Param \| rouge-1 \|rouge-2\|rouge-l\| \|-------\|--------:\|--------:\|--------:\|--------:\| \| pko-t5-small \| 95M \| 51.48 \| 33.18 \| 44.96 \| \| NewsT5-small \| 61M \| 52.15 \| 33.59 \| 45.41 \| - AI-Hub 문서요약 텍스트 [Dataset](https://www.aihub.or.kr/aihubdata/data/view.do?currMenu=115&topMenu=100&aihubDataSe=realm&dataSetSn=97) - Training: 245,626 - Validation: 20,296 - Test: 9,931 \| \| #Param \| rouge-1 \|rouge-2\|rouge-l\| \|-------\|--------:\|--------:\|--------:\|--------:\| \| pko-t5-small \| 95M \| 53.44 \| 34.03 \| 45.36 \| \| NewsT5-small \| 61M \| 53.74 \| 34.27 \| 45.52 \| - [pko-t5-small](https://github.com/paust-team/pko-t5)
BM-K/KoMiniLM	BM-K	2023-08-30T05:37:05Z	134	2	transformers	[ "transformers", "pytorch", "safetensors", "bert", "feature-extraction", "arxiv:2002.10957", "text-embeddings-inference", "endpoints_compatible", "region:us" ]	feature-extraction	2022-05-23T04:26:31Z	# KoMiniLM 🐣 Korean mini language model ## Overview Current language models usually consist of hundreds of millions of parameters which brings challenges for fine-tuning and online serving in real-life applications due to latency and capacity constraints. In this project, we release a light weight korean language model to address the aforementioned shortcomings of existing language models. ## Quick tour ```python from transformers import AutoTokenizer, AutoModel tokenizer = AutoTokenizer.from_pretrained("BM-K/KoMiniLM") # 23M model model = AutoModel.from_pretrained("BM-K/KoMiniLM") inputs = tokenizer("안녕 세상아!", return_tensors="pt") outputs = model(inputs) ``` ## Update history Updates on 2022.06.20 - Release KoMiniLM-bert-68M Updates on 2022.05.24 - Release KoMiniLM-bert-23M ## Pre-training `Teacher Model`: [KLUE-BERT(base)](https://github.com/KLUE-benchmark/KLUE) ### Object Self-Attention Distribution and Self-Attention Value-Relation [[Wang et al., 2020]](https://arxiv.org/abs/2002.10957) were distilled from each discrete layer of the teacher model to the student model. Wang et al. distilled in the last layer of the transformer, but that was not the case in this project. ### Data sets \|Data\|News comments\|News article\| \|:----:\|:----:\|:----:\| \|size\|10G\|10G\| ### Config - KoMiniLM-23M ```json { "architectures": [ "BertForPreTraining" ], "attention_probs_dropout_prob": 0.1, "classifier_dropout": null, "hidden_act": "gelu", "hidden_dropout_prob": 0.1, "hidden_size": 384, "initializer_range": 0.02, "intermediate_size": 1536, "layer_norm_eps": 1e-12, "max_position_embeddings": 512, "model_type": "bert", "num_attention_heads": 12, "num_hidden_layers": 6, "output_attentions": true, "pad_token_id": 0, "position_embedding_type": "absolute", "return_dict": false, "torch_dtype": "float32", "transformers_version": "4.13.0", "type_vocab_size": 2, "use_cache": true, "vocab_size": 32000 } ``` ### Performance on subtasks - The results of our fine-tuning experiments are an average of 3 runs for each task. ``` cd KoMiniLM-Finetune bash scripts/run_all_kominilm.sh ``` \|\| #Param \| Average \| NSMC<br>(Acc) \| Naver NER<br>(F1) \| PAWS<br>(Acc) \| KorNLI<br>(Acc) \| KorSTS<br>(Spearman) \| Question Pair<br>(Acc) \| KorQuaD<br>(Dev)<br>(EM/F1) \| \|:----:\|:----:\|:----:\|:----:\|:----:\|:----:\|:----:\|:----:\|:----:\|:----:\| \|KoBERT(KLUE)\| 110M \| 86.84 \| 90.20±0.07 \| 87.11±0.05 \| 81.36±0.21 \| 81.06±0.33 \| 82.47±0.14 \| 95.03±0.44 \| 84.43±0.18 / <br>93.05±0.04 \| \|KcBERT\| 108M \| 78.94 \| 89.60±0.10 \| 84.34±0.13 \| 67.02±0.42\| 74.17±0.52 \| 76.57±0.51 \| 93.97±0.27 \| 60.87±0.27 / <br>85.01±0.14 \| \|KoBERT(SKT)\| 92M \| 79.73 \| 89.28±0.42 \| 87.54±0.04 \| 80.93±0.91 \| 78.18±0.45 \| 75.98±2.81 \| 94.37±0.31 \| 51.94±0.60 / <br>79.69±0.66 \| \|DistilKoBERT\| 28M \| 74.73 \| 88.39±0.08 \| 84.22±0.01 \| 61.74±0.45 \| 70.22±0.14 \| 72.11±0.27 \| 92.65±0.16 \| 52.52±0.48 / <br>76.00±0.71 \| \| \| \| \| \| \| \| \| \| \| \|KoMiniLM<sup>†</sup>\| 68M \| 85.90 \| 89.84±0.02 \| 85.98±0.09 \| 80.78±0.30 \| 79.28±0.17 \| 81.00±0.07 \| 94.89±0.37 \| 83.27±0.08 / <br>92.08±0.06 \| \|KoMiniLM<sup>†</sup>\| 23M** \| 84.79 \| 89.67±0.03 \| 84.79±0.09 \| 78.67±0.45 \| 78.10±0.07 \| 78.90±0.11 \| 94.81±0.12 \| 82.11±0.42 / <br>91.21±0.29 \| - [NSMC](https://github.com/e9t/nsmc) (Naver Sentiment Movie Corpus) - [Naver NER](https://github.com/naver/nlp-challenge) (NER task on Naver NLP Challenge 2018) - [PAWS](https://github.com/google-research-datasets/paws) (Korean Paraphrase Adversaries from Word Scrambling) - [KorNLI/KorSTS](https://github.com/kakaobrain/KorNLUDatasets) (Korean Natural Language Understanding) - [Question Pair](https://github.com/songys/Question_pair) (Paired Question) - [KorQuAD](https://korquad.github.io/) (The Korean Question Answering Dataset) <img src = "https://user-images.githubusercontent.com/55969260/174229747-279122dc-9d27-4da9-a6e7-f9f1fe1651f7.png"> <br> ### User Contributed Examples - ## Reference - [KLUE BERT](https://github.com/KLUE-benchmark/KLUE) - [KcBERT](https://github.com/Beomi/KcBERT) - [SKT KoBERT](https://github.com/SKTBrain/KoBERT) - [DistilKoBERT](https://github.com/monologg/DistilKoBERT) - [lassl](https://github.com/lassl/lassl)
BM-K/KoDiffCSE-RoBERTa	BM-K	2023-08-30T05:36:55Z	27,758	4	transformers	[ "transformers", "pytorch", "safetensors", "roberta", "feature-extraction", "arxiv:2204.10298", "arxiv:2004.03289", "arxiv:2105.09680", "text-embeddings-inference", "endpoints_compatible", "region:us" ]	feature-extraction	2023-02-28T04:17:36Z	# KoDiffCSE Difference-based Contrastive Learning for Korean Sentence Embeddings <br> - [DiffCSE-[NAACL 2022]](https://arxiv.org/abs/2204.10298) <br> - [[Github]](https://github.com/voidism/DiffCSE) Official implementation of DiffCSE <br> <img src=https://user-images.githubusercontent.com/55969260/201829550-9674a3ac-cb9b-4e17-b777-7d96fdf5c633.png> ## Quick tour ```python import torch from transformers import AutoModel, AutoTokenizer def cal_score(a, b): if len(a.shape) == 1: a = a.unsqueeze(0) if len(b.shape) == 1: b = b.unsqueeze(0) a_norm = a / a.norm(dim=1)[:, None] b_norm = b / b.norm(dim=1)[:, None] return torch.mm(a_norm, b_norm.transpose(0, 1)) * 100 model = AutoModel.from_pretrained('BM-K/KoDiffCSE-RoBERTa') tokenizer = AutoTokenizer.from_pretrained('BM-K/KoDiffCSE-RoBERTa') sentences = ['치타가 들판을 가로 질러 먹이를 쫓는다.', '치타 한 마리가 먹이 뒤에서 달리고 있다.', '원숭이 한 마리가 드럼을 연주한다.'] inputs = tokenizer(sentences, padding=True, truncation=True, return_tensors="pt") embeddings, _ = model(inputs, return_dict=False) score01 = cal_score(embeddings[0][0], embeddings[1][0]) # 84.56 # '치타가 들판을 가로 질러 먹이를 쫓는다.' @ '치타 한 마리가 먹이 뒤에서 달리고 있다.' score02 = cal_score(embeddings[0][0], embeddings[2][0]) # 48.06 # '치타가 들판을 가로 질러 먹이를 쫓는다.' @ '원숭이 한 마리가 드럼을 연주한다.' ``` ## Setups [![Python](https://img.shields.io/badge/python-3.8.5-blue?logo=python&logoColor=FED643)](https://www.python.org/downloads/release/python-385/) [![Pytorch](https://img.shields.io/badge/pytorch-1.7.1-red?logo=pytorch)](https://pytorch.org/get-started/previous-versions/) ## Encoder Models Baseline encoders used for korean sentence embedding - [KLUE-PLMs](https://github.com/KLUE-benchmark/KLUE/blob/main/README.md) \| Model \| Embedding size \| Hidden size \| # Layers \| # Heads \| \|----------------------\|----------------\|-------------\|----------\|---------\| \| KLUE-BERT-base \| 768 \| 768 \| 12 \| 12 \| \| KLUE-RoBERTa-base \| 768 \| 768 \| 12 \| 12 \| > Warning <br> > Large pre-trained models need a lot of GPU memory to train ## Datasets The data must exist in the "--path_to_data" folder - [wiki-corpus](https://github.com/jeongukjae/korean-wikipedia-corpus) (Unsupervised Training) - [KorSTS](https://github.com/kakaobrain/KorNLUDatasets) (Validation & Testing) ## Training - unsupervised ``` python main.py \ --model klue/roberta-base \ --generator_name klue/roberta-small \ --multi_gpu True \ --train True \ --test False \ --max_len 64 \ --batch_size 256 \ --epochs 1 \ --eval_steps 125 \ --lr 0.00005 \ --masking_ratio 0.15 \ --lambda_weight 0.005 \ --warmup_ratio 0.05 \ --temperature 0.05 \ --path_to_data Dataset/ \ --train_data wiki_corpus_examples.txt \ --valid_data valid_sts.tsv \ --ckpt best_checkpoint.pt ``` ``` bash run_diff.sh ``` > Note** <br> > Using roberta as an encoder is beneficial for training because the KoBERT model cannot build a small-sized generator. ## Evaluation ``` python main.py \ --model klue/roberta-base \ --generator klue/roberta-small \ --train False \ --test True \ --max_len 64 \ --batch_size 256 \ --path_to_data Dataset/ \ --test_data test_sts.tsv \ --path_to_saved_model output/best_checkpoint.pt ``` ## Performance - unsupervised \| Model \| Average \| Cosine Pearson \| Cosine Spearman \| Euclidean Pearson \| Euclidean Spearman \| Manhattan Pearson \| Manhattan Spearman \| Dot Pearson \| Dot Spearman \| \|------------------------\|:----:\|:----:\|:----:\|:----:\|:----:\|:----:\|:----:\|:----:\|:----:\| \| KoSRoBERTa-base<sup>†</sup> \| N/A \| N/A \| 48.96 \| N/A \| N/A \| N/A \| N/A \| N/A \| N/A \| \| KoSRoBERTa-large<sup>†</sup> \| N/A \| N/A \| 51.35 \| N/A \| N/A \| N/A \| N/A \| N/A \| N/A \| \| \| \| \| \| \| \| \| \| \| \| \| KoSimCSE-BERT \| 74.08 \| 74.92 \| 73.98 \| 74.15 \| 74.22 \| 74.07 \| 74.07 \| 74.15 \| 73.14 \| \| KoSimCSE-RoBERTa \| 75.27 \| 75.93 \| 75.00 \| 75.28 \| 75.01 \| 75.17 \| 74.83 \| 75.95 \| 75.01 \| \| \| \| \| \| \| \| \| \| \| \| \| KoDiffCSE-RoBERTa \| 77.17 \| 77.73 \| 76.96 \| 77.21 \| 76.89 \| 77.11 \| 76.81 \| 77.74 \| 76.97 \| - [Korean-SRoBERTa<sup>†</sup>](https://arxiv.org/abs/2004.03289) ## License This work is licensed under a <a rel="license" href="http://creativecommons.org/licenses/by-sa/4.0/">Creative Commons Attribution-ShareAlike 4.0 International License</a>. <a rel="license" href="http://creativecommons.org/licenses/by-sa/4.0/"><img alt="Creative Commons License" style="border-width:0" src="https://i.creativecommons.org/l/by-sa/4.0/88x31.png" /></a><br /> ## References ```bibtex @inproceedings{chuang2022diffcse, title={{DiffCSE}: Difference-based Contrastive Learning for Sentence Embeddings}, author={Chuang, Yung-Sung and Dangovski, Rumen and Luo, Hongyin and Zhang, Yang and Chang, Shiyu and Soljacic, Marin and Li, Shang-Wen and Yih, Wen-tau and Kim, Yoon and Glass, James}, booktitle={Annual Conference of the North American Chapter of the Association for Computational Linguistics (NAACL)}, year={2022} } @misc{park2021klue, title={KLUE: Korean Language Understanding Evaluation}, author={Sungjoon Park and Jihyung Moon and Sungdong Kim and Won Ik Cho and Jiyoon Han and Jangwon Park and Chisung Song and Junseong Kim and Yongsook Song and Taehwan Oh and Joohong Lee and Juhyun Oh and Sungwon Lyu and Younghoon Jeong and Inkwon Lee and Sangwoo Seo and Dongjun Lee and Hyunwoo Kim and Myeonghwa Lee and Seongbo Jang and Seungwon Do and Sunkyoung Kim and Kyungtae Lim and Jongwon Lee and Kyumin Park and Jamin Shin and Seonghyun Kim and Lucy Park and Alice Oh and Jungwoo Ha and Kyunghyun Cho}, year={2021}, eprint={2105.09680}, archivePrefix={arXiv}, primaryClass={cs.CL} } @article{ham2020kornli, title={KorNLI and KorSTS: New Benchmark Datasets for Korean Natural Language Understanding}, author={Ham, Jiyeon and Choe, Yo Joong and Park, Kyubyong and Choi, Ilji and Soh, Hyungjoon}, journal={arXiv preprint arXiv:2004.03289}, year={2020} } ```
Shivam098/opt-Translator	Shivam098	2023-08-30T05:16:01Z	79	0	transformers	[ "transformers", "pytorch", "mbart", "text2text-generation", "generated_from_trainer", "dataset:opus100", "base_model:Shivam098/opt-translation", "base_model:quantized:Shivam098/opt-translation", "autotrain_compatible", "endpoints_compatible", "4-bit", "gptq", "region:us" ]	text2text-generation	2023-08-29T13:17:01Z	--- base_model: Shivam098/opt-translation tags: - generated_from_trainer datasets: - opus100 model-index: - name: opt-Translator results: [] --- <!-- This model card has been generated automatically according to the information the Trainer had access to. You should probably proofread and complete it, then remove this comment. --> # opt-Translator This model is a fine-tuned version of [Shivam098/opt-translation](https://huggingface.co/Shivam098/opt-translation) on the opus100 dataset. ## Model description More information needed ## Intended uses & limitations More information needed ## Training and evaluation data More information needed ## Training procedure ### Training hyperparameters The following hyperparameters were used during training: - learning_rate: 5e-05 - train_batch_size: 2 - eval_batch_size: 8 - seed: 42 - optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08 - lr_scheduler_type: linear - training_steps: 1000 ### Training results ### Framework versions - Transformers 4.33.0.dev0 - Pytorch 2.0.1+cu118 - Datasets 2.14.4 - Tokenizers 0.13.3
AmlanSamanta/ludwig-webinar	AmlanSamanta	2023-08-30T04:40:12Z	1	0	peft	[ "peft", "region:us" ]	null	2023-08-30T04:40:06Z	--- library_name: peft --- ## Training procedure The following `bitsandbytes` quantization config was used during training: - quant_method: bitsandbytes - load_in_8bit: False - load_in_4bit: True - llm_int8_threshold: 6.0 - llm_int8_skip_modules: None - llm_int8_enable_fp32_cpu_offload: False - llm_int8_has_fp16_weight: False - bnb_4bit_quant_type: nf4 - bnb_4bit_use_double_quant: True - bnb_4bit_compute_dtype: float16 The following `bitsandbytes` quantization config was used during training: - quant_method: bitsandbytes - load_in_8bit: False - load_in_4bit: True - llm_int8_threshold: 6.0 - llm_int8_skip_modules: None - llm_int8_enable_fp32_cpu_offload: False - llm_int8_has_fp16_weight: False - bnb_4bit_quant_type: nf4 - bnb_4bit_use_double_quant: True - bnb_4bit_compute_dtype: float16 ### Framework versions - PEFT 0.5.0 - PEFT 0.5.0
MohanaSri/ppo-Pendulum-v2	MohanaSri	2023-08-30T03:51:27Z	1	0	stable-baselines3	[ "stable-baselines3", "Pendulum-v1", "deep-reinforcement-learning", "reinforcement-learning", "model-index", "region:us" ]	reinforcement-learning	2023-08-30T03:51:08Z	--- library_name: stable-baselines3 tags: - Pendulum-v1 - deep-reinforcement-learning - reinforcement-learning - stable-baselines3 model-index: - name: PPO results: - task: type: reinforcement-learning name: reinforcement-learning dataset: name: Pendulum-v1 type: Pendulum-v1 metrics: - type: mean_reward value: -1348.57 +/- 247.55 name: mean_reward verified: false --- # PPO Agent playing Pendulum-v1 This is a trained model of a PPO agent playing Pendulum-v1 using the [stable-baselines3 library](https://github.com/DLR-RM/stable-baselines3). ## Usage (with Stable-baselines3) TODO: Add your code ```python from stable_baselines3 import ... from huggingface_sb3 import load_from_hub ... ```
RajkNakka/a2c-PandaReachDense-v3	RajkNakka	2023-08-30T03:41:38Z	3	0	stable-baselines3	[ "stable-baselines3", "PandaReachDense-v3", "deep-reinforcement-learning", "reinforcement-learning", "model-index", "region:us" ]	reinforcement-learning	2023-08-30T03:36:11Z	--- library_name: stable-baselines3 tags: - PandaReachDense-v3 - deep-reinforcement-learning - reinforcement-learning - stable-baselines3 model-index: - name: A2C results: - task: type: reinforcement-learning name: reinforcement-learning dataset: name: PandaReachDense-v3 type: PandaReachDense-v3 metrics: - type: mean_reward value: -0.27 +/- 0.14 name: mean_reward verified: false --- # A2C Agent playing PandaReachDense-v3 This is a trained model of a A2C agent playing PandaReachDense-v3 using the [stable-baselines3 library](https://github.com/DLR-RM/stable-baselines3). ## Usage (with Stable-baselines3) TODO: Add your code ```python from stable_baselines3 import ... from huggingface_sb3 import load_from_hub ... ```
nightdude/config_611	nightdude	2023-08-30T03:40:54Z	0	0	peft	[ "peft", "region:us" ]	null	2023-08-30T03:40:29Z	--- library_name: peft --- ## Training procedure The following `bitsandbytes` quantization config was used during training: - quant_method: bitsandbytes - load_in_8bit: False - load_in_4bit: True - llm_int8_threshold: 6.0 - llm_int8_skip_modules: None - llm_int8_enable_fp32_cpu_offload: False - llm_int8_has_fp16_weight: False - bnb_4bit_quant_type: nf4 - bnb_4bit_use_double_quant: True - bnb_4bit_compute_dtype: bfloat16 ### Framework versions - PEFT 0.5.0.dev0
sang-kyung/train_monster_toy1	sang-kyung	2023-08-30T03:38:19Z	1	0	diffusers	[ "diffusers", "stable-diffusion", "stable-diffusion-diffusers", "text-to-image", "lora", "base_model:stabilityai/stable-diffusion-2-1-base", "base_model:adapter:stabilityai/stable-diffusion-2-1-base", "license:creativeml-openrail-m", "region:us" ]	text-to-image	2023-08-30T03:29:39Z	--- license: creativeml-openrail-m base_model: stabilityai/stable-diffusion-2-1-base instance_prompt: a photo of sks monster toy tags: - stable-diffusion - stable-diffusion-diffusers - text-to-image - diffusers - lora inference: true --- # LoRA DreamBooth - sang-kyung/train_monster_toy1 These are LoRA adaption weights for stabilityai/stable-diffusion-2-1-base. The weights were trained on a photo of sks monster toy using [DreamBooth](https://dreambooth.github.io/). You can find some example images in the following. ![img_0](./image_0.png) ![img_1](./image_1.png) ![img_2](./image_2.png) ![img_3](./image_3.png) LoRA for the text encoder was enabled: False.
nightdude/config_601	nightdude	2023-08-30T03:35:59Z	0	0	peft	[ "peft", "region:us" ]	null	2023-08-30T03:34:59Z	--- library_name: peft --- ## Training procedure The following `bitsandbytes` quantization config was used during training: - quant_method: bitsandbytes - load_in_8bit: False - load_in_4bit: True - llm_int8_threshold: 6.0 - llm_int8_skip_modules: None - llm_int8_enable_fp32_cpu_offload: False - llm_int8_has_fp16_weight: False - bnb_4bit_quant_type: nf4 - bnb_4bit_use_double_quant: True - bnb_4bit_compute_dtype: bfloat16 ### Framework versions - PEFT 0.5.0.dev0
gigant/distilhubert-audio-course-finetuned-gtzan-v5	gigant	2023-08-30T03:32:09Z	159	0	transformers	[ "transformers", "pytorch", "hubert", "audio-classification", "generated_from_trainer", "dataset:marsyas/gtzan", "base_model:ntu-spml/distilhubert", "base_model:finetune:ntu-spml/distilhubert", "license:apache-2.0", "model-index", "endpoints_compatible", "region:us" ]	audio-classification	2023-08-29T23:41:25Z	--- license: apache-2.0 base_model: ntu-spml/distilhubert tags: - generated_from_trainer datasets: - marsyas/gtzan metrics: - accuracy model-index: - name: distilhubert-audio-course-finetuned-gtzan-v5 results: - task: name: Audio Classification type: audio-classification dataset: name: GTZAN type: marsyas/gtzan config: all split: train args: all metrics: - name: Accuracy type: accuracy value: 0.87 --- <!-- This model card has been generated automatically according to the information the Trainer had access to. You should probably proofread and complete it, then remove this comment. --> # distilhubert-audio-course-finetuned-gtzan-v5 This model is a fine-tuned version of [ntu-spml/distilhubert](https://huggingface.co/ntu-spml/distilhubert) on the GTZAN dataset. It achieves the following results on the evaluation set: - Loss: 0.9236 - Accuracy: 0.87 ## Model description More information needed ## Intended uses & limitations More information needed ## Training and evaluation data More information needed ## Training procedure ### Training hyperparameters The following hyperparameters were used during training: - learning_rate: 0.0001 - train_batch_size: 8 - eval_batch_size: 8 - seed: 42 - gradient_accumulation_steps: 2 - total_train_batch_size: 16 - optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08 - lr_scheduler_type: linear - lr_scheduler_warmup_ratio: 0.7 - num_epochs: 50 ### Training results \| Training Loss \| Epoch \| Step \| Validation Loss \| Accuracy \| \|:-------------:\|:-----:\|:----:\|:---------------:\|:--------:\| \| 2.2989 \| 0.99 \| 56 \| 2.2882 \| 0.11 \| \| 2.2716 \| 2.0 \| 113 \| 2.2469 \| 0.31 \| \| 2.1919 \| 2.99 \| 169 \| 2.1317 \| 0.4 \| \| 2.0117 \| 4.0 \| 226 \| 1.9244 \| 0.53 \| \| 1.7966 \| 4.99 \| 282 \| 1.7315 \| 0.65 \| \| 1.6379 \| 6.0 \| 339 \| 1.5920 \| 0.59 \| \| 1.4496 \| 6.99 \| 395 \| 1.3539 \| 0.71 \| \| 1.3264 \| 8.0 \| 452 \| 1.1879 \| 0.7 \| \| 1.0601 \| 8.99 \| 508 \| 1.1342 \| 0.7 \| \| 0.9737 \| 10.0 \| 565 \| 0.9209 \| 0.79 \| \| 0.7915 \| 10.99 \| 621 \| 0.8768 \| 0.74 \| \| 0.6432 \| 12.0 \| 678 \| 0.8060 \| 0.8 \| \| 0.5217 \| 12.99 \| 734 \| 0.6562 \| 0.85 \| \| 0.3335 \| 14.0 \| 791 \| 0.7744 \| 0.76 \| \| 0.2866 \| 14.99 \| 847 \| 0.6969 \| 0.82 \| \| 0.1425 \| 16.0 \| 904 \| 0.6378 \| 0.82 \| \| 0.1278 \| 16.99 \| 960 \| 0.6972 \| 0.82 \| \| 0.0706 \| 18.0 \| 1017 \| 0.7328 \| 0.84 \| \| 0.0301 \| 18.99 \| 1073 \| 0.9245 \| 0.76 \| \| 0.0379 \| 20.0 \| 1130 \| 0.8437 \| 0.85 \| \| 0.0147 \| 20.99 \| 1186 \| 0.7489 \| 0.83 \| \| 0.0067 \| 22.0 \| 1243 \| 0.8975 \| 0.83 \| \| 0.0049 \| 22.99 \| 1299 \| 1.1788 \| 0.81 \| \| 0.0038 \| 24.0 \| 1356 \| 1.1146 \| 0.81 \| \| 0.0028 \| 24.99 \| 1412 \| 1.0270 \| 0.85 \| \| 0.0027 \| 26.0 \| 1469 \| 1.0634 \| 0.83 \| \| 0.0024 \| 26.99 \| 1525 \| 1.0220 \| 0.84 \| \| 0.0023 \| 28.0 \| 1582 \| 1.0282 \| 0.83 \| \| 0.0487 \| 28.99 \| 1638 \| 1.0735 \| 0.82 \| \| 0.0458 \| 30.0 \| 1695 \| 1.1198 \| 0.82 \| \| 0.2453 \| 30.99 \| 1751 \| 1.1154 \| 0.81 \| \| 0.0552 \| 32.0 \| 1808 \| 1.1630 \| 0.79 \| \| 0.1202 \| 32.99 \| 1864 \| 1.2746 \| 0.81 \| \| 0.2709 \| 34.0 \| 1921 \| 1.3797 \| 0.79 \| \| 0.275 \| 34.99 \| 1977 \| 1.5372 \| 0.75 \| \| 0.1268 \| 36.0 \| 2034 \| 0.8140 \| 0.86 \| \| 0.1582 \| 36.99 \| 2090 \| 1.4153 \| 0.77 \| \| 0.0054 \| 38.0 \| 2147 \| 1.3796 \| 0.79 \| \| 0.0299 \| 38.99 \| 2203 \| 1.3653 \| 0.78 \| \| 0.0199 \| 40.0 \| 2260 \| 0.9987 \| 0.87 \| \| 0.0021 \| 40.99 \| 2316 \| 1.0689 \| 0.84 \| \| 0.0007 \| 42.0 \| 2373 \| 1.0383 \| 0.85 \| \| 0.0006 \| 42.99 \| 2429 \| 1.0493 \| 0.84 \| \| 0.0006 \| 44.0 \| 2486 \| 1.0744 \| 0.85 \| \| 0.0005 \| 44.99 \| 2542 \| 0.9151 \| 0.86 \| \| 0.0004 \| 46.0 \| 2599 \| 0.8946 \| 0.87 \| \| 0.01 \| 46.99 \| 2655 \| 0.8960 \| 0.88 \| \| 0.0073 \| 48.0 \| 2712 \| 0.9485 \| 0.87 \| \| 0.0004 \| 48.99 \| 2768 \| 0.9247 \| 0.87 \| \| 0.0004 \| 49.56 \| 2800 \| 0.9236 \| 0.87 \| ### Framework versions - Transformers 4.32.1 - Pytorch 2.0.1+cu117 - Datasets 2.14.4 - Tokenizers 0.13.3
abeiler/huggingface-goatLora-goatV10-testData	abeiler	2023-08-30T03:19:52Z	0	0	null	[ "pytorch", "tensorboard", "generated_from_trainer", "region:us" ]	null	2023-08-30T03:07:07Z	--- tags: - generated_from_trainer model-index: - name: huggingface-goatLora-goatV10-testData results: [] --- <!-- This model card has been generated automatically according to the information the Trainer had access to. You should probably proofread and complete it, then remove this comment. --> # huggingface-goatLora-goatV10-testData This model is a fine-tuned version of [meta-llama/Llama-2-7b-hf](https://huggingface.co/meta-llama/Llama-2-7b-hf) on an unknown dataset. ## Model description More information needed ## Intended uses & limitations More information needed ## Training and evaluation data More information needed ## Training procedure ### Training hyperparameters The following hyperparameters were used during training: - learning_rate: 0.0003 - train_batch_size: 16 - eval_batch_size: 8 - seed: 42 - gradient_accumulation_steps: 8 - total_train_batch_size: 128 - optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08 - lr_scheduler_type: linear - lr_scheduler_warmup_steps: 100 - num_epochs: 1 - mixed_precision_training: Native AMP ### Training results ### Framework versions - Transformers 4.28.1 - Pytorch 2.0.0 - Datasets 2.12.0 - Tokenizers 0.13.3
Adrian101-hnd/ppo-Huggy	Adrian101-hnd	2023-08-30T03:13:37Z	1	0	ml-agents	[ "ml-agents", "tensorboard", "onnx", "Huggy", "deep-reinforcement-learning", "reinforcement-learning", "ML-Agents-Huggy", "region:us" ]	reinforcement-learning	2023-08-30T03:13:32Z	--- library_name: ml-agents tags: - Huggy - deep-reinforcement-learning - reinforcement-learning - ML-Agents-Huggy --- # ppo Agent playing Huggy This is a trained model of a ppo agent playing Huggy using the [Unity ML-Agents Library](https://github.com/Unity-Technologies/ml-agents). ## Usage (with ML-Agents) The Documentation: https://unity-technologies.github.io/ml-agents/ML-Agents-Toolkit-Documentation/ We wrote a complete tutorial to learn to train your first agent using ML-Agents and publish it to the Hub: - A short tutorial where you teach Huggy the Dog 🐶 to fetch the stick and then play with him directly in your browser: https://huggingface.co/learn/deep-rl-course/unitbonus1/introduction - A longer tutorial to understand how works ML-Agents: https://huggingface.co/learn/deep-rl-course/unit5/introduction ### Resume the training ```bash mlagents-learn <your_configuration_file_path.yaml> --run-id=<run_id> --resume ``` ### Watch your Agent play You can watch your agent playing directly in your browser 1. If the environment is part of ML-Agents official environments, go to https://huggingface.co/unity 2. Step 1: Find your model_id: Adrian101-hnd/ppo-Huggy 3. Step 2: Select your .nn /.onnx file 4. Click on Watch the agent play 👀
hecool108/ct-p1	hecool108	2023-08-30T03:05:55Z	2	0	diffusers	[ "diffusers", "safetensors", "text-generation-inference", "ctee", "text-to-image", "en", "license:openrail", "region:us" ]	text-to-image	2023-08-30T02:58:26Z	--- license: openrail language: - en tags: - text-generation-inference - ctee pipeline_tag: text-to-image ---
uer/chinese_roberta_L-12_H-256	uer	2023-08-30T02:30:45Z	121	0	transformers	[ "transformers", "pytorch", "tf", "jax", "bert", "fill-mask", "zh", "dataset:CLUECorpusSmall", "arxiv:1909.05658", "arxiv:2212.06385", "arxiv:1908.08962", "arxiv:1810.04805", "arxiv:1907.11692", "autotrain_compatible", "endpoints_compatible", "region:us" ]	fill-mask	2022-03-02T23:29:05Z	--- language: zh datasets: CLUECorpusSmall widget: - text: "北京是[MASK]国的首都。" --- # Chinese RoBERTa Miniatures ## Model description This is the set of 24 Chinese RoBERTa models pre-trained by [UER-py](https://github.com/dbiir/UER-py/), which is introduced in [this paper](https://arxiv.org/abs/1909.05658). Besides, the models could also be pre-trained by [TencentPretrain](https://github.com/Tencent/TencentPretrain) introduced in [this paper](https://arxiv.org/abs/2212.06385), which inherits UER-py to support models with parameters above one billion, and extends it to a multimodal pre-training framework. [Turc et al.](https://arxiv.org/abs/1908.08962) have shown that the standard BERT recipe is effective on a wide range of model sizes. Following their paper, we released the 24 Chinese RoBERTa models. In order to facilitate users in reproducing the results, we used a publicly available corpus and provided all training details. You can download the 24 Chinese RoBERTa miniatures either from the [UER-py Modelzoo page](https://github.com/dbiir/UER-py/wiki/Modelzoo), or via HuggingFace from the links below: \| \| H=128 \| H=256 \| H=512 \| H=768 \| \| -------- \| :-----------------------: \| :-----------------------: \| :-------------------------: \| :-------------------------: \| \| L=2 \| [2/128 (Tiny)][2_128] \| [2/256][2_256] \| [2/512][2_512] \| [2/768][2_768] \| \| L=4 \| [4/128][4_128] \| [4/256 (Mini)][4_256] \| [4/512 (Small)][4_512] \| [4/768][4_768] \| \| L=6 \| [6/128][6_128] \| [6/256][6_256] \| [6/512][6_512] \| [6/768][6_768] \| \| L=8 \| [8/128][8_128] \| [8/256][8_256] \| [8/512 (Medium)][8_512] \| [8/768][8_768] \| \| L=10 \| [10/128][10_128] \| [10/256][10_256] \| [10/512][10_512] \| [10/768][10_768] \| \| L=12 \| [12/128][12_128] \| [12/256][12_256] \| [12/512][12_512] \| [12/768 (Base)][12_768] \| Here are scores on the devlopment set of six Chinese tasks: \| Model \| Score \| book_review \| chnsenticorp \| lcqmc \| tnews(CLUE) \| iflytek(CLUE) \| ocnli(CLUE) \| \| -------------- \| :---: \| :----: \| :----------: \| :---: \| :---------: \| :-----------: \| :---------: \| \| RoBERTa-Tiny \| 72.3 \| 83.4 \| 91.4 \| 81.8 \| 62.0 \| 55.0 \| 60.3 \| \| RoBERTa-Mini \| 75.9 \| 85.7 \| 93.7 \| 86.1 \| 63.9 \| 58.3 \| 67.4 \| \| RoBERTa-Small \| 76.9 \| 87.5 \| 93.4 \| 86.5 \| 65.1 \| 59.4 \| 69.7 \| \| RoBERTa-Medium \| 78.0 \| 88.7 \| 94.8 \| 88.1 \| 65.6 \| 59.5 \| 71.2 \| \| RoBERTa-Base \| 79.7 \| 90.1 \| 95.2 \| 89.2 \| 67.0 \| 60.9 \| 75.5 \| For each task, we selected the best fine-tuning hyperparameters from the lists below, and trained with the sequence length of 128: - epochs: 3, 5, 8 - batch sizes: 32, 64 - learning rates: 3e-5, 1e-4, 3e-4 ## How to use You can use this model directly with a pipeline for masked language modeling (take the case of RoBERTa-Medium): ```python >>> from transformers import pipeline >>> unmasker = pipeline('fill-mask', model='uer/chinese_roberta_L-8_H-512') >>> unmasker("中国的首都是[MASK]京。") [ {'sequence': '[CLS] 中国的首都是北京。 [SEP]', 'score': 0.8701988458633423, 'token': 1266, 'token_str': '北'}, {'sequence': '[CLS] 中国的首都是南京。 [SEP]', 'score': 0.1194809079170227, 'token': 1298, 'token_str': '南'}, {'sequence': '[CLS] 中国的首都是东京。 [SEP]', 'score': 0.0037803512532263994, 'token': 691, 'token_str': '东'}, {'sequence': '[CLS] 中国的首都是普京。 [SEP]', 'score': 0.0017127094324678183, 'token': 3249, 'token_str': '普'}, {'sequence': '[CLS] 中国的首都是望京。 [SEP]', 'score': 0.001687526935711503, 'token': 3307, 'token_str': '望'} ] ``` Here is how to use this model to get the features of a given text in PyTorch: ```python from transformers import BertTokenizer, BertModel tokenizer = BertTokenizer.from_pretrained('uer/chinese_roberta_L-8_H-512') model = BertModel.from_pretrained("uer/chinese_roberta_L-8_H-512") text = "用你喜欢的任何文本替换我。" encoded_input = tokenizer(text, return_tensors='pt') output = model(**encoded_input) ``` and in TensorFlow: ```python from transformers import BertTokenizer, TFBertModel tokenizer = BertTokenizer.from_pretrained('uer/chinese_roberta_L-8_H-512') model = TFBertModel.from_pretrained("uer/chinese_roberta_L-8_H-512") text = "用你喜欢的任何文本替换我。" encoded_input = tokenizer(text, return_tensors='tf') output = model(encoded_input) ``` ## Training data [CLUECorpusSmall](https://github.com/CLUEbenchmark/CLUECorpus2020/) is used as training data. We found that models pre-trained on CLUECorpusSmall outperform those pre-trained on CLUECorpus2020, although CLUECorpus2020 is much larger than CLUECorpusSmall. ## Training procedure Models are pre-trained by [UER-py](https://github.com/dbiir/UER-py/) on [Tencent Cloud](https://cloud.tencent.com/). We pre-train 1,000,000 steps with a sequence length of 128 and then pre-train 250,000 additional steps with a sequence length of 512. We use the same hyper-parameters on different model sizes. Taking the case of RoBERTa-Medium Stage1: ``` python3 preprocess.py --corpus_path corpora/cluecorpussmall.txt \ --vocab_path models/google_zh_vocab.txt \ --dataset_path cluecorpussmall_seq128_dataset.pt \ --processes_num 32 --seq_length 128 \ --dynamic_masking --data_processor mlm ``` ``` python3 pretrain.py --dataset_path cluecorpussmall_seq128_dataset.pt \ --vocab_path models/google_zh_vocab.txt \ --config_path models/bert/medium_config.json \ --output_model_path models/cluecorpussmall_roberta_medium_seq128_model.bin \ --world_size 8 --gpu_ranks 0 1 2 3 4 5 6 7 \ --total_steps 1000000 --save_checkpoint_steps 100000 --report_steps 50000 \ --learning_rate 1e-4 --batch_size 64 \ --data_processor mlm --target mlm ``` Stage2: ``` python3 preprocess.py --corpus_path corpora/cluecorpussmall.txt \ --vocab_path models/google_zh_vocab.txt \ --dataset_path cluecorpussmall_seq512_dataset.pt \ --processes_num 32 --seq_length 512 \ --dynamic_masking --data_processor mlm ``` ``` python3 pretrain.py --dataset_path cluecorpussmall_seq512_dataset.pt \ --vocab_path models/google_zh_vocab.txt \ --pretrained_model_path models/cluecorpussmall_roberta_medium_seq128_model.bin-1000000 \ --config_path models/bert/medium_config.json \ --output_model_path models/cluecorpussmall_roberta_medium_seq512_model.bin \ --world_size 8 --gpu_ranks 0 1 2 3 4 5 6 7 \ --total_steps 250000 --save_checkpoint_steps 50000 --report_steps 10000 \ --learning_rate 5e-5 --batch_size 16 \ --data_processor mlm --target mlm ``` Finally, we convert the pre-trained model into Huggingface's format: ``` python3 scripts/convert_bert_from_uer_to_huggingface.py --input_model_path models/cluecorpussmall_roberta_medium_seq512_model.bin-250000 \ --output_model_path pytorch_model.bin \ --layers_num 8 --type mlm ``` ### BibTeX entry and citation info ``` @article{devlin2018bert, title={Bert: Pre-training of deep bidirectional transformers for language understanding}, author={Devlin, Jacob and Chang, Ming-Wei and Lee, Kenton and Toutanova, Kristina}, journal={arXiv preprint arXiv:1810.04805}, year={2018} } @article{liu2019roberta, title={Roberta: A robustly optimized bert pretraining approach}, author={Liu, Yinhan and Ott, Myle and Goyal, Naman and Du, Jingfei and Joshi, Mandar and Chen, Danqi and Levy, Omer and Lewis, Mike and Zettlemoyer, Luke and Stoyanov, Veselin}, journal={arXiv preprint arXiv:1907.11692}, year={2019} } @article{turc2019, title={Well-Read Students Learn Better: On the Importance of Pre-training Compact Models}, author={Turc, Iulia and Chang, Ming-Wei and Lee, Kenton and Toutanova, Kristina}, journal={arXiv preprint arXiv:1908.08962v2 }, year={2019} } @article{zhao2019uer, title={UER: An Open-Source Toolkit for Pre-training Models}, author={Zhao, Zhe and Chen, Hui and Zhang, Jinbin and Zhao, Xin and Liu, Tao and Lu, Wei and Chen, Xi and Deng, Haotang and Ju, Qi and Du, Xiaoyong}, journal={EMNLP-IJCNLP 2019}, pages={241}, year={2019} } @article{zhao2023tencentpretrain, title={TencentPretrain: A Scalable and Flexible Toolkit for Pre-training Models of Different Modalities}, author={Zhao, Zhe and Li, Yudong and Hou, Cheng and Zhao, Jing and others}, journal={ACL 2023}, pages={217}, year={2023} } ``` [2_128]:https://huggingface.co/uer/chinese_roberta_L-2_H-128 [2_256]:https://huggingface.co/uer/chinese_roberta_L-2_H-256 [2_512]:https://huggingface.co/uer/chinese_roberta_L-2_H-512 [2_768]:https://huggingface.co/uer/chinese_roberta_L-2_H-768 [4_128]:https://huggingface.co/uer/chinese_roberta_L-4_H-128 [4_256]:https://huggingface.co/uer/chinese_roberta_L-4_H-256 [4_512]:https://huggingface.co/uer/chinese_roberta_L-4_H-512 [4_768]:https://huggingface.co/uer/chinese_roberta_L-4_H-768 [6_128]:https://huggingface.co/uer/chinese_roberta_L-6_H-128 [6_256]:https://huggingface.co/uer/chinese_roberta_L-6_H-256 [6_512]:https://huggingface.co/uer/chinese_roberta_L-6_H-512 [6_768]:https://huggingface.co/uer/chinese_roberta_L-6_H-768 [8_128]:https://huggingface.co/uer/chinese_roberta_L-8_H-128 [8_256]:https://huggingface.co/uer/chinese_roberta_L-8_H-256 [8_512]:https://huggingface.co/uer/chinese_roberta_L-8_H-512 [8_768]:https://huggingface.co/uer/chinese_roberta_L-8_H-768 [10_128]:https://huggingface.co/uer/chinese_roberta_L-10_H-128 [10_256]:https://huggingface.co/uer/chinese_roberta_L-10_H-256 [10_512]:https://huggingface.co/uer/chinese_roberta_L-10_H-512 [10_768]:https://huggingface.co/uer/chinese_roberta_L-10_H-768 [12_128]:https://huggingface.co/uer/chinese_roberta_L-12_H-128 [12_256]:https://huggingface.co/uer/chinese_roberta_L-12_H-256 [12_512]:https://huggingface.co/uer/chinese_roberta_L-12_H-512 [12_768]:https://huggingface.co/uer/chinese_roberta_L-12_H-768
uer/chinese_roberta_L-10_H-512	uer	2023-08-30T02:29:34Z	111	0	transformers	[ "transformers", "pytorch", "tf", "jax", "bert", "fill-mask", "zh", "dataset:CLUECorpusSmall", "arxiv:1909.05658", "arxiv:2212.06385", "arxiv:1908.08962", "arxiv:1810.04805", "arxiv:1907.11692", "autotrain_compatible", "endpoints_compatible", "region:us" ]	fill-mask	2022-03-02T23:29:05Z	--- language: zh datasets: CLUECorpusSmall widget: - text: "北京是[MASK]国的首都。" --- # Chinese RoBERTa Miniatures ## Model description This is the set of 24 Chinese RoBERTa models pre-trained by [UER-py](https://github.com/dbiir/UER-py/), which is introduced in [this paper](https://arxiv.org/abs/1909.05658). Besides, the models could also be pre-trained by [TencentPretrain](https://github.com/Tencent/TencentPretrain) introduced in [this paper](https://arxiv.org/abs/2212.06385), which inherits UER-py to support models with parameters above one billion, and extends it to a multimodal pre-training framework. [Turc et al.](https://arxiv.org/abs/1908.08962) have shown that the standard BERT recipe is effective on a wide range of model sizes. Following their paper, we released the 24 Chinese RoBERTa models. In order to facilitate users in reproducing the results, we used a publicly available corpus and provided all training details. You can download the 24 Chinese RoBERTa miniatures either from the [UER-py Modelzoo page](https://github.com/dbiir/UER-py/wiki/Modelzoo), or via HuggingFace from the links below: \| \| H=128 \| H=256 \| H=512 \| H=768 \| \| -------- \| :-----------------------: \| :-----------------------: \| :-------------------------: \| :-------------------------: \| \| L=2 \| [2/128 (Tiny)][2_128] \| [2/256][2_256] \| [2/512][2_512] \| [2/768][2_768] \| \| L=4 \| [4/128][4_128] \| [4/256 (Mini)][4_256] \| [4/512 (Small)][4_512] \| [4/768][4_768] \| \| L=6 \| [6/128][6_128] \| [6/256][6_256] \| [6/512][6_512] \| [6/768][6_768] \| \| L=8 \| [8/128][8_128] \| [8/256][8_256] \| [8/512 (Medium)][8_512] \| [8/768][8_768] \| \| L=10 \| [10/128][10_128] \| [10/256][10_256] \| [10/512][10_512] \| [10/768][10_768] \| \| L=12 \| [12/128][12_128] \| [12/256][12_256] \| [12/512][12_512] \| [12/768 (Base)][12_768] \| Here are scores on the devlopment set of six Chinese tasks: \| Model \| Score \| book_review \| chnsenticorp \| lcqmc \| tnews(CLUE) \| iflytek(CLUE) \| ocnli(CLUE) \| \| -------------- \| :---: \| :----: \| :----------: \| :---: \| :---------: \| :-----------: \| :---------: \| \| RoBERTa-Tiny \| 72.3 \| 83.4 \| 91.4 \| 81.8 \| 62.0 \| 55.0 \| 60.3 \| \| RoBERTa-Mini \| 75.9 \| 85.7 \| 93.7 \| 86.1 \| 63.9 \| 58.3 \| 67.4 \| \| RoBERTa-Small \| 76.9 \| 87.5 \| 93.4 \| 86.5 \| 65.1 \| 59.4 \| 69.7 \| \| RoBERTa-Medium \| 78.0 \| 88.7 \| 94.8 \| 88.1 \| 65.6 \| 59.5 \| 71.2 \| \| RoBERTa-Base \| 79.7 \| 90.1 \| 95.2 \| 89.2 \| 67.0 \| 60.9 \| 75.5 \| For each task, we selected the best fine-tuning hyperparameters from the lists below, and trained with the sequence length of 128: - epochs: 3, 5, 8 - batch sizes: 32, 64 - learning rates: 3e-5, 1e-4, 3e-4 ## How to use You can use this model directly with a pipeline for masked language modeling (take the case of RoBERTa-Medium): ```python >>> from transformers import pipeline >>> unmasker = pipeline('fill-mask', model='uer/chinese_roberta_L-8_H-512') >>> unmasker("中国的首都是[MASK]京。") [ {'sequence': '[CLS] 中国的首都是北京。 [SEP]', 'score': 0.8701988458633423, 'token': 1266, 'token_str': '北'}, {'sequence': '[CLS] 中国的首都是南京。 [SEP]', 'score': 0.1194809079170227, 'token': 1298, 'token_str': '南'}, {'sequence': '[CLS] 中国的首都是东京。 [SEP]', 'score': 0.0037803512532263994, 'token': 691, 'token_str': '东'}, {'sequence': '[CLS] 中国的首都是普京。 [SEP]', 'score': 0.0017127094324678183, 'token': 3249, 'token_str': '普'}, {'sequence': '[CLS] 中国的首都是望京。 [SEP]', 'score': 0.001687526935711503, 'token': 3307, 'token_str': '望'} ] ``` Here is how to use this model to get the features of a given text in PyTorch: ```python from transformers import BertTokenizer, BertModel tokenizer = BertTokenizer.from_pretrained('uer/chinese_roberta_L-8_H-512') model = BertModel.from_pretrained("uer/chinese_roberta_L-8_H-512") text = "用你喜欢的任何文本替换我。" encoded_input = tokenizer(text, return_tensors='pt') output = model(**encoded_input) ``` and in TensorFlow: ```python from transformers import BertTokenizer, TFBertModel tokenizer = BertTokenizer.from_pretrained('uer/chinese_roberta_L-8_H-512') model = TFBertModel.from_pretrained("uer/chinese_roberta_L-8_H-512") text = "用你喜欢的任何文本替换我。" encoded_input = tokenizer(text, return_tensors='tf') output = model(encoded_input) ``` ## Training data [CLUECorpusSmall](https://github.com/CLUEbenchmark/CLUECorpus2020/) is used as training data. We found that models pre-trained on CLUECorpusSmall outperform those pre-trained on CLUECorpus2020, although CLUECorpus2020 is much larger than CLUECorpusSmall. ## Training procedure Models are pre-trained by [UER-py](https://github.com/dbiir/UER-py/) on [Tencent Cloud](https://cloud.tencent.com/). We pre-train 1,000,000 steps with a sequence length of 128 and then pre-train 250,000 additional steps with a sequence length of 512. We use the same hyper-parameters on different model sizes. Taking the case of RoBERTa-Medium Stage1: ``` python3 preprocess.py --corpus_path corpora/cluecorpussmall.txt \ --vocab_path models/google_zh_vocab.txt \ --dataset_path cluecorpussmall_seq128_dataset.pt \ --processes_num 32 --seq_length 128 \ --dynamic_masking --data_processor mlm ``` ``` python3 pretrain.py --dataset_path cluecorpussmall_seq128_dataset.pt \ --vocab_path models/google_zh_vocab.txt \ --config_path models/bert/medium_config.json \ --output_model_path models/cluecorpussmall_roberta_medium_seq128_model.bin \ --world_size 8 --gpu_ranks 0 1 2 3 4 5 6 7 \ --total_steps 1000000 --save_checkpoint_steps 100000 --report_steps 50000 \ --learning_rate 1e-4 --batch_size 64 \ --data_processor mlm --target mlm ``` Stage2: ``` python3 preprocess.py --corpus_path corpora/cluecorpussmall.txt \ --vocab_path models/google_zh_vocab.txt \ --dataset_path cluecorpussmall_seq512_dataset.pt \ --processes_num 32 --seq_length 512 \ --dynamic_masking --data_processor mlm ``` ``` python3 pretrain.py --dataset_path cluecorpussmall_seq512_dataset.pt \ --vocab_path models/google_zh_vocab.txt \ --pretrained_model_path models/cluecorpussmall_roberta_medium_seq128_model.bin-1000000 \ --config_path models/bert/medium_config.json \ --output_model_path models/cluecorpussmall_roberta_medium_seq512_model.bin \ --world_size 8 --gpu_ranks 0 1 2 3 4 5 6 7 \ --total_steps 250000 --save_checkpoint_steps 50000 --report_steps 10000 \ --learning_rate 5e-5 --batch_size 16 \ --data_processor mlm --target mlm ``` Finally, we convert the pre-trained model into Huggingface's format: ``` python3 scripts/convert_bert_from_uer_to_huggingface.py --input_model_path models/cluecorpussmall_roberta_medium_seq512_model.bin-250000 \ --output_model_path pytorch_model.bin \ --layers_num 8 --type mlm ``` ### BibTeX entry and citation info ``` @article{devlin2018bert, title={Bert: Pre-training of deep bidirectional transformers for language understanding}, author={Devlin, Jacob and Chang, Ming-Wei and Lee, Kenton and Toutanova, Kristina}, journal={arXiv preprint arXiv:1810.04805}, year={2018} } @article{liu2019roberta, title={Roberta: A robustly optimized bert pretraining approach}, author={Liu, Yinhan and Ott, Myle and Goyal, Naman and Du, Jingfei and Joshi, Mandar and Chen, Danqi and Levy, Omer and Lewis, Mike and Zettlemoyer, Luke and Stoyanov, Veselin}, journal={arXiv preprint arXiv:1907.11692}, year={2019} } @article{turc2019, title={Well-Read Students Learn Better: On the Importance of Pre-training Compact Models}, author={Turc, Iulia and Chang, Ming-Wei and Lee, Kenton and Toutanova, Kristina}, journal={arXiv preprint arXiv:1908.08962v2 }, year={2019} } @article{zhao2019uer, title={UER: An Open-Source Toolkit for Pre-training Models}, author={Zhao, Zhe and Chen, Hui and Zhang, Jinbin and Zhao, Xin and Liu, Tao and Lu, Wei and Chen, Xi and Deng, Haotang and Ju, Qi and Du, Xiaoyong}, journal={EMNLP-IJCNLP 2019}, pages={241}, year={2019} } @article{zhao2023tencentpretrain, title={TencentPretrain: A Scalable and Flexible Toolkit for Pre-training Models of Different Modalities}, author={Zhao, Zhe and Li, Yudong and Hou, Cheng and Zhao, Jing and others}, journal={ACL 2023}, pages={217}, year={2023} } ``` [2_128]:https://huggingface.co/uer/chinese_roberta_L-2_H-128 [2_256]:https://huggingface.co/uer/chinese_roberta_L-2_H-256 [2_512]:https://huggingface.co/uer/chinese_roberta_L-2_H-512 [2_768]:https://huggingface.co/uer/chinese_roberta_L-2_H-768 [4_128]:https://huggingface.co/uer/chinese_roberta_L-4_H-128 [4_256]:https://huggingface.co/uer/chinese_roberta_L-4_H-256 [4_512]:https://huggingface.co/uer/chinese_roberta_L-4_H-512 [4_768]:https://huggingface.co/uer/chinese_roberta_L-4_H-768 [6_128]:https://huggingface.co/uer/chinese_roberta_L-6_H-128 [6_256]:https://huggingface.co/uer/chinese_roberta_L-6_H-256 [6_512]:https://huggingface.co/uer/chinese_roberta_L-6_H-512 [6_768]:https://huggingface.co/uer/chinese_roberta_L-6_H-768 [8_128]:https://huggingface.co/uer/chinese_roberta_L-8_H-128 [8_256]:https://huggingface.co/uer/chinese_roberta_L-8_H-256 [8_512]:https://huggingface.co/uer/chinese_roberta_L-8_H-512 [8_768]:https://huggingface.co/uer/chinese_roberta_L-8_H-768 [10_128]:https://huggingface.co/uer/chinese_roberta_L-10_H-128 [10_256]:https://huggingface.co/uer/chinese_roberta_L-10_H-256 [10_512]:https://huggingface.co/uer/chinese_roberta_L-10_H-512 [10_768]:https://huggingface.co/uer/chinese_roberta_L-10_H-768 [12_128]:https://huggingface.co/uer/chinese_roberta_L-12_H-128 [12_256]:https://huggingface.co/uer/chinese_roberta_L-12_H-256 [12_512]:https://huggingface.co/uer/chinese_roberta_L-12_H-512 [12_768]:https://huggingface.co/uer/chinese_roberta_L-12_H-768
yumjunstar/trocr-small-printedkorean-deleteunusedchar_noise	yumjunstar	2023-08-30T02:28:54Z	51	0	transformers	[ "transformers", "pytorch", "vision-encoder-decoder", "image-text-to-text", "generated_from_trainer", "endpoints_compatible", "region:us" ]	image-text-to-text	2023-08-29T17:52:22Z	--- tags: - generated_from_trainer metrics: - wer - accuracy model-index: - name: trocr-small-printedkorean-deleteunusedchar_noise results: [] --- <!-- This model card has been generated automatically according to the information the Trainer had access to. You should probably proofread and complete it, then remove this comment. --> # trocr-small-printedkorean-deleteunusedchar_noise This model was trained from scratch on an unknown dataset. It achieves the following results on the evaluation set: - Loss: 1.3375 - Cer: 0.2783 - Wer: 0.2975 - Accuracy: 45.6667 ## Model description More information needed ## Intended uses & limitations More information needed ## Training and evaluation data More information needed ## Training procedure ### Training hyperparameters The following hyperparameters were used during training: - learning_rate: 4e-05 - train_batch_size: 128 - eval_batch_size: 192 - seed: 42 - optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08 - lr_scheduler_type: linear - num_epochs: 10 - mixed_precision_training: Native AMP ### Training results \| Training Loss \| Epoch \| Step \| Validation Loss \| Cer \| Wer \| Accuracy \| \|:-------------:\|:-----:\|:-----:\|:---------------:\|:------:\|:------:\|:--------:\| \| 1.711 \| 0.43 \| 1000 \| 1.6485 \| 0.3288 \| 0.3944 \| 30.6667 \| \| 1.6849 \| 0.85 \| 2000 \| 1.5361 \| 0.3098 \| 0.3809 \| 32.3333 \| \| 1.4933 \| 1.28 \| 3000 \| 1.4302 \| 0.2935 \| 0.3533 \| 34.6667 \| \| 1.526 \| 1.71 \| 4000 \| 1.4010 \| 0.2922 \| 0.3400 \| 35.6667 \| \| 1.3422 \| 2.13 \| 5000 \| 1.3883 \| 0.2846 \| 0.3331 \| 36.0 \| \| 1.333 \| 2.56 \| 6000 \| 1.3790 \| 0.2871 \| 0.3308 \| 34.0 \| \| 1.3295 \| 2.99 \| 7000 \| 1.3644 \| 0.2876 \| 0.3294 \| 35.6667 \| \| 1.3294 \| 3.42 \| 8000 \| 1.3588 \| 0.2824 \| 0.3202 \| 36.6667 \| \| 1.3578 \| 3.84 \| 9000 \| 1.3502 \| 0.2823 \| 0.3162 \| 40.6667 \| \| 1.3029 \| 4.27 \| 10000 \| 1.3514 \| 0.2879 \| 0.3228 \| 37.0 \| \| 1.2777 \| 4.7 \| 11000 \| 1.3507 \| 0.2813 \| 0.3168 \| 38.3333 \| \| 1.1781 \| 5.12 \| 12000 \| 1.3507 \| 0.2791 \| 0.3150 \| 40.3333 \| \| 1.3025 \| 5.55 \| 13000 \| 1.3459 \| 0.2818 \| 0.3099 \| 41.6667 \| \| 1.2024 \| 5.98 \| 14000 \| 1.3401 \| 0.2801 \| 0.3061 \| 41.6667 \| \| 1.1792 \| 6.4 \| 15000 \| 1.3412 \| 0.2763 \| 0.3015 \| 44.6667 \| \| 1.1586 \| 6.83 \| 16000 \| 1.3410 \| 0.2799 \| 0.3064 \| 43.3333 \| \| 1.2098 \| 7.26 \| 17000 \| 1.3439 \| 0.2777 \| 0.3030 \| 43.6667 \| \| 1.2122 \| 7.69 \| 18000 \| 1.3418 \| 0.2816 \| 0.3050 \| 43.3333 \| \| 1.1323 \| 8.11 \| 19000 \| 1.3409 \| 0.2767 \| 0.2981 \| 45.3333 \| \| 1.2215 \| 8.54 \| 20000 \| 1.3386 \| 0.2781 \| 0.3004 \| 44.0 \| \| 1.2068 \| 8.97 \| 21000 \| 1.3375 \| 0.2762 \| 0.2972 \| 45.0 \| \| 1.0847 \| 9.39 \| 22000 \| 1.3366 \| 0.2765 \| 0.2969 \| 46.0 \| \| 1.1791 \| 9.82 \| 23000 \| 1.3375 \| 0.2783 \| 0.2975 \| 45.6667 \| ### Framework versions - Transformers 4.28.0 - Pytorch 1.13.1+cu116 - Datasets 2.14.4 - Tokenizers 0.13.3
uer/chinese_roberta_L-8_H-768	uer	2023-08-30T02:28:34Z	110	0	transformers	[ "transformers", "pytorch", "tf", "jax", "bert", "fill-mask", "zh", "dataset:CLUECorpusSmall", "arxiv:1909.05658", "arxiv:2212.06385", "arxiv:1908.08962", "arxiv:1810.04805", "arxiv:1907.11692", "autotrain_compatible", "endpoints_compatible", "region:us" ]	fill-mask	2022-03-02T23:29:05Z	--- language: zh datasets: CLUECorpusSmall widget: - text: "北京是[MASK]国的首都。" --- # Chinese RoBERTa Miniatures ## Model description This is the set of 24 Chinese RoBERTa models pre-trained by [UER-py](https://github.com/dbiir/UER-py/), which is introduced in [this paper](https://arxiv.org/abs/1909.05658). Besides, the models could also be pre-trained by [TencentPretrain](https://github.com/Tencent/TencentPretrain) introduced in [this paper](https://arxiv.org/abs/2212.06385), which inherits UER-py to support models with parameters above one billion, and extends it to a multimodal pre-training framework. [Turc et al.](https://arxiv.org/abs/1908.08962) have shown that the standard BERT recipe is effective on a wide range of model sizes. Following their paper, we released the 24 Chinese RoBERTa models. In order to facilitate users in reproducing the results, we used a publicly available corpus and provided all training details. You can download the 24 Chinese RoBERTa miniatures either from the [UER-py Modelzoo page](https://github.com/dbiir/UER-py/wiki/Modelzoo), or via HuggingFace from the links below: \| \| H=128 \| H=256 \| H=512 \| H=768 \| \| -------- \| :-----------------------: \| :-----------------------: \| :-------------------------: \| :-------------------------: \| \| L=2 \| [2/128 (Tiny)][2_128] \| [2/256][2_256] \| [2/512][2_512] \| [2/768][2_768] \| \| L=4 \| [4/128][4_128] \| [4/256 (Mini)][4_256] \| [4/512 (Small)][4_512] \| [4/768][4_768] \| \| L=6 \| [6/128][6_128] \| [6/256][6_256] \| [6/512][6_512] \| [6/768][6_768] \| \| L=8 \| [8/128][8_128] \| [8/256][8_256] \| [8/512 (Medium)][8_512] \| [8/768][8_768] \| \| L=10 \| [10/128][10_128] \| [10/256][10_256] \| [10/512][10_512] \| [10/768][10_768] \| \| L=12 \| [12/128][12_128] \| [12/256][12_256] \| [12/512][12_512] \| [12/768 (Base)][12_768] \| Here are scores on the devlopment set of six Chinese tasks: \| Model \| Score \| book_review \| chnsenticorp \| lcqmc \| tnews(CLUE) \| iflytek(CLUE) \| ocnli(CLUE) \| \| -------------- \| :---: \| :----: \| :----------: \| :---: \| :---------: \| :-----------: \| :---------: \| \| RoBERTa-Tiny \| 72.3 \| 83.4 \| 91.4 \| 81.8 \| 62.0 \| 55.0 \| 60.3 \| \| RoBERTa-Mini \| 75.9 \| 85.7 \| 93.7 \| 86.1 \| 63.9 \| 58.3 \| 67.4 \| \| RoBERTa-Small \| 76.9 \| 87.5 \| 93.4 \| 86.5 \| 65.1 \| 59.4 \| 69.7 \| \| RoBERTa-Medium \| 78.0 \| 88.7 \| 94.8 \| 88.1 \| 65.6 \| 59.5 \| 71.2 \| \| RoBERTa-Base \| 79.7 \| 90.1 \| 95.2 \| 89.2 \| 67.0 \| 60.9 \| 75.5 \| For each task, we selected the best fine-tuning hyperparameters from the lists below, and trained with the sequence length of 128: - epochs: 3, 5, 8 - batch sizes: 32, 64 - learning rates: 3e-5, 1e-4, 3e-4 ## How to use You can use this model directly with a pipeline for masked language modeling (take the case of RoBERTa-Medium): ```python >>> from transformers import pipeline >>> unmasker = pipeline('fill-mask', model='uer/chinese_roberta_L-8_H-512') >>> unmasker("中国的首都是[MASK]京。") [ {'sequence': '[CLS] 中国的首都是北京。 [SEP]', 'score': 0.8701988458633423, 'token': 1266, 'token_str': '北'}, {'sequence': '[CLS] 中国的首都是南京。 [SEP]', 'score': 0.1194809079170227, 'token': 1298, 'token_str': '南'}, {'sequence': '[CLS] 中国的首都是东京。 [SEP]', 'score': 0.0037803512532263994, 'token': 691, 'token_str': '东'}, {'sequence': '[CLS] 中国的首都是普京。 [SEP]', 'score': 0.0017127094324678183, 'token': 3249, 'token_str': '普'}, {'sequence': '[CLS] 中国的首都是望京。 [SEP]', 'score': 0.001687526935711503, 'token': 3307, 'token_str': '望'} ] ``` Here is how to use this model to get the features of a given text in PyTorch: ```python from transformers import BertTokenizer, BertModel tokenizer = BertTokenizer.from_pretrained('uer/chinese_roberta_L-8_H-512') model = BertModel.from_pretrained("uer/chinese_roberta_L-8_H-512") text = "用你喜欢的任何文本替换我。" encoded_input = tokenizer(text, return_tensors='pt') output = model(**encoded_input) ``` and in TensorFlow: ```python from transformers import BertTokenizer, TFBertModel tokenizer = BertTokenizer.from_pretrained('uer/chinese_roberta_L-8_H-512') model = TFBertModel.from_pretrained("uer/chinese_roberta_L-8_H-512") text = "用你喜欢的任何文本替换我。" encoded_input = tokenizer(text, return_tensors='tf') output = model(encoded_input) ``` ## Training data [CLUECorpusSmall](https://github.com/CLUEbenchmark/CLUECorpus2020/) is used as training data. We found that models pre-trained on CLUECorpusSmall outperform those pre-trained on CLUECorpus2020, although CLUECorpus2020 is much larger than CLUECorpusSmall. ## Training procedure Models are pre-trained by [UER-py](https://github.com/dbiir/UER-py/) on [Tencent Cloud](https://cloud.tencent.com/). We pre-train 1,000,000 steps with a sequence length of 128 and then pre-train 250,000 additional steps with a sequence length of 512. We use the same hyper-parameters on different model sizes. Taking the case of RoBERTa-Medium Stage1: ``` python3 preprocess.py --corpus_path corpora/cluecorpussmall.txt \ --vocab_path models/google_zh_vocab.txt \ --dataset_path cluecorpussmall_seq128_dataset.pt \ --processes_num 32 --seq_length 128 \ --dynamic_masking --data_processor mlm ``` ``` python3 pretrain.py --dataset_path cluecorpussmall_seq128_dataset.pt \ --vocab_path models/google_zh_vocab.txt \ --config_path models/bert/medium_config.json \ --output_model_path models/cluecorpussmall_roberta_medium_seq128_model.bin \ --world_size 8 --gpu_ranks 0 1 2 3 4 5 6 7 \ --total_steps 1000000 --save_checkpoint_steps 100000 --report_steps 50000 \ --learning_rate 1e-4 --batch_size 64 \ --data_processor mlm --target mlm ``` Stage2: ``` python3 preprocess.py --corpus_path corpora/cluecorpussmall.txt \ --vocab_path models/google_zh_vocab.txt \ --dataset_path cluecorpussmall_seq512_dataset.pt \ --processes_num 32 --seq_length 512 \ --dynamic_masking --data_processor mlm ``` ``` python3 pretrain.py --dataset_path cluecorpussmall_seq512_dataset.pt \ --vocab_path models/google_zh_vocab.txt \ --pretrained_model_path models/cluecorpussmall_roberta_medium_seq128_model.bin-1000000 \ --config_path models/bert/medium_config.json \ --output_model_path models/cluecorpussmall_roberta_medium_seq512_model.bin \ --world_size 8 --gpu_ranks 0 1 2 3 4 5 6 7 \ --total_steps 250000 --save_checkpoint_steps 50000 --report_steps 10000 \ --learning_rate 5e-5 --batch_size 16 \ --data_processor mlm --target mlm ``` Finally, we convert the pre-trained model into Huggingface's format: ``` python3 scripts/convert_bert_from_uer_to_huggingface.py --input_model_path models/cluecorpussmall_roberta_medium_seq512_model.bin-250000 \ --output_model_path pytorch_model.bin \ --layers_num 8 --type mlm ``` ### BibTeX entry and citation info ``` @article{devlin2018bert, title={Bert: Pre-training of deep bidirectional transformers for language understanding}, author={Devlin, Jacob and Chang, Ming-Wei and Lee, Kenton and Toutanova, Kristina}, journal={arXiv preprint arXiv:1810.04805}, year={2018} } @article{liu2019roberta, title={Roberta: A robustly optimized bert pretraining approach}, author={Liu, Yinhan and Ott, Myle and Goyal, Naman and Du, Jingfei and Joshi, Mandar and Chen, Danqi and Levy, Omer and Lewis, Mike and Zettlemoyer, Luke and Stoyanov, Veselin}, journal={arXiv preprint arXiv:1907.11692}, year={2019} } @article{turc2019, title={Well-Read Students Learn Better: On the Importance of Pre-training Compact Models}, author={Turc, Iulia and Chang, Ming-Wei and Lee, Kenton and Toutanova, Kristina}, journal={arXiv preprint arXiv:1908.08962v2 }, year={2019} } @article{zhao2019uer, title={UER: An Open-Source Toolkit for Pre-training Models}, author={Zhao, Zhe and Chen, Hui and Zhang, Jinbin and Zhao, Xin and Liu, Tao and Lu, Wei and Chen, Xi and Deng, Haotang and Ju, Qi and Du, Xiaoyong}, journal={EMNLP-IJCNLP 2019}, pages={241}, year={2019} } @article{zhao2023tencentpretrain, title={TencentPretrain: A Scalable and Flexible Toolkit for Pre-training Models of Different Modalities}, author={Zhao, Zhe and Li, Yudong and Hou, Cheng and Zhao, Jing and others}, journal={ACL 2023}, pages={217}, year={2023} } ``` [2_128]:https://huggingface.co/uer/chinese_roberta_L-2_H-128 [2_256]:https://huggingface.co/uer/chinese_roberta_L-2_H-256 [2_512]:https://huggingface.co/uer/chinese_roberta_L-2_H-512 [2_768]:https://huggingface.co/uer/chinese_roberta_L-2_H-768 [4_128]:https://huggingface.co/uer/chinese_roberta_L-4_H-128 [4_256]:https://huggingface.co/uer/chinese_roberta_L-4_H-256 [4_512]:https://huggingface.co/uer/chinese_roberta_L-4_H-512 [4_768]:https://huggingface.co/uer/chinese_roberta_L-4_H-768 [6_128]:https://huggingface.co/uer/chinese_roberta_L-6_H-128 [6_256]:https://huggingface.co/uer/chinese_roberta_L-6_H-256 [6_512]:https://huggingface.co/uer/chinese_roberta_L-6_H-512 [6_768]:https://huggingface.co/uer/chinese_roberta_L-6_H-768 [8_128]:https://huggingface.co/uer/chinese_roberta_L-8_H-128 [8_256]:https://huggingface.co/uer/chinese_roberta_L-8_H-256 [8_512]:https://huggingface.co/uer/chinese_roberta_L-8_H-512 [8_768]:https://huggingface.co/uer/chinese_roberta_L-8_H-768 [10_128]:https://huggingface.co/uer/chinese_roberta_L-10_H-128 [10_256]:https://huggingface.co/uer/chinese_roberta_L-10_H-256 [10_512]:https://huggingface.co/uer/chinese_roberta_L-10_H-512 [10_768]:https://huggingface.co/uer/chinese_roberta_L-10_H-768 [12_128]:https://huggingface.co/uer/chinese_roberta_L-12_H-128 [12_256]:https://huggingface.co/uer/chinese_roberta_L-12_H-256 [12_512]:https://huggingface.co/uer/chinese_roberta_L-12_H-512 [12_768]:https://huggingface.co/uer/chinese_roberta_L-12_H-768
uer/chinese_roberta_L-8_H-512	uer	2023-08-30T02:28:15Z	146	3	transformers	[ "transformers", "pytorch", "tf", "jax", "bert", "fill-mask", "zh", "dataset:CLUECorpusSmall", "arxiv:1909.05658", "arxiv:2212.06385", "arxiv:1908.08962", "arxiv:1810.04805", "arxiv:1907.11692", "autotrain_compatible", "endpoints_compatible", "region:us" ]	fill-mask	2022-03-02T23:29:05Z	--- language: zh datasets: CLUECorpusSmall widget: - text: "北京是[MASK]国的首都。" --- # Chinese RoBERTa Miniatures ## Model description This is the set of 24 Chinese RoBERTa models pre-trained by [UER-py](https://github.com/dbiir/UER-py/), which is introduced in [this paper](https://arxiv.org/abs/1909.05658). Besides, the models could also be pre-trained by [TencentPretrain](https://github.com/Tencent/TencentPretrain) introduced in [this paper](https://arxiv.org/abs/2212.06385), which inherits UER-py to support models with parameters above one billion, and extends it to a multimodal pre-training framework. [Turc et al.](https://arxiv.org/abs/1908.08962) have shown that the standard BERT recipe is effective on a wide range of model sizes. Following their paper, we released the 24 Chinese RoBERTa models. In order to facilitate users in reproducing the results, we used a publicly available corpus and provided all training details. You can download the 24 Chinese RoBERTa miniatures either from the [UER-py Modelzoo page](https://github.com/dbiir/UER-py/wiki/Modelzoo), or via HuggingFace from the links below: \| \| H=128 \| H=256 \| H=512 \| H=768 \| \| -------- \| :-----------------------: \| :-----------------------: \| :-------------------------: \| :-------------------------: \| \| L=2 \| [2/128 (Tiny)][2_128] \| [2/256][2_256] \| [2/512][2_512] \| [2/768][2_768] \| \| L=4 \| [4/128][4_128] \| [4/256 (Mini)][4_256] \| [4/512 (Small)][4_512] \| [4/768][4_768] \| \| L=6 \| [6/128][6_128] \| [6/256][6_256] \| [6/512][6_512] \| [6/768][6_768] \| \| L=8 \| [8/128][8_128] \| [8/256][8_256] \| [8/512 (Medium)][8_512] \| [8/768][8_768] \| \| L=10 \| [10/128][10_128] \| [10/256][10_256] \| [10/512][10_512] \| [10/768][10_768] \| \| L=12 \| [12/128][12_128] \| [12/256][12_256] \| [12/512][12_512] \| [12/768 (Base)][12_768] \| Here are scores on the devlopment set of six Chinese tasks: \| Model \| Score \| book_review \| chnsenticorp \| lcqmc \| tnews(CLUE) \| iflytek(CLUE) \| ocnli(CLUE) \| \| -------------- \| :---: \| :----: \| :----------: \| :---: \| :---------: \| :-----------: \| :---------: \| \| RoBERTa-Tiny \| 72.3 \| 83.4 \| 91.4 \| 81.8 \| 62.0 \| 55.0 \| 60.3 \| \| RoBERTa-Mini \| 75.9 \| 85.7 \| 93.7 \| 86.1 \| 63.9 \| 58.3 \| 67.4 \| \| RoBERTa-Small \| 76.9 \| 87.5 \| 93.4 \| 86.5 \| 65.1 \| 59.4 \| 69.7 \| \| RoBERTa-Medium \| 78.0 \| 88.7 \| 94.8 \| 88.1 \| 65.6 \| 59.5 \| 71.2 \| \| RoBERTa-Base \| 79.7 \| 90.1 \| 95.2 \| 89.2 \| 67.0 \| 60.9 \| 75.5 \| For each task, we selected the best fine-tuning hyperparameters from the lists below, and trained with the sequence length of 128: - epochs: 3, 5, 8 - batch sizes: 32, 64 - learning rates: 3e-5, 1e-4, 3e-4 ## How to use You can use this model directly with a pipeline for masked language modeling (take the case of RoBERTa-Medium): ```python >>> from transformers import pipeline >>> unmasker = pipeline('fill-mask', model='uer/chinese_roberta_L-8_H-512') >>> unmasker("中国的首都是[MASK]京。") [ {'sequence': '[CLS] 中国的首都是北京。 [SEP]', 'score': 0.8701988458633423, 'token': 1266, 'token_str': '北'}, {'sequence': '[CLS] 中国的首都是南京。 [SEP]', 'score': 0.1194809079170227, 'token': 1298, 'token_str': '南'}, {'sequence': '[CLS] 中国的首都是东京。 [SEP]', 'score': 0.0037803512532263994, 'token': 691, 'token_str': '东'}, {'sequence': '[CLS] 中国的首都是普京。 [SEP]', 'score': 0.0017127094324678183, 'token': 3249, 'token_str': '普'}, {'sequence': '[CLS] 中国的首都是望京。 [SEP]', 'score': 0.001687526935711503, 'token': 3307, 'token_str': '望'} ] ``` Here is how to use this model to get the features of a given text in PyTorch: ```python from transformers import BertTokenizer, BertModel tokenizer = BertTokenizer.from_pretrained('uer/chinese_roberta_L-8_H-512') model = BertModel.from_pretrained("uer/chinese_roberta_L-8_H-512") text = "用你喜欢的任何文本替换我。" encoded_input = tokenizer(text, return_tensors='pt') output = model(**encoded_input) ``` and in TensorFlow: ```python from transformers import BertTokenizer, TFBertModel tokenizer = BertTokenizer.from_pretrained('uer/chinese_roberta_L-8_H-512') model = TFBertModel.from_pretrained("uer/chinese_roberta_L-8_H-512") text = "用你喜欢的任何文本替换我。" encoded_input = tokenizer(text, return_tensors='tf') output = model(encoded_input) ``` ## Training data [CLUECorpusSmall](https://github.com/CLUEbenchmark/CLUECorpus2020/) is used as training data. We found that models pre-trained on CLUECorpusSmall outperform those pre-trained on CLUECorpus2020, although CLUECorpus2020 is much larger than CLUECorpusSmall. ## Training procedure Models are pre-trained by [UER-py](https://github.com/dbiir/UER-py/) on [Tencent Cloud](https://cloud.tencent.com/). We pre-train 1,000,000 steps with a sequence length of 128 and then pre-train 250,000 additional steps with a sequence length of 512. We use the same hyper-parameters on different model sizes. Taking the case of RoBERTa-Medium Stage1: ``` python3 preprocess.py --corpus_path corpora/cluecorpussmall.txt \ --vocab_path models/google_zh_vocab.txt \ --dataset_path cluecorpussmall_seq128_dataset.pt \ --processes_num 32 --seq_length 128 \ --dynamic_masking --data_processor mlm ``` ``` python3 pretrain.py --dataset_path cluecorpussmall_seq128_dataset.pt \ --vocab_path models/google_zh_vocab.txt \ --config_path models/bert/medium_config.json \ --output_model_path models/cluecorpussmall_roberta_medium_seq128_model.bin \ --world_size 8 --gpu_ranks 0 1 2 3 4 5 6 7 \ --total_steps 1000000 --save_checkpoint_steps 100000 --report_steps 50000 \ --learning_rate 1e-4 --batch_size 64 \ --data_processor mlm --target mlm ``` Stage2: ``` python3 preprocess.py --corpus_path corpora/cluecorpussmall.txt \ --vocab_path models/google_zh_vocab.txt \ --dataset_path cluecorpussmall_seq512_dataset.pt \ --processes_num 32 --seq_length 512 \ --dynamic_masking --data_processor mlm ``` ``` python3 pretrain.py --dataset_path cluecorpussmall_seq512_dataset.pt \ --vocab_path models/google_zh_vocab.txt \ --pretrained_model_path models/cluecorpussmall_roberta_medium_seq128_model.bin-1000000 \ --config_path models/bert/medium_config.json \ --output_model_path models/cluecorpussmall_roberta_medium_seq512_model.bin \ --world_size 8 --gpu_ranks 0 1 2 3 4 5 6 7 \ --total_steps 250000 --save_checkpoint_steps 50000 --report_steps 10000 \ --learning_rate 5e-5 --batch_size 16 \ --data_processor mlm --target mlm ``` Finally, we convert the pre-trained model into Huggingface's format: ``` python3 scripts/convert_bert_from_uer_to_huggingface.py --input_model_path models/cluecorpussmall_roberta_medium_seq512_model.bin-250000 \ --output_model_path pytorch_model.bin \ --layers_num 8 --type mlm ``` ### BibTeX entry and citation info ``` @article{devlin2018bert, title={Bert: Pre-training of deep bidirectional transformers for language understanding}, author={Devlin, Jacob and Chang, Ming-Wei and Lee, Kenton and Toutanova, Kristina}, journal={arXiv preprint arXiv:1810.04805}, year={2018} } @article{liu2019roberta, title={Roberta: A robustly optimized bert pretraining approach}, author={Liu, Yinhan and Ott, Myle and Goyal, Naman and Du, Jingfei and Joshi, Mandar and Chen, Danqi and Levy, Omer and Lewis, Mike and Zettlemoyer, Luke and Stoyanov, Veselin}, journal={arXiv preprint arXiv:1907.11692}, year={2019} } @article{turc2019, title={Well-Read Students Learn Better: On the Importance of Pre-training Compact Models}, author={Turc, Iulia and Chang, Ming-Wei and Lee, Kenton and Toutanova, Kristina}, journal={arXiv preprint arXiv:1908.08962v2 }, year={2019} } @article{zhao2019uer, title={UER: An Open-Source Toolkit for Pre-training Models}, author={Zhao, Zhe and Chen, Hui and Zhang, Jinbin and Zhao, Xin and Liu, Tao and Lu, Wei and Chen, Xi and Deng, Haotang and Ju, Qi and Du, Xiaoyong}, journal={EMNLP-IJCNLP 2019}, pages={241}, year={2019} } @article{zhao2023tencentpretrain, title={TencentPretrain: A Scalable and Flexible Toolkit for Pre-training Models of Different Modalities}, author={Zhao, Zhe and Li, Yudong and Hou, Cheng and Zhao, Jing and others}, journal={ACL 2023}, pages={217}, year={2023} } ``` [2_128]:https://huggingface.co/uer/chinese_roberta_L-2_H-128 [2_256]:https://huggingface.co/uer/chinese_roberta_L-2_H-256 [2_512]:https://huggingface.co/uer/chinese_roberta_L-2_H-512 [2_768]:https://huggingface.co/uer/chinese_roberta_L-2_H-768 [4_128]:https://huggingface.co/uer/chinese_roberta_L-4_H-128 [4_256]:https://huggingface.co/uer/chinese_roberta_L-4_H-256 [4_512]:https://huggingface.co/uer/chinese_roberta_L-4_H-512 [4_768]:https://huggingface.co/uer/chinese_roberta_L-4_H-768 [6_128]:https://huggingface.co/uer/chinese_roberta_L-6_H-128 [6_256]:https://huggingface.co/uer/chinese_roberta_L-6_H-256 [6_512]:https://huggingface.co/uer/chinese_roberta_L-6_H-512 [6_768]:https://huggingface.co/uer/chinese_roberta_L-6_H-768 [8_128]:https://huggingface.co/uer/chinese_roberta_L-8_H-128 [8_256]:https://huggingface.co/uer/chinese_roberta_L-8_H-256 [8_512]:https://huggingface.co/uer/chinese_roberta_L-8_H-512 [8_768]:https://huggingface.co/uer/chinese_roberta_L-8_H-768 [10_128]:https://huggingface.co/uer/chinese_roberta_L-10_H-128 [10_256]:https://huggingface.co/uer/chinese_roberta_L-10_H-256 [10_512]:https://huggingface.co/uer/chinese_roberta_L-10_H-512 [10_768]:https://huggingface.co/uer/chinese_roberta_L-10_H-768 [12_128]:https://huggingface.co/uer/chinese_roberta_L-12_H-128 [12_256]:https://huggingface.co/uer/chinese_roberta_L-12_H-256 [12_512]:https://huggingface.co/uer/chinese_roberta_L-12_H-512 [12_768]:https://huggingface.co/uer/chinese_roberta_L-12_H-768
fuguizhen/q-FrozenLake-v1-4x4-noSlippery	fuguizhen	2023-08-30T02:26:51Z	0	0	null	[ "FrozenLake-v1-4x4-no_slippery", "q-learning", "reinforcement-learning", "custom-implementation", "model-index", "region:us" ]	reinforcement-learning	2023-08-30T02:26:49Z	--- tags: - FrozenLake-v1-4x4-no_slippery - q-learning - reinforcement-learning - custom-implementation model-index: - name: q-FrozenLake-v1-4x4-noSlippery results: - task: type: reinforcement-learning name: reinforcement-learning dataset: name: FrozenLake-v1-4x4-no_slippery type: FrozenLake-v1-4x4-no_slippery metrics: - type: mean_reward value: 1.00 +/- 0.00 name: mean_reward verified: false --- # Q-Learning Agent playing1 FrozenLake-v1 This is a trained model of a Q-Learning agent playing FrozenLake-v1 . ## Usage model = load_from_hub(repo_id="fuguizhen/q-FrozenLake-v1-4x4-noSlippery", filename="q-learning.pkl") # Don't forget to check if you need to add additional attributes (is_slippery=False etc) env = gym.make(model["env_id"])
uer/chinese_roberta_L-6_H-256	uer	2023-08-30T02:26:30Z	108	1	transformers	[ "transformers", "pytorch", "tf", "jax", "bert", "fill-mask", "zh", "dataset:CLUECorpusSmall", "arxiv:1909.05658", "arxiv:2212.06385", "arxiv:1908.08962", "arxiv:1810.04805", "arxiv:1907.11692", "autotrain_compatible", "endpoints_compatible", "region:us" ]	fill-mask	2022-03-02T23:29:05Z	--- language: zh datasets: CLUECorpusSmall widget: - text: "北京是[MASK]国的首都。" --- # Chinese RoBERTa Miniatures ## Model description This is the set of 24 Chinese RoBERTa models pre-trained by [UER-py](https://github.com/dbiir/UER-py/), which is introduced in [this paper](https://arxiv.org/abs/1909.05658). Besides, the models could also be pre-trained by [TencentPretrain](https://github.com/Tencent/TencentPretrain) introduced in [this paper](https://arxiv.org/abs/2212.06385), which inherits UER-py to support models with parameters above one billion, and extends it to a multimodal pre-training framework. [Turc et al.](https://arxiv.org/abs/1908.08962) have shown that the standard BERT recipe is effective on a wide range of model sizes. Following their paper, we released the 24 Chinese RoBERTa models. In order to facilitate users in reproducing the results, we used a publicly available corpus and provided all training details. You can download the 24 Chinese RoBERTa miniatures either from the [UER-py Modelzoo page](https://github.com/dbiir/UER-py/wiki/Modelzoo), or via HuggingFace from the links below: \| \| H=128 \| H=256 \| H=512 \| H=768 \| \| -------- \| :-----------------------: \| :-----------------------: \| :-------------------------: \| :-------------------------: \| \| L=2 \| [2/128 (Tiny)][2_128] \| [2/256][2_256] \| [2/512][2_512] \| [2/768][2_768] \| \| L=4 \| [4/128][4_128] \| [4/256 (Mini)][4_256] \| [4/512 (Small)][4_512] \| [4/768][4_768] \| \| L=6 \| [6/128][6_128] \| [6/256][6_256] \| [6/512][6_512] \| [6/768][6_768] \| \| L=8 \| [8/128][8_128] \| [8/256][8_256] \| [8/512 (Medium)][8_512] \| [8/768][8_768] \| \| L=10 \| [10/128][10_128] \| [10/256][10_256] \| [10/512][10_512] \| [10/768][10_768] \| \| L=12 \| [12/128][12_128] \| [12/256][12_256] \| [12/512][12_512] \| [12/768 (Base)][12_768] \| Here are scores on the devlopment set of six Chinese tasks: \| Model \| Score \| book_review \| chnsenticorp \| lcqmc \| tnews(CLUE) \| iflytek(CLUE) \| ocnli(CLUE) \| \| -------------- \| :---: \| :----: \| :----------: \| :---: \| :---------: \| :-----------: \| :---------: \| \| RoBERTa-Tiny \| 72.3 \| 83.4 \| 91.4 \| 81.8 \| 62.0 \| 55.0 \| 60.3 \| \| RoBERTa-Mini \| 75.9 \| 85.7 \| 93.7 \| 86.1 \| 63.9 \| 58.3 \| 67.4 \| \| RoBERTa-Small \| 76.9 \| 87.5 \| 93.4 \| 86.5 \| 65.1 \| 59.4 \| 69.7 \| \| RoBERTa-Medium \| 78.0 \| 88.7 \| 94.8 \| 88.1 \| 65.6 \| 59.5 \| 71.2 \| \| RoBERTa-Base \| 79.7 \| 90.1 \| 95.2 \| 89.2 \| 67.0 \| 60.9 \| 75.5 \| For each task, we selected the best fine-tuning hyperparameters from the lists below, and trained with the sequence length of 128: - epochs: 3, 5, 8 - batch sizes: 32, 64 - learning rates: 3e-5, 1e-4, 3e-4 ## How to use You can use this model directly with a pipeline for masked language modeling (take the case of RoBERTa-Medium): ```python >>> from transformers import pipeline >>> unmasker = pipeline('fill-mask', model='uer/chinese_roberta_L-8_H-512') >>> unmasker("中国的首都是[MASK]京。") [ {'sequence': '[CLS] 中国的首都是北京。 [SEP]', 'score': 0.8701988458633423, 'token': 1266, 'token_str': '北'}, {'sequence': '[CLS] 中国的首都是南京。 [SEP]', 'score': 0.1194809079170227, 'token': 1298, 'token_str': '南'}, {'sequence': '[CLS] 中国的首都是东京。 [SEP]', 'score': 0.0037803512532263994, 'token': 691, 'token_str': '东'}, {'sequence': '[CLS] 中国的首都是普京。 [SEP]', 'score': 0.0017127094324678183, 'token': 3249, 'token_str': '普'}, {'sequence': '[CLS] 中国的首都是望京。 [SEP]', 'score': 0.001687526935711503, 'token': 3307, 'token_str': '望'} ] ``` Here is how to use this model to get the features of a given text in PyTorch: ```python from transformers import BertTokenizer, BertModel tokenizer = BertTokenizer.from_pretrained('uer/chinese_roberta_L-8_H-512') model = BertModel.from_pretrained("uer/chinese_roberta_L-8_H-512") text = "用你喜欢的任何文本替换我。" encoded_input = tokenizer(text, return_tensors='pt') output = model(**encoded_input) ``` and in TensorFlow: ```python from transformers import BertTokenizer, TFBertModel tokenizer = BertTokenizer.from_pretrained('uer/chinese_roberta_L-8_H-512') model = TFBertModel.from_pretrained("uer/chinese_roberta_L-8_H-512") text = "用你喜欢的任何文本替换我。" encoded_input = tokenizer(text, return_tensors='tf') output = model(encoded_input) ``` ## Training data [CLUECorpusSmall](https://github.com/CLUEbenchmark/CLUECorpus2020/) is used as training data. We found that models pre-trained on CLUECorpusSmall outperform those pre-trained on CLUECorpus2020, although CLUECorpus2020 is much larger than CLUECorpusSmall. ## Training procedure Models are pre-trained by [UER-py](https://github.com/dbiir/UER-py/) on [Tencent Cloud](https://cloud.tencent.com/). We pre-train 1,000,000 steps with a sequence length of 128 and then pre-train 250,000 additional steps with a sequence length of 512. We use the same hyper-parameters on different model sizes. Taking the case of RoBERTa-Medium Stage1: ``` python3 preprocess.py --corpus_path corpora/cluecorpussmall.txt \ --vocab_path models/google_zh_vocab.txt \ --dataset_path cluecorpussmall_seq128_dataset.pt \ --processes_num 32 --seq_length 128 \ --dynamic_masking --data_processor mlm ``` ``` python3 pretrain.py --dataset_path cluecorpussmall_seq128_dataset.pt \ --vocab_path models/google_zh_vocab.txt \ --config_path models/bert/medium_config.json \ --output_model_path models/cluecorpussmall_roberta_medium_seq128_model.bin \ --world_size 8 --gpu_ranks 0 1 2 3 4 5 6 7 \ --total_steps 1000000 --save_checkpoint_steps 100000 --report_steps 50000 \ --learning_rate 1e-4 --batch_size 64 \ --data_processor mlm --target mlm ``` Stage2: ``` python3 preprocess.py --corpus_path corpora/cluecorpussmall.txt \ --vocab_path models/google_zh_vocab.txt \ --dataset_path cluecorpussmall_seq512_dataset.pt \ --processes_num 32 --seq_length 512 \ --dynamic_masking --data_processor mlm ``` ``` python3 pretrain.py --dataset_path cluecorpussmall_seq512_dataset.pt \ --vocab_path models/google_zh_vocab.txt \ --pretrained_model_path models/cluecorpussmall_roberta_medium_seq128_model.bin-1000000 \ --config_path models/bert/medium_config.json \ --output_model_path models/cluecorpussmall_roberta_medium_seq512_model.bin \ --world_size 8 --gpu_ranks 0 1 2 3 4 5 6 7 \ --total_steps 250000 --save_checkpoint_steps 50000 --report_steps 10000 \ --learning_rate 5e-5 --batch_size 16 \ --data_processor mlm --target mlm ``` Finally, we convert the pre-trained model into Huggingface's format: ``` python3 scripts/convert_bert_from_uer_to_huggingface.py --input_model_path models/cluecorpussmall_roberta_medium_seq512_model.bin-250000 \ --output_model_path pytorch_model.bin \ --layers_num 8 --type mlm ``` ### BibTeX entry and citation info ``` @article{devlin2018bert, title={Bert: Pre-training of deep bidirectional transformers for language understanding}, author={Devlin, Jacob and Chang, Ming-Wei and Lee, Kenton and Toutanova, Kristina}, journal={arXiv preprint arXiv:1810.04805}, year={2018} } @article{liu2019roberta, title={Roberta: A robustly optimized bert pretraining approach}, author={Liu, Yinhan and Ott, Myle and Goyal, Naman and Du, Jingfei and Joshi, Mandar and Chen, Danqi and Levy, Omer and Lewis, Mike and Zettlemoyer, Luke and Stoyanov, Veselin}, journal={arXiv preprint arXiv:1907.11692}, year={2019} } @article{turc2019, title={Well-Read Students Learn Better: On the Importance of Pre-training Compact Models}, author={Turc, Iulia and Chang, Ming-Wei and Lee, Kenton and Toutanova, Kristina}, journal={arXiv preprint arXiv:1908.08962v2 }, year={2019} } @article{zhao2019uer, title={UER: An Open-Source Toolkit for Pre-training Models}, author={Zhao, Zhe and Chen, Hui and Zhang, Jinbin and Zhao, Xin and Liu, Tao and Lu, Wei and Chen, Xi and Deng, Haotang and Ju, Qi and Du, Xiaoyong}, journal={EMNLP-IJCNLP 2019}, pages={241}, year={2019} } @article{zhao2023tencentpretrain, title={TencentPretrain: A Scalable and Flexible Toolkit for Pre-training Models of Different Modalities}, author={Zhao, Zhe and Li, Yudong and Hou, Cheng and Zhao, Jing and others}, journal={ACL 2023}, pages={217}, year={2023} } ``` [2_128]:https://huggingface.co/uer/chinese_roberta_L-2_H-128 [2_256]:https://huggingface.co/uer/chinese_roberta_L-2_H-256 [2_512]:https://huggingface.co/uer/chinese_roberta_L-2_H-512 [2_768]:https://huggingface.co/uer/chinese_roberta_L-2_H-768 [4_128]:https://huggingface.co/uer/chinese_roberta_L-4_H-128 [4_256]:https://huggingface.co/uer/chinese_roberta_L-4_H-256 [4_512]:https://huggingface.co/uer/chinese_roberta_L-4_H-512 [4_768]:https://huggingface.co/uer/chinese_roberta_L-4_H-768 [6_128]:https://huggingface.co/uer/chinese_roberta_L-6_H-128 [6_256]:https://huggingface.co/uer/chinese_roberta_L-6_H-256 [6_512]:https://huggingface.co/uer/chinese_roberta_L-6_H-512 [6_768]:https://huggingface.co/uer/chinese_roberta_L-6_H-768 [8_128]:https://huggingface.co/uer/chinese_roberta_L-8_H-128 [8_256]:https://huggingface.co/uer/chinese_roberta_L-8_H-256 [8_512]:https://huggingface.co/uer/chinese_roberta_L-8_H-512 [8_768]:https://huggingface.co/uer/chinese_roberta_L-8_H-768 [10_128]:https://huggingface.co/uer/chinese_roberta_L-10_H-128 [10_256]:https://huggingface.co/uer/chinese_roberta_L-10_H-256 [10_512]:https://huggingface.co/uer/chinese_roberta_L-10_H-512 [10_768]:https://huggingface.co/uer/chinese_roberta_L-10_H-768 [12_128]:https://huggingface.co/uer/chinese_roberta_L-12_H-128 [12_256]:https://huggingface.co/uer/chinese_roberta_L-12_H-256 [12_512]:https://huggingface.co/uer/chinese_roberta_L-12_H-512 [12_768]:https://huggingface.co/uer/chinese_roberta_L-12_H-768
uer/chinese_roberta_L-6_H-128	uer	2023-08-30T02:26:10Z	119	0	transformers	[ "transformers", "pytorch", "tf", "jax", "bert", "fill-mask", "zh", "dataset:CLUECorpusSmall", "arxiv:1909.05658", "arxiv:2212.06385", "arxiv:1908.08962", "arxiv:1810.04805", "arxiv:1907.11692", "autotrain_compatible", "endpoints_compatible", "region:us" ]	fill-mask	2022-03-02T23:29:05Z	--- language: zh datasets: CLUECorpusSmall widget: - text: "北京是[MASK]国的首都。" --- # Chinese RoBERTa Miniatures ## Model description This is the set of 24 Chinese RoBERTa models pre-trained by [UER-py](https://github.com/dbiir/UER-py/), which is introduced in [this paper](https://arxiv.org/abs/1909.05658). Besides, the models could also be pre-trained by [TencentPretrain](https://github.com/Tencent/TencentPretrain) introduced in [this paper](https://arxiv.org/abs/2212.06385), which inherits UER-py to support models with parameters above one billion, and extends it to a multimodal pre-training framework. [Turc et al.](https://arxiv.org/abs/1908.08962) have shown that the standard BERT recipe is effective on a wide range of model sizes. Following their paper, we released the 24 Chinese RoBERTa models. In order to facilitate users in reproducing the results, we used a publicly available corpus and provided all training details. You can download the 24 Chinese RoBERTa miniatures either from the [UER-py Modelzoo page](https://github.com/dbiir/UER-py/wiki/Modelzoo), or via HuggingFace from the links below: \| \| H=128 \| H=256 \| H=512 \| H=768 \| \| -------- \| :-----------------------: \| :-----------------------: \| :-------------------------: \| :-------------------------: \| \| L=2 \| [2/128 (Tiny)][2_128] \| [2/256][2_256] \| [2/512][2_512] \| [2/768][2_768] \| \| L=4 \| [4/128][4_128] \| [4/256 (Mini)][4_256] \| [4/512 (Small)][4_512] \| [4/768][4_768] \| \| L=6 \| [6/128][6_128] \| [6/256][6_256] \| [6/512][6_512] \| [6/768][6_768] \| \| L=8 \| [8/128][8_128] \| [8/256][8_256] \| [8/512 (Medium)][8_512] \| [8/768][8_768] \| \| L=10 \| [10/128][10_128] \| [10/256][10_256] \| [10/512][10_512] \| [10/768][10_768] \| \| L=12 \| [12/128][12_128] \| [12/256][12_256] \| [12/512][12_512] \| [12/768 (Base)][12_768] \| Here are scores on the devlopment set of six Chinese tasks: \| Model \| Score \| book_review \| chnsenticorp \| lcqmc \| tnews(CLUE) \| iflytek(CLUE) \| ocnli(CLUE) \| \| -------------- \| :---: \| :----: \| :----------: \| :---: \| :---------: \| :-----------: \| :---------: \| \| RoBERTa-Tiny \| 72.3 \| 83.4 \| 91.4 \| 81.8 \| 62.0 \| 55.0 \| 60.3 \| \| RoBERTa-Mini \| 75.9 \| 85.7 \| 93.7 \| 86.1 \| 63.9 \| 58.3 \| 67.4 \| \| RoBERTa-Small \| 76.9 \| 87.5 \| 93.4 \| 86.5 \| 65.1 \| 59.4 \| 69.7 \| \| RoBERTa-Medium \| 78.0 \| 88.7 \| 94.8 \| 88.1 \| 65.6 \| 59.5 \| 71.2 \| \| RoBERTa-Base \| 79.7 \| 90.1 \| 95.2 \| 89.2 \| 67.0 \| 60.9 \| 75.5 \| For each task, we selected the best fine-tuning hyperparameters from the lists below, and trained with the sequence length of 128: - epochs: 3, 5, 8 - batch sizes: 32, 64 - learning rates: 3e-5, 1e-4, 3e-4 ## How to use You can use this model directly with a pipeline for masked language modeling (take the case of RoBERTa-Medium): ```python >>> from transformers import pipeline >>> unmasker = pipeline('fill-mask', model='uer/chinese_roberta_L-8_H-512') >>> unmasker("中国的首都是[MASK]京。") [ {'sequence': '[CLS] 中国的首都是北京。 [SEP]', 'score': 0.8701988458633423, 'token': 1266, 'token_str': '北'}, {'sequence': '[CLS] 中国的首都是南京。 [SEP]', 'score': 0.1194809079170227, 'token': 1298, 'token_str': '南'}, {'sequence': '[CLS] 中国的首都是东京。 [SEP]', 'score': 0.0037803512532263994, 'token': 691, 'token_str': '东'}, {'sequence': '[CLS] 中国的首都是普京。 [SEP]', 'score': 0.0017127094324678183, 'token': 3249, 'token_str': '普'}, {'sequence': '[CLS] 中国的首都是望京。 [SEP]', 'score': 0.001687526935711503, 'token': 3307, 'token_str': '望'} ] ``` Here is how to use this model to get the features of a given text in PyTorch: ```python from transformers import BertTokenizer, BertModel tokenizer = BertTokenizer.from_pretrained('uer/chinese_roberta_L-8_H-512') model = BertModel.from_pretrained("uer/chinese_roberta_L-8_H-512") text = "用你喜欢的任何文本替换我。" encoded_input = tokenizer(text, return_tensors='pt') output = model(**encoded_input) ``` and in TensorFlow: ```python from transformers import BertTokenizer, TFBertModel tokenizer = BertTokenizer.from_pretrained('uer/chinese_roberta_L-8_H-512') model = TFBertModel.from_pretrained("uer/chinese_roberta_L-8_H-512") text = "用你喜欢的任何文本替换我。" encoded_input = tokenizer(text, return_tensors='tf') output = model(encoded_input) ``` ## Training data [CLUECorpusSmall](https://github.com/CLUEbenchmark/CLUECorpus2020/) is used as training data. We found that models pre-trained on CLUECorpusSmall outperform those pre-trained on CLUECorpus2020, although CLUECorpus2020 is much larger than CLUECorpusSmall. ## Training procedure Models are pre-trained by [UER-py](https://github.com/dbiir/UER-py/) on [Tencent Cloud](https://cloud.tencent.com/). We pre-train 1,000,000 steps with a sequence length of 128 and then pre-train 250,000 additional steps with a sequence length of 512. We use the same hyper-parameters on different model sizes. Taking the case of RoBERTa-Medium Stage1: ``` python3 preprocess.py --corpus_path corpora/cluecorpussmall.txt \ --vocab_path models/google_zh_vocab.txt \ --dataset_path cluecorpussmall_seq128_dataset.pt \ --processes_num 32 --seq_length 128 \ --dynamic_masking --data_processor mlm ``` ``` python3 pretrain.py --dataset_path cluecorpussmall_seq128_dataset.pt \ --vocab_path models/google_zh_vocab.txt \ --config_path models/bert/medium_config.json \ --output_model_path models/cluecorpussmall_roberta_medium_seq128_model.bin \ --world_size 8 --gpu_ranks 0 1 2 3 4 5 6 7 \ --total_steps 1000000 --save_checkpoint_steps 100000 --report_steps 50000 \ --learning_rate 1e-4 --batch_size 64 \ --data_processor mlm --target mlm ``` Stage2: ``` python3 preprocess.py --corpus_path corpora/cluecorpussmall.txt \ --vocab_path models/google_zh_vocab.txt \ --dataset_path cluecorpussmall_seq512_dataset.pt \ --processes_num 32 --seq_length 512 \ --dynamic_masking --data_processor mlm ``` ``` python3 pretrain.py --dataset_path cluecorpussmall_seq512_dataset.pt \ --vocab_path models/google_zh_vocab.txt \ --pretrained_model_path models/cluecorpussmall_roberta_medium_seq128_model.bin-1000000 \ --config_path models/bert/medium_config.json \ --output_model_path models/cluecorpussmall_roberta_medium_seq512_model.bin \ --world_size 8 --gpu_ranks 0 1 2 3 4 5 6 7 \ --total_steps 250000 --save_checkpoint_steps 50000 --report_steps 10000 \ --learning_rate 5e-5 --batch_size 16 \ --data_processor mlm --target mlm ``` Finally, we convert the pre-trained model into Huggingface's format: ``` python3 scripts/convert_bert_from_uer_to_huggingface.py --input_model_path models/cluecorpussmall_roberta_medium_seq512_model.bin-250000 \ --output_model_path pytorch_model.bin \ --layers_num 8 --type mlm ``` ### BibTeX entry and citation info ``` @article{devlin2018bert, title={Bert: Pre-training of deep bidirectional transformers for language understanding}, author={Devlin, Jacob and Chang, Ming-Wei and Lee, Kenton and Toutanova, Kristina}, journal={arXiv preprint arXiv:1810.04805}, year={2018} } @article{liu2019roberta, title={Roberta: A robustly optimized bert pretraining approach}, author={Liu, Yinhan and Ott, Myle and Goyal, Naman and Du, Jingfei and Joshi, Mandar and Chen, Danqi and Levy, Omer and Lewis, Mike and Zettlemoyer, Luke and Stoyanov, Veselin}, journal={arXiv preprint arXiv:1907.11692}, year={2019} } @article{turc2019, title={Well-Read Students Learn Better: On the Importance of Pre-training Compact Models}, author={Turc, Iulia and Chang, Ming-Wei and Lee, Kenton and Toutanova, Kristina}, journal={arXiv preprint arXiv:1908.08962v2 }, year={2019} } @article{zhao2019uer, title={UER: An Open-Source Toolkit for Pre-training Models}, author={Zhao, Zhe and Chen, Hui and Zhang, Jinbin and Zhao, Xin and Liu, Tao and Lu, Wei and Chen, Xi and Deng, Haotang and Ju, Qi and Du, Xiaoyong}, journal={EMNLP-IJCNLP 2019}, pages={241}, year={2019} } @article{zhao2023tencentpretrain, title={TencentPretrain: A Scalable and Flexible Toolkit for Pre-training Models of Different Modalities}, author={Zhao, Zhe and Li, Yudong and Hou, Cheng and Zhao, Jing and others}, journal={ACL 2023}, pages={217}, year={2023} } ``` [2_128]:https://huggingface.co/uer/chinese_roberta_L-2_H-128 [2_256]:https://huggingface.co/uer/chinese_roberta_L-2_H-256 [2_512]:https://huggingface.co/uer/chinese_roberta_L-2_H-512 [2_768]:https://huggingface.co/uer/chinese_roberta_L-2_H-768 [4_128]:https://huggingface.co/uer/chinese_roberta_L-4_H-128 [4_256]:https://huggingface.co/uer/chinese_roberta_L-4_H-256 [4_512]:https://huggingface.co/uer/chinese_roberta_L-4_H-512 [4_768]:https://huggingface.co/uer/chinese_roberta_L-4_H-768 [6_128]:https://huggingface.co/uer/chinese_roberta_L-6_H-128 [6_256]:https://huggingface.co/uer/chinese_roberta_L-6_H-256 [6_512]:https://huggingface.co/uer/chinese_roberta_L-6_H-512 [6_768]:https://huggingface.co/uer/chinese_roberta_L-6_H-768 [8_128]:https://huggingface.co/uer/chinese_roberta_L-8_H-128 [8_256]:https://huggingface.co/uer/chinese_roberta_L-8_H-256 [8_512]:https://huggingface.co/uer/chinese_roberta_L-8_H-512 [8_768]:https://huggingface.co/uer/chinese_roberta_L-8_H-768 [10_128]:https://huggingface.co/uer/chinese_roberta_L-10_H-128 [10_256]:https://huggingface.co/uer/chinese_roberta_L-10_H-256 [10_512]:https://huggingface.co/uer/chinese_roberta_L-10_H-512 [10_768]:https://huggingface.co/uer/chinese_roberta_L-10_H-768 [12_128]:https://huggingface.co/uer/chinese_roberta_L-12_H-128 [12_256]:https://huggingface.co/uer/chinese_roberta_L-12_H-256 [12_512]:https://huggingface.co/uer/chinese_roberta_L-12_H-512 [12_768]:https://huggingface.co/uer/chinese_roberta_L-12_H-768
uer/chinese_roberta_L-4_H-768	uer	2023-08-30T02:25:50Z	109	0	transformers	[ "transformers", "pytorch", "tf", "jax", "bert", "fill-mask", "zh", "dataset:CLUECorpusSmall", "arxiv:1909.05658", "arxiv:2212.06385", "arxiv:1908.08962", "arxiv:1810.04805", "arxiv:1907.11692", "autotrain_compatible", "endpoints_compatible", "region:us" ]	fill-mask	2022-03-02T23:29:05Z	--- language: zh datasets: CLUECorpusSmall widget: - text: "北京是[MASK]国的首都。" --- # Chinese RoBERTa Miniatures ## Model description This is the set of 24 Chinese RoBERTa models pre-trained by [UER-py](https://github.com/dbiir/UER-py/), which is introduced in [this paper](https://arxiv.org/abs/1909.05658). Besides, the models could also be pre-trained by [TencentPretrain](https://github.com/Tencent/TencentPretrain) introduced in [this paper](https://arxiv.org/abs/2212.06385), which inherits UER-py to support models with parameters above one billion, and extends it to a multimodal pre-training framework. [Turc et al.](https://arxiv.org/abs/1908.08962) have shown that the standard BERT recipe is effective on a wide range of model sizes. Following their paper, we released the 24 Chinese RoBERTa models. In order to facilitate users in reproducing the results, we used a publicly available corpus and provided all training details. You can download the 24 Chinese RoBERTa miniatures either from the [UER-py Modelzoo page](https://github.com/dbiir/UER-py/wiki/Modelzoo), or via HuggingFace from the links below: \| \| H=128 \| H=256 \| H=512 \| H=768 \| \| -------- \| :-----------------------: \| :-----------------------: \| :-------------------------: \| :-------------------------: \| \| L=2 \| [2/128 (Tiny)][2_128] \| [2/256][2_256] \| [2/512][2_512] \| [2/768][2_768] \| \| L=4 \| [4/128][4_128] \| [4/256 (Mini)][4_256] \| [4/512 (Small)][4_512] \| [4/768][4_768] \| \| L=6 \| [6/128][6_128] \| [6/256][6_256] \| [6/512][6_512] \| [6/768][6_768] \| \| L=8 \| [8/128][8_128] \| [8/256][8_256] \| [8/512 (Medium)][8_512] \| [8/768][8_768] \| \| L=10 \| [10/128][10_128] \| [10/256][10_256] \| [10/512][10_512] \| [10/768][10_768] \| \| L=12 \| [12/128][12_128] \| [12/256][12_256] \| [12/512][12_512] \| [12/768 (Base)][12_768] \| Here are scores on the devlopment set of six Chinese tasks: \| Model \| Score \| book_review \| chnsenticorp \| lcqmc \| tnews(CLUE) \| iflytek(CLUE) \| ocnli(CLUE) \| \| -------------- \| :---: \| :----: \| :----------: \| :---: \| :---------: \| :-----------: \| :---------: \| \| RoBERTa-Tiny \| 72.3 \| 83.4 \| 91.4 \| 81.8 \| 62.0 \| 55.0 \| 60.3 \| \| RoBERTa-Mini \| 75.9 \| 85.7 \| 93.7 \| 86.1 \| 63.9 \| 58.3 \| 67.4 \| \| RoBERTa-Small \| 76.9 \| 87.5 \| 93.4 \| 86.5 \| 65.1 \| 59.4 \| 69.7 \| \| RoBERTa-Medium \| 78.0 \| 88.7 \| 94.8 \| 88.1 \| 65.6 \| 59.5 \| 71.2 \| \| RoBERTa-Base \| 79.7 \| 90.1 \| 95.2 \| 89.2 \| 67.0 \| 60.9 \| 75.5 \| For each task, we selected the best fine-tuning hyperparameters from the lists below, and trained with the sequence length of 128: - epochs: 3, 5, 8 - batch sizes: 32, 64 - learning rates: 3e-5, 1e-4, 3e-4 ## How to use You can use this model directly with a pipeline for masked language modeling (take the case of RoBERTa-Medium): ```python >>> from transformers import pipeline >>> unmasker = pipeline('fill-mask', model='uer/chinese_roberta_L-8_H-512') >>> unmasker("中国的首都是[MASK]京。") [ {'sequence': '[CLS] 中国的首都是北京。 [SEP]', 'score': 0.8701988458633423, 'token': 1266, 'token_str': '北'}, {'sequence': '[CLS] 中国的首都是南京。 [SEP]', 'score': 0.1194809079170227, 'token': 1298, 'token_str': '南'}, {'sequence': '[CLS] 中国的首都是东京。 [SEP]', 'score': 0.0037803512532263994, 'token': 691, 'token_str': '东'}, {'sequence': '[CLS] 中国的首都是普京。 [SEP]', 'score': 0.0017127094324678183, 'token': 3249, 'token_str': '普'}, {'sequence': '[CLS] 中国的首都是望京。 [SEP]', 'score': 0.001687526935711503, 'token': 3307, 'token_str': '望'} ] ``` Here is how to use this model to get the features of a given text in PyTorch: ```python from transformers import BertTokenizer, BertModel tokenizer = BertTokenizer.from_pretrained('uer/chinese_roberta_L-8_H-512') model = BertModel.from_pretrained("uer/chinese_roberta_L-8_H-512") text = "用你喜欢的任何文本替换我。" encoded_input = tokenizer(text, return_tensors='pt') output = model(**encoded_input) ``` and in TensorFlow: ```python from transformers import BertTokenizer, TFBertModel tokenizer = BertTokenizer.from_pretrained('uer/chinese_roberta_L-8_H-512') model = TFBertModel.from_pretrained("uer/chinese_roberta_L-8_H-512") text = "用你喜欢的任何文本替换我。" encoded_input = tokenizer(text, return_tensors='tf') output = model(encoded_input) ``` ## Training data [CLUECorpusSmall](https://github.com/CLUEbenchmark/CLUECorpus2020/) is used as training data. We found that models pre-trained on CLUECorpusSmall outperform those pre-trained on CLUECorpus2020, although CLUECorpus2020 is much larger than CLUECorpusSmall. ## Training procedure Models are pre-trained by [UER-py](https://github.com/dbiir/UER-py/) on [Tencent Cloud](https://cloud.tencent.com/). We pre-train 1,000,000 steps with a sequence length of 128 and then pre-train 250,000 additional steps with a sequence length of 512. We use the same hyper-parameters on different model sizes. Taking the case of RoBERTa-Medium Stage1: ``` python3 preprocess.py --corpus_path corpora/cluecorpussmall.txt \ --vocab_path models/google_zh_vocab.txt \ --dataset_path cluecorpussmall_seq128_dataset.pt \ --processes_num 32 --seq_length 128 \ --dynamic_masking --data_processor mlm ``` ``` python3 pretrain.py --dataset_path cluecorpussmall_seq128_dataset.pt \ --vocab_path models/google_zh_vocab.txt \ --config_path models/bert/medium_config.json \ --output_model_path models/cluecorpussmall_roberta_medium_seq128_model.bin \ --world_size 8 --gpu_ranks 0 1 2 3 4 5 6 7 \ --total_steps 1000000 --save_checkpoint_steps 100000 --report_steps 50000 \ --learning_rate 1e-4 --batch_size 64 \ --data_processor mlm --target mlm ``` Stage2: ``` python3 preprocess.py --corpus_path corpora/cluecorpussmall.txt \ --vocab_path models/google_zh_vocab.txt \ --dataset_path cluecorpussmall_seq512_dataset.pt \ --processes_num 32 --seq_length 512 \ --dynamic_masking --data_processor mlm ``` ``` python3 pretrain.py --dataset_path cluecorpussmall_seq512_dataset.pt \ --vocab_path models/google_zh_vocab.txt \ --pretrained_model_path models/cluecorpussmall_roberta_medium_seq128_model.bin-1000000 \ --config_path models/bert/medium_config.json \ --output_model_path models/cluecorpussmall_roberta_medium_seq512_model.bin \ --world_size 8 --gpu_ranks 0 1 2 3 4 5 6 7 \ --total_steps 250000 --save_checkpoint_steps 50000 --report_steps 10000 \ --learning_rate 5e-5 --batch_size 16 \ --data_processor mlm --target mlm ``` Finally, we convert the pre-trained model into Huggingface's format: ``` python3 scripts/convert_bert_from_uer_to_huggingface.py --input_model_path models/cluecorpussmall_roberta_medium_seq512_model.bin-250000 \ --output_model_path pytorch_model.bin \ --layers_num 8 --type mlm ``` ### BibTeX entry and citation info ``` @article{devlin2018bert, title={Bert: Pre-training of deep bidirectional transformers for language understanding}, author={Devlin, Jacob and Chang, Ming-Wei and Lee, Kenton and Toutanova, Kristina}, journal={arXiv preprint arXiv:1810.04805}, year={2018} } @article{liu2019roberta, title={Roberta: A robustly optimized bert pretraining approach}, author={Liu, Yinhan and Ott, Myle and Goyal, Naman and Du, Jingfei and Joshi, Mandar and Chen, Danqi and Levy, Omer and Lewis, Mike and Zettlemoyer, Luke and Stoyanov, Veselin}, journal={arXiv preprint arXiv:1907.11692}, year={2019} } @article{turc2019, title={Well-Read Students Learn Better: On the Importance of Pre-training Compact Models}, author={Turc, Iulia and Chang, Ming-Wei and Lee, Kenton and Toutanova, Kristina}, journal={arXiv preprint arXiv:1908.08962v2 }, year={2019} } @article{zhao2019uer, title={UER: An Open-Source Toolkit for Pre-training Models}, author={Zhao, Zhe and Chen, Hui and Zhang, Jinbin and Zhao, Xin and Liu, Tao and Lu, Wei and Chen, Xi and Deng, Haotang and Ju, Qi and Du, Xiaoyong}, journal={EMNLP-IJCNLP 2019}, pages={241}, year={2019} } @article{zhao2023tencentpretrain, title={TencentPretrain: A Scalable and Flexible Toolkit for Pre-training Models of Different Modalities}, author={Zhao, Zhe and Li, Yudong and Hou, Cheng and Zhao, Jing and others}, journal={ACL 2023}, pages={217}, year={2023} } ``` [2_128]:https://huggingface.co/uer/chinese_roberta_L-2_H-128 [2_256]:https://huggingface.co/uer/chinese_roberta_L-2_H-256 [2_512]:https://huggingface.co/uer/chinese_roberta_L-2_H-512 [2_768]:https://huggingface.co/uer/chinese_roberta_L-2_H-768 [4_128]:https://huggingface.co/uer/chinese_roberta_L-4_H-128 [4_256]:https://huggingface.co/uer/chinese_roberta_L-4_H-256 [4_512]:https://huggingface.co/uer/chinese_roberta_L-4_H-512 [4_768]:https://huggingface.co/uer/chinese_roberta_L-4_H-768 [6_128]:https://huggingface.co/uer/chinese_roberta_L-6_H-128 [6_256]:https://huggingface.co/uer/chinese_roberta_L-6_H-256 [6_512]:https://huggingface.co/uer/chinese_roberta_L-6_H-512 [6_768]:https://huggingface.co/uer/chinese_roberta_L-6_H-768 [8_128]:https://huggingface.co/uer/chinese_roberta_L-8_H-128 [8_256]:https://huggingface.co/uer/chinese_roberta_L-8_H-256 [8_512]:https://huggingface.co/uer/chinese_roberta_L-8_H-512 [8_768]:https://huggingface.co/uer/chinese_roberta_L-8_H-768 [10_128]:https://huggingface.co/uer/chinese_roberta_L-10_H-128 [10_256]:https://huggingface.co/uer/chinese_roberta_L-10_H-256 [10_512]:https://huggingface.co/uer/chinese_roberta_L-10_H-512 [10_768]:https://huggingface.co/uer/chinese_roberta_L-10_H-768 [12_128]:https://huggingface.co/uer/chinese_roberta_L-12_H-128 [12_256]:https://huggingface.co/uer/chinese_roberta_L-12_H-256 [12_512]:https://huggingface.co/uer/chinese_roberta_L-12_H-512 [12_768]:https://huggingface.co/uer/chinese_roberta_L-12_H-768
uer/chinese_roberta_L-4_H-256	uer	2023-08-30T02:25:07Z	117	3	transformers	[ "transformers", "pytorch", "tf", "jax", "bert", "fill-mask", "zh", "dataset:CLUECorpusSmall", "arxiv:1909.05658", "arxiv:2212.06385", "arxiv:1908.08962", "arxiv:1810.04805", "arxiv:1907.11692", "autotrain_compatible", "endpoints_compatible", "region:us" ]	fill-mask	2022-03-02T23:29:05Z	--- language: zh datasets: CLUECorpusSmall widget: - text: "北京是[MASK]国的首都。" --- # Chinese RoBERTa Miniatures ## Model description This is the set of 24 Chinese RoBERTa models pre-trained by [UER-py](https://github.com/dbiir/UER-py/), which is introduced in [this paper](https://arxiv.org/abs/1909.05658). Besides, the models could also be pre-trained by [TencentPretrain](https://github.com/Tencent/TencentPretrain) introduced in [this paper](https://arxiv.org/abs/2212.06385), which inherits UER-py to support models with parameters above one billion, and extends it to a multimodal pre-training framework. [Turc et al.](https://arxiv.org/abs/1908.08962) have shown that the standard BERT recipe is effective on a wide range of model sizes. Following their paper, we released the 24 Chinese RoBERTa models. In order to facilitate users in reproducing the results, we used a publicly available corpus and provided all training details. You can download the 24 Chinese RoBERTa miniatures either from the [UER-py Modelzoo page](https://github.com/dbiir/UER-py/wiki/Modelzoo), or via HuggingFace from the links below: \| \| H=128 \| H=256 \| H=512 \| H=768 \| \| -------- \| :-----------------------: \| :-----------------------: \| :-------------------------: \| :-------------------------: \| \| L=2 \| [2/128 (Tiny)][2_128] \| [2/256][2_256] \| [2/512][2_512] \| [2/768][2_768] \| \| L=4 \| [4/128][4_128] \| [4/256 (Mini)][4_256] \| [4/512 (Small)][4_512] \| [4/768][4_768] \| \| L=6 \| [6/128][6_128] \| [6/256][6_256] \| [6/512][6_512] \| [6/768][6_768] \| \| L=8 \| [8/128][8_128] \| [8/256][8_256] \| [8/512 (Medium)][8_512] \| [8/768][8_768] \| \| L=10 \| [10/128][10_128] \| [10/256][10_256] \| [10/512][10_512] \| [10/768][10_768] \| \| L=12 \| [12/128][12_128] \| [12/256][12_256] \| [12/512][12_512] \| [12/768 (Base)][12_768] \| Here are scores on the devlopment set of six Chinese tasks: \| Model \| Score \| book_review \| chnsenticorp \| lcqmc \| tnews(CLUE) \| iflytek(CLUE) \| ocnli(CLUE) \| \| -------------- \| :---: \| :----: \| :----------: \| :---: \| :---------: \| :-----------: \| :---------: \| \| RoBERTa-Tiny \| 72.3 \| 83.4 \| 91.4 \| 81.8 \| 62.0 \| 55.0 \| 60.3 \| \| RoBERTa-Mini \| 75.9 \| 85.7 \| 93.7 \| 86.1 \| 63.9 \| 58.3 \| 67.4 \| \| RoBERTa-Small \| 76.9 \| 87.5 \| 93.4 \| 86.5 \| 65.1 \| 59.4 \| 69.7 \| \| RoBERTa-Medium \| 78.0 \| 88.7 \| 94.8 \| 88.1 \| 65.6 \| 59.5 \| 71.2 \| \| RoBERTa-Base \| 79.7 \| 90.1 \| 95.2 \| 89.2 \| 67.0 \| 60.9 \| 75.5 \| For each task, we selected the best fine-tuning hyperparameters from the lists below, and trained with the sequence length of 128: - epochs: 3, 5, 8 - batch sizes: 32, 64 - learning rates: 3e-5, 1e-4, 3e-4 ## How to use You can use this model directly with a pipeline for masked language modeling (take the case of RoBERTa-Medium): ```python >>> from transformers import pipeline >>> unmasker = pipeline('fill-mask', model='uer/chinese_roberta_L-8_H-512') >>> unmasker("中国的首都是[MASK]京。") [ {'sequence': '[CLS] 中国的首都是北京。 [SEP]', 'score': 0.8701988458633423, 'token': 1266, 'token_str': '北'}, {'sequence': '[CLS] 中国的首都是南京。 [SEP]', 'score': 0.1194809079170227, 'token': 1298, 'token_str': '南'}, {'sequence': '[CLS] 中国的首都是东京。 [SEP]', 'score': 0.0037803512532263994, 'token': 691, 'token_str': '东'}, {'sequence': '[CLS] 中国的首都是普京。 [SEP]', 'score': 0.0017127094324678183, 'token': 3249, 'token_str': '普'}, {'sequence': '[CLS] 中国的首都是望京。 [SEP]', 'score': 0.001687526935711503, 'token': 3307, 'token_str': '望'} ] ``` Here is how to use this model to get the features of a given text in PyTorch: ```python from transformers import BertTokenizer, BertModel tokenizer = BertTokenizer.from_pretrained('uer/chinese_roberta_L-8_H-512') model = BertModel.from_pretrained("uer/chinese_roberta_L-8_H-512") text = "用你喜欢的任何文本替换我。" encoded_input = tokenizer(text, return_tensors='pt') output = model(**encoded_input) ``` and in TensorFlow: ```python from transformers import BertTokenizer, TFBertModel tokenizer = BertTokenizer.from_pretrained('uer/chinese_roberta_L-8_H-512') model = TFBertModel.from_pretrained("uer/chinese_roberta_L-8_H-512") text = "用你喜欢的任何文本替换我。" encoded_input = tokenizer(text, return_tensors='tf') output = model(encoded_input) ``` ## Training data [CLUECorpusSmall](https://github.com/CLUEbenchmark/CLUECorpus2020/) is used as training data. We found that models pre-trained on CLUECorpusSmall outperform those pre-trained on CLUECorpus2020, although CLUECorpus2020 is much larger than CLUECorpusSmall. ## Training procedure Models are pre-trained by [UER-py](https://github.com/dbiir/UER-py/) on [Tencent Cloud](https://cloud.tencent.com/). We pre-train 1,000,000 steps with a sequence length of 128 and then pre-train 250,000 additional steps with a sequence length of 512. We use the same hyper-parameters on different model sizes. Taking the case of RoBERTa-Medium Stage1: ``` python3 preprocess.py --corpus_path corpora/cluecorpussmall.txt \ --vocab_path models/google_zh_vocab.txt \ --dataset_path cluecorpussmall_seq128_dataset.pt \ --processes_num 32 --seq_length 128 \ --dynamic_masking --data_processor mlm ``` ``` python3 pretrain.py --dataset_path cluecorpussmall_seq128_dataset.pt \ --vocab_path models/google_zh_vocab.txt \ --config_path models/bert/medium_config.json \ --output_model_path models/cluecorpussmall_roberta_medium_seq128_model.bin \ --world_size 8 --gpu_ranks 0 1 2 3 4 5 6 7 \ --total_steps 1000000 --save_checkpoint_steps 100000 --report_steps 50000 \ --learning_rate 1e-4 --batch_size 64 \ --data_processor mlm --target mlm ``` Stage2: ``` python3 preprocess.py --corpus_path corpora/cluecorpussmall.txt \ --vocab_path models/google_zh_vocab.txt \ --dataset_path cluecorpussmall_seq512_dataset.pt \ --processes_num 32 --seq_length 512 \ --dynamic_masking --data_processor mlm ``` ``` python3 pretrain.py --dataset_path cluecorpussmall_seq512_dataset.pt \ --vocab_path models/google_zh_vocab.txt \ --pretrained_model_path models/cluecorpussmall_roberta_medium_seq128_model.bin-1000000 \ --config_path models/bert/medium_config.json \ --output_model_path models/cluecorpussmall_roberta_medium_seq512_model.bin \ --world_size 8 --gpu_ranks 0 1 2 3 4 5 6 7 \ --total_steps 250000 --save_checkpoint_steps 50000 --report_steps 10000 \ --learning_rate 5e-5 --batch_size 16 \ --data_processor mlm --target mlm ``` Finally, we convert the pre-trained model into Huggingface's format: ``` python3 scripts/convert_bert_from_uer_to_huggingface.py --input_model_path models/cluecorpussmall_roberta_medium_seq512_model.bin-250000 \ --output_model_path pytorch_model.bin \ --layers_num 8 --type mlm ``` ### BibTeX entry and citation info ``` @article{devlin2018bert, title={Bert: Pre-training of deep bidirectional transformers for language understanding}, author={Devlin, Jacob and Chang, Ming-Wei and Lee, Kenton and Toutanova, Kristina}, journal={arXiv preprint arXiv:1810.04805}, year={2018} } @article{liu2019roberta, title={Roberta: A robustly optimized bert pretraining approach}, author={Liu, Yinhan and Ott, Myle and Goyal, Naman and Du, Jingfei and Joshi, Mandar and Chen, Danqi and Levy, Omer and Lewis, Mike and Zettlemoyer, Luke and Stoyanov, Veselin}, journal={arXiv preprint arXiv:1907.11692}, year={2019} } @article{turc2019, title={Well-Read Students Learn Better: On the Importance of Pre-training Compact Models}, author={Turc, Iulia and Chang, Ming-Wei and Lee, Kenton and Toutanova, Kristina}, journal={arXiv preprint arXiv:1908.08962v2 }, year={2019} } @article{zhao2019uer, title={UER: An Open-Source Toolkit for Pre-training Models}, author={Zhao, Zhe and Chen, Hui and Zhang, Jinbin and Zhao, Xin and Liu, Tao and Lu, Wei and Chen, Xi and Deng, Haotang and Ju, Qi and Du, Xiaoyong}, journal={EMNLP-IJCNLP 2019}, pages={241}, year={2019} } @article{zhao2023tencentpretrain, title={TencentPretrain: A Scalable and Flexible Toolkit for Pre-training Models of Different Modalities}, author={Zhao, Zhe and Li, Yudong and Hou, Cheng and Zhao, Jing and others}, journal={ACL 2023}, pages={217}, year={2023} } ``` [2_128]:https://huggingface.co/uer/chinese_roberta_L-2_H-128 [2_256]:https://huggingface.co/uer/chinese_roberta_L-2_H-256 [2_512]:https://huggingface.co/uer/chinese_roberta_L-2_H-512 [2_768]:https://huggingface.co/uer/chinese_roberta_L-2_H-768 [4_128]:https://huggingface.co/uer/chinese_roberta_L-4_H-128 [4_256]:https://huggingface.co/uer/chinese_roberta_L-4_H-256 [4_512]:https://huggingface.co/uer/chinese_roberta_L-4_H-512 [4_768]:https://huggingface.co/uer/chinese_roberta_L-4_H-768 [6_128]:https://huggingface.co/uer/chinese_roberta_L-6_H-128 [6_256]:https://huggingface.co/uer/chinese_roberta_L-6_H-256 [6_512]:https://huggingface.co/uer/chinese_roberta_L-6_H-512 [6_768]:https://huggingface.co/uer/chinese_roberta_L-6_H-768 [8_128]:https://huggingface.co/uer/chinese_roberta_L-8_H-128 [8_256]:https://huggingface.co/uer/chinese_roberta_L-8_H-256 [8_512]:https://huggingface.co/uer/chinese_roberta_L-8_H-512 [8_768]:https://huggingface.co/uer/chinese_roberta_L-8_H-768 [10_128]:https://huggingface.co/uer/chinese_roberta_L-10_H-128 [10_256]:https://huggingface.co/uer/chinese_roberta_L-10_H-256 [10_512]:https://huggingface.co/uer/chinese_roberta_L-10_H-512 [10_768]:https://huggingface.co/uer/chinese_roberta_L-10_H-768 [12_128]:https://huggingface.co/uer/chinese_roberta_L-12_H-128 [12_256]:https://huggingface.co/uer/chinese_roberta_L-12_H-256 [12_512]:https://huggingface.co/uer/chinese_roberta_L-12_H-512 [12_768]:https://huggingface.co/uer/chinese_roberta_L-12_H-768
yoo11/taskC_distilbert_model	yoo11	2023-08-30T02:21:22Z	61	0	transformers	[ "transformers", "tf", "distilbert", "text-classification", "generated_from_keras_callback", "base_model:distilbert/distilbert-base-uncased", "base_model:finetune:distilbert/distilbert-base-uncased", "license:apache-2.0", "autotrain_compatible", "endpoints_compatible", "region:us" ]	text-classification	2023-08-30T02:17:13Z	--- license: apache-2.0 base_model: distilbert-base-uncased tags: - generated_from_keras_callback model-index: - name: yoo11/taskC_distilbert_model results: [] --- <!-- This model card has been generated automatically according to the information Keras had access to. You should probably proofread and complete it, then remove this comment. --> # yoo11/taskC_distilbert_model This model is a fine-tuned version of [distilbert-base-uncased](https://huggingface.co/distilbert-base-uncased) on an unknown dataset. It achieves the following results on the evaluation set: - Train Loss: 1.2046 - Validation Loss: 1.3859 - Train Accuracy: 0.5247 - Epoch: 2 ## Model description More information needed ## Intended uses & limitations More information needed ## Training and evaluation data More information needed ## Training procedure ### Training hyperparameters The following hyperparameters were used during training: - optimizer: {'name': 'Adam', 'weight_decay': None, 'clipnorm': None, 'global_clipnorm': None, 'clipvalue': None, 'use_ema': False, 'ema_momentum': 0.99, 'ema_overwrite_frequency': None, 'jit_compile': True, 'is_legacy_optimizer': False, 'learning_rate': {'class_name': 'PolynomialDecay', 'config': {'initial_learning_rate': 2e-05, 'decay_steps': 1060, 'end_learning_rate': 0.0, 'power': 1.0, 'cycle': False, 'name': None}}, 'beta_1': 0.9, 'beta_2': 0.999, 'epsilon': 1e-08, 'amsgrad': False} - training_precision: float32 ### Training results \| Train Loss \| Validation Loss \| Train Accuracy \| Epoch \| \|:----------:\|:---------------:\|:--------------:\|:-----:\| \| 1.9115 \| 1.6364 \| 0.4444 \| 0 \| \| 1.4956 \| 1.4539 \| 0.4835 \| 1 \| \| 1.2046 \| 1.3859 \| 0.5247 \| 2 \| ### Framework versions - Transformers 4.32.1 - TensorFlow 2.12.0 - Datasets 2.14.4 - Tokenizers 0.13.3
Xieyiyiyi/ceshi	Xieyiyiyi	2023-08-30T02:12:23Z	0	0	null	[ "license:bsl-1.0", "region:us" ]	null	2023-02-21T08:33:29Z	--- license: bsl-1.0 --- [[toc]] - [Refrence111](#Refrence111) ## Refrence 1111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111 1111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111 1111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111 ## Refrence111
Chanblock/Llama-2-7b-chat-hf-250_data_final	Chanblock	2023-08-30T02:08:21Z	0	0	null	[ "tensorboard", "generated_from_trainer", "base_model:NousResearch/Llama-2-7b-hf", "base_model:finetune:NousResearch/Llama-2-7b-hf", "region:us" ]	null	2023-08-30T02:03:56Z	--- base_model: NousResearch/Llama-2-7b-hf tags: - generated_from_trainer model-index: - name: Llama-2-7b-chat-hf-250_data_final results: [] --- <!-- This model card has been generated automatically according to the information the Trainer had access to. You should probably proofread and complete it, then remove this comment. --> # Llama-2-7b-chat-hf-250_data_final This model is a fine-tuned version of [NousResearch/Llama-2-7b-hf](https://huggingface.co/NousResearch/Llama-2-7b-hf) on an unknown dataset. ## Model description More information needed ## Intended uses & limitations More information needed ## Training and evaluation data More information needed ## Training procedure ### Training hyperparameters The following hyperparameters were used during training: - learning_rate: 0.0002 - train_batch_size: 3 - eval_batch_size: 3 - seed: 42 - optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08 - lr_scheduler_type: linear - num_epochs: 1 ### Training results ### Framework versions - Transformers 4.31.0 - Pytorch 2.0.1+cu118 - Datasets 2.14.4 - Tokenizers 0.13.3
jason-lee08/my_model	jason-lee08	2023-08-30T01:21:53Z	186	0	transformers	[ "transformers", "safetensors", "gpt_neo", "text-generation", "generated_from_trainer", "base_model:roneneldan/TinyStories-33M", "base_model:finetune:roneneldan/TinyStories-33M", "autotrain_compatible", "endpoints_compatible", "region:us" ]	text-generation	2023-08-30T01:20:23Z	--- base_model: roneneldan/TinyStories-33M tags: - generated_from_trainer model-index: - name: my_model results: [] --- <!-- This model card has been generated automatically according to the information the Trainer had access to. You should probably proofread and complete it, then remove this comment. --> # my_model This model is a fine-tuned version of [roneneldan/TinyStories-33M](https://huggingface.co/roneneldan/TinyStories-33M) on the None dataset. It achieves the following results on the evaluation set: - Loss: 1.0283 ## Model description More information needed ## Intended uses & limitations More information needed ## Training and evaluation data More information needed ## Training procedure ### Training hyperparameters The following hyperparameters were used during training: - learning_rate: 3e-05 - train_batch_size: 32 - eval_batch_size: 8 - seed: 42 - gradient_accumulation_steps: 16 - total_train_batch_size: 512 - optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08 - lr_scheduler_type: linear - training_steps: 8 ### Training results \| Training Loss \| Epoch \| Step \| Validation Loss \| \|:-------------:\|:-----:\|:----:\|:---------------:\| \| No log \| 0.11 \| 2 \| 1.1872 \| \| No log \| 0.22 \| 4 \| 1.0699 \| \| No log \| 0.33 \| 6 \| 1.0373 \| \| No log \| 0.44 \| 8 \| 1.0283 \| ### Framework versions - Transformers 4.32.1 - Pytorch 2.0.1 - Datasets 2.12.0 - Tokenizers 0.13.2
TomyAI/jojimizugi	TomyAI	2023-08-30T01:18:03Z	0	7	null	[ "license:creativeml-openrail-m", "region:us" ]	null	2023-08-30T00:03:53Z	--- license: creativeml-openrail-m --- 可愛い水着のLoRAです。プロンプト：形状： swimsuit:ワンピース型 tankini:タンクトップ型～チューブトップ型 bikini:三角ビキニ装飾： frilled:主に肩～胸のフリフリ skirt:スカート型模様： printed:模様入 gingham check:ギンガムチェック polca dot:水玉模様 striped:横縞 floral printed:花柄 scale printed:鱗柄（人魚風）これ以外でも(模様の名前) printed swimsuitとすれば大体反応します。ウェイトを1,1,1,1,1,1,1,1,1,0,0,0,1,1,1,0,0 にすると顔への影響が減りますが、柄の精度が落ちます。 ![jojimizugi](jojimizugi.png) ![jojimizugi_grid](jojimizugi_grid.jpg)
Adrian101-hnd/ppo-LunarLander-v2	Adrian101-hnd	2023-08-30T01:05:09Z	1	0	stable-baselines3	[ "stable-baselines3", "LunarLander-v2", "deep-reinforcement-learning", "reinforcement-learning", "model-index", "region:us" ]	reinforcement-learning	2023-08-30T01:04:46Z	--- library_name: stable-baselines3 tags: - LunarLander-v2 - deep-reinforcement-learning - reinforcement-learning - stable-baselines3 model-index: - name: PPO results: - task: type: reinforcement-learning name: reinforcement-learning dataset: name: LunarLander-v2 type: LunarLander-v2 metrics: - type: mean_reward value: 252.32 +/- 21.07 name: mean_reward verified: false --- # PPO Agent playing LunarLander-v2 This is a trained model of a PPO agent playing LunarLander-v2 using the [stable-baselines3 library](https://github.com/DLR-RM/stable-baselines3). ## Usage (with Stable-baselines3) TODO: Add your code ```python from stable_baselines3 import ... from huggingface_sb3 import load_from_hub ... ```
AzulkerAI/Crickey	AzulkerAI	2023-08-30T00:50:12Z	0	0	null	[ "region:us" ]	null	2023-08-30T00:38:11Z	blahblahblahblah look its funny thingy
touchtech/fashion-images-perspectives-vit-huge-patch14-224-in21k	touchtech	2023-08-30T00:48:28Z	6	0	transformers	[ "transformers", "pytorch", "vit", "image-classification", "vision", "generated_from_trainer", "dataset:imagefolder", "base_model:google/vit-huge-patch14-224-in21k", "base_model:finetune:google/vit-huge-patch14-224-in21k", "license:apache-2.0", "model-index", "autotrain_compatible", "endpoints_compatible", "region:us" ]	image-classification	2023-08-29T19:30:32Z	--- license: apache-2.0 base_model: google/vit-huge-patch14-224-in21k tags: - image-classification - vision - generated_from_trainer datasets: - imagefolder metrics: - accuracy model-index: - name: fashion-images-perspectives-vit-huge-patch14-224-in21k results: - task: name: Image Classification type: image-classification dataset: name: touchtech/fashion-images-perspectives type: imagefolder config: default split: train args: default metrics: - name: Accuracy type: accuracy value: 0.9149976711690732 --- <!-- This model card has been generated automatically according to the information the Trainer had access to. You should probably proofread and complete it, then remove this comment. --> # fashion-images-perspectives-vit-huge-patch14-224-in21k This model is a fine-tuned version of [google/vit-huge-patch14-224-in21k](https://huggingface.co/google/vit-huge-patch14-224-in21k) on the touchtech/fashion-images-perspectives dataset. It achieves the following results on the evaluation set: - Loss: 0.2604 - Accuracy: 0.9150 ## Model description More information needed ## Intended uses & limitations More information needed ## Training and evaluation data More information needed ## Training procedure ### Training hyperparameters The following hyperparameters were used during training: - learning_rate: 2e-05 - train_batch_size: 8 - eval_batch_size: 8 - seed: 1337 - optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08 - lr_scheduler_type: linear - num_epochs: 5.0 ### Training results \| Training Loss \| Epoch \| Step \| Validation Loss \| Accuracy \| \|:-------------:\|:-----:\|:-----:\|:---------------:\|:--------:\| \| 0.5932 \| 1.0 \| 3042 \| 0.3863 \| 0.8889 \| \| 0.4286 \| 2.0 \| 6084 \| 0.4201 \| 0.8642 \| \| 0.3628 \| 3.0 \| 9126 \| 0.2820 \| 0.9101 \| \| 0.3183 \| 4.0 \| 12168 \| 0.2604 \| 0.9150 \| \| 0.2648 \| 5.0 \| 15210 \| 0.2710 \| 0.9150 \| ### Framework versions - Transformers 4.33.0.dev0 - Pytorch 2.0.1+cu118 - Datasets 2.14.4 - Tokenizers 0.13.3
yoo11/taskA_distilbert_model	yoo11	2023-08-30T00:23:08Z	61	0	transformers	[ "transformers", "tf", "distilbert", "text-classification", "generated_from_keras_callback", "base_model:distilbert/distilbert-base-uncased", "base_model:finetune:distilbert/distilbert-base-uncased", "license:apache-2.0", "autotrain_compatible", "endpoints_compatible", "region:us" ]	text-classification	2023-08-30T00:15:17Z	--- license: apache-2.0 base_model: distilbert-base-uncased tags: - generated_from_keras_callback model-index: - name: yoo11/taskA_distilbert_model results: [] --- <!-- This model card has been generated automatically according to the information Keras had access to. You should probably proofread and complete it, then remove this comment. --> # yoo11/taskA_distilbert_model This model is a fine-tuned version of [distilbert-base-uncased](https://huggingface.co/distilbert-base-uncased) on an unknown dataset. It achieves the following results on the evaluation set: - Train Loss: 0.1501 - Validation Loss: 0.4055 - Train Accuracy: 0.863 - Epoch: 2 ## Model description More information needed ## Intended uses & limitations More information needed ## Training and evaluation data More information needed ## Training procedure ### Training hyperparameters The following hyperparameters were used during training: - optimizer: {'name': 'Adam', 'weight_decay': None, 'clipnorm': None, 'global_clipnorm': None, 'clipvalue': None, 'use_ema': False, 'ema_momentum': 0.99, 'ema_overwrite_frequency': None, 'jit_compile': True, 'is_legacy_optimizer': False, 'learning_rate': {'class_name': 'PolynomialDecay', 'config': {'initial_learning_rate': 2e-05, 'decay_steps': 4375, 'end_learning_rate': 0.0, 'power': 1.0, 'cycle': False, 'name': None}}, 'beta_1': 0.9, 'beta_2': 0.999, 'epsilon': 1e-08, 'amsgrad': False} - training_precision: float32 ### Training results \| Train Loss \| Validation Loss \| Train Accuracy \| Epoch \| \|:----------:\|:---------------:\|:--------------:\|:-----:\| \| 0.3800 \| 0.3247 \| 0.861 \| 0 \| \| 0.2582 \| 0.3318 \| 0.861 \| 1 \| \| 0.1501 \| 0.4055 \| 0.863 \| 2 \| ### Framework versions - Transformers 4.32.1 - TensorFlow 2.12.0 - Datasets 2.14.4 - Tokenizers 0.13.3

Ryukijano/masked-lm-tpu

Ryukijano

2023-08-30T12:30:04Z

5

0

transformers

[ "transformers", "tf", "roberta", "fill-mask", "generated_from_keras_callback", "base_model:FacebookAI/roberta-base", "base_model:finetune:FacebookAI/roberta-base", "license:mit", "autotrain_compatible", "endpoints_compatible", "region:us" ]

fill-mask

2023-08-30T07:41:30Z

--- license: mit base_model: roberta-base tags: - generated_from_keras_callback model-index: - name: Ryukijano/masked-lm-tpu results: [] ---  # Ryukijano/masked-lm-tpu This model is a fine-tuned version of [roberta-base](https://huggingface.co/roberta-base) on an unknown dataset. It achieves the following results on the evaluation set: - Train Loss: 5.8422 - Train Accuracy: 0.0344 - Validation Loss: 5.8152 - Validation Accuracy: 0.0340 - Epoch: 48 ## Model description More information needed ## Intended uses & limitations More information needed ## Training and evaluation data More information needed ## Training procedure ### Training hyperparameters The following hyperparameters were used during training: - optimizer: {'name': 'AdamWeightDecay', 'learning_rate': {'class_name': 'WarmUp', 'config': {'initial_learning_rate': 0.0001, 'decay_schedule_fn': {'class_name': 'PolynomialDecay', 'config': {'initial_learning_rate': 0.0001, 'decay_steps': 111625, 'end_learning_rate': 0.0, 'power': 1.0, 'cycle': False, 'name': None}, '__passive_serialization__': True}, 'warmup_steps': 5875, 'power': 1.0, 'name': None}}, 'decay': 0.0, 'beta_1': 0.9, 'beta_2': 0.999, 'epsilon': 1e-08, 'amsgrad': False, 'weight_decay_rate': 0.001} - training_precision: float32 ### Training results | Train Loss | Train Accuracy | Validation Loss | Validation Accuracy | Epoch | |:----------:|:--------------:|:---------------:|:-------------------:|:-----:| | 10.2437 | 0.0000 | 10.1909 | 0.0000 | 0 | | 10.1151 | 0.0001 | 9.9763 | 0.0016 | 1 | | 9.8665 | 0.0107 | 9.6535 | 0.0215 | 2 | | 9.5331 | 0.0230 | 9.2992 | 0.0223 | 3 | | 9.2000 | 0.0231 | 8.9944 | 0.0222 | 4 | | 8.9195 | 0.0229 | 8.7450 | 0.0224 | 5 | | 8.6997 | 0.0231 | 8.6124 | 0.0219 | 6 | | 8.5689 | 0.0229 | 8.4904 | 0.0222 | 7 | | 8.4525 | 0.0230 | 8.3865 | 0.0223 | 8 | | 8.3594 | 0.0230 | 8.3069 | 0.0221 | 9 | | 8.2662 | 0.0231 | 8.2092 | 0.0224 | 10 | | 8.1956 | 0.0231 | 8.1208 | 0.0222 | 11 | | 8.1285 | 0.0229 | 8.0806 | 0.0219 | 12 | | 8.0345 | 0.0234 | 8.0030 | 0.0220 | 13 | | 7.9960 | 0.0228 | 7.9144 | 0.0224 | 14 | | 7.9065 | 0.0231 | 7.8661 | 0.0221 | 15 | | 7.8449 | 0.0229 | 7.7873 | 0.0219 | 16 | | 7.7673 | 0.0232 | 7.6903 | 0.0229 | 17 | | 7.6868 | 0.0242 | 7.6129 | 0.0243 | 18 | | 7.6206 | 0.0250 | 7.5579 | 0.0246 | 19 | | 7.5231 | 0.0258 | 7.4564 | 0.0254 | 20 | | 7.4589 | 0.0262 | 7.4136 | 0.0255 | 21 | | 7.3658 | 0.0269 | 7.2941 | 0.0265 | 22 | | 7.2832 | 0.0274 | 7.1998 | 0.0270 | 23 | | 7.2035 | 0.0275 | 7.1203 | 0.0271 | 24 | | 7.1116 | 0.0280 | 7.0582 | 0.0269 | 25 | | 7.0099 | 0.0287 | 6.9567 | 0.0287 | 26 | | 6.9296 | 0.0294 | 6.8759 | 0.0287 | 27 | | 6.8524 | 0.0296 | 6.8272 | 0.0285 | 28 | | 6.7757 | 0.0300 | 6.7311 | 0.0291 | 29 | | 6.7031 | 0.0304 | 6.6316 | 0.0305 | 30 | | 6.6361 | 0.0306 | 6.5744 | 0.0307 | 31 | | 6.5578 | 0.0312 | 6.4946 | 0.0312 | 32 | | 6.4674 | 0.0319 | 6.4212 | 0.0314 | 33 | | 6.4096 | 0.0322 | 6.3557 | 0.0320 | 34 | | 6.3614 | 0.0321 | 6.3093 | 0.0322 | 35 | | 6.2754 | 0.0329 | 6.2240 | 0.0326 | 36 | | 6.2609 | 0.0326 | 6.2114 | 0.0321 | 37 | | 6.1866 | 0.0329 | 6.1645 | 0.0320 | 38 | | 6.1470 | 0.0330 | 6.1193 | 0.0323 | 39 | | 6.0936 | 0.0329 | 6.0600 | 0.0324 | 40 | | 6.0625 | 0.0330 | 6.0282 | 0.0323 | 41 | | 6.0062 | 0.0335 | 5.9649 | 0.0329 | 42 | | 5.9731 | 0.0339 | 5.9661 | 0.0330 | 43 | | 5.9460 | 0.0335 | 5.9259 | 0.0330 | 44 | | 5.9206 | 0.0338 | 5.8926 | 0.0333 | 45 | | 5.8734 | 0.0343 | 5.8471 | 0.0340 | 46 | | 5.8663 | 0.0341 | 5.8561 | 0.0337 | 47 | | 5.8422 | 0.0344 | 5.8152 | 0.0340 | 48 | ### Framework versions - Transformers 4.32.1 - TensorFlow 2.12.0 - Tokenizers 0.13.3

PlaceReporter99/Utility_Bot_Chat

PlaceReporter99

2023-08-30T12:23:38Z

0

transformers

[ "transformers", "conversational", "en", "dataset:fka/awesome-chatgpt-prompts", "license:mit", "endpoints_compatible", "region:us" ]

text-generation

2023-08-30T12:18:00Z

--- license: mit datasets: - fka/awesome-chatgpt-prompts language: - en metrics: - accuracy library_name: transformers pipeline_tag: conversational ---

Norod78/SDXL-jojoso_style-Lora

Norod78

2023-08-30T12:22:51Z

520

3

diffusers

[ "diffusers", "text-to-image", "stable-diffusion", "lora", "en", "base_model:stabilityai/stable-diffusion-xl-base-1.0", "base_model:adapter:stabilityai/stable-diffusion-xl-base-1.0", "region:us" ]

text-to-image

2023-08-17T12:33:25Z

--- base_model: stabilityai/stable-diffusion-xl-base-1.0 instance_prompt: jojoso style tags: - text-to-image - stable-diffusion - lora - diffusers widget: - text: jojoso style cute dog - text: jojoso style picure of a Pokemon - text: jojoso style Cthulhu rising from the sea in a great storm - text: jojoso style girl with a pearl earring by johannes vermeer inference: true language: - en --- # Trigger words Use "jojoso style" in your prompts # Examples The girl with a pearl earring by johannes vermeer, Very detailed, clean, high quality, sharp image, jojoso style ![The girl with a pearl earring by johannes vermeer](https://huggingface.co/Norod78/SDXL-jojoso_style-Lora/resolve/main/Examples/00189-20230816194930-558-the%20girl%20with%20a%20pearl%20earring%20by%20johannes%20vermeer%2C%20Very%20detailed%2C%20clean%2C%20high%20quality%2C%20sharp%20image%2C%20jojoso%20style%20_lora_SDXL-jojo.jpeg) Cthulhu rising from the sea in a great storm, Very detailed, clean, high quality, sharp image, jojoso style ![Cthulhu rising from the sea in a great storm](https://huggingface.co/Norod78/SDXL-jojoso_style-Lora/resolve/main/Examples/00224-20230816201509-557-Cthulhu%20rising%20from%20the%20sea%20in%20a%20great%20storm%2C%20Very%20detailed%2C%20clean%2C%20high%20quality%2C%20sharp%20image%2C%20jojoso%20style%20_lora_SDXL-jojoso_st.jpeg)

umerah/Task3

umerah

2023-08-30T12:14:26Z

10

0

transformers

[ "transformers", "pytorch", "gpt2", "text-generation", "generated_from_trainer", "base_model:distilbert/distilgpt2", "base_model:finetune:distilbert/distilgpt2", "license:apache-2.0", "autotrain_compatible", "text-generation-inference", "endpoints_compatible", "region:us" ]

text-generation

2023-08-30T09:05:51Z

--- license: apache-2.0 base_model: distilgpt2 tags: - generated_from_trainer model-index: - name: Task3 results: [] ---  # Task3 This model is a fine-tuned version of [distilgpt2](https://huggingface.co/distilgpt2) on the None dataset. It achieves the following results on the evaluation set: - Loss: 3.0246 ## Model description More information needed ## Intended uses & limitations More information needed ## Training and evaluation data More information needed ## Training procedure ### Training hyperparameters The following hyperparameters were used during training: - learning_rate: 0.0001 - train_batch_size: 16 - eval_batch_size: 16 - seed: 42 - optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08 - lr_scheduler_type: linear - num_epochs: 10 ### Training results | Training Loss | Epoch | Step | Validation Loss | |:-------------:|:-----:|:-----:|:---------------:| | 3.1753 | 1.0 | 1741 | 3.0775 | | 3.0194 | 2.0 | 3482 | 3.0338 | | 2.9194 | 3.0 | 5223 | 3.0157 | | 2.8401 | 4.0 | 6964 | 3.0063 | | 2.7765 | 5.0 | 8705 | 3.0064 | | 2.7266 | 6.0 | 10446 | 3.0093 | | 2.6817 | 7.0 | 12187 | 3.0105 | | 2.6458 | 8.0 | 13928 | 3.0156 | | 2.6195 | 9.0 | 15669 | 3.0205 | | 2.5997 | 10.0 | 17410 | 3.0246 | ### Framework versions - Transformers 4.32.1 - Pytorch 2.0.1+cu118 - Datasets 2.14.4 - Tokenizers 0.13.3

cvapict/yhi-message-type-paraphrase-mpnet-base-v2

cvapict

2023-08-30T12:11:46Z

4

0

sentence-transformers

[ "sentence-transformers", "pytorch", "mpnet", "setfit", "text-classification", "arxiv:2209.11055", "license:apache-2.0", "region:us" ]

text-classification

2023-08-30T12:11:04Z

--- license: apache-2.0 tags: - setfit - sentence-transformers - text-classification pipeline_tag: text-classification --- # cvapict/yhi-message-type-paraphrase-mpnet-base-v2 {'accuracy': 0.8536585365853658} This is a [SetFit model](https://github.com/huggingface/setfit) that can be used for text classification. The model has been trained using an efficient few-shot learning technique that involves: 1. Fine-tuning a [Sentence Transformer](https://www.sbert.net) with contrastive learning. 2. Training a classification head with features from the fine-tuned Sentence Transformer. ## Usage To use this model for inference, first install the SetFit library: ```bash python -m pip install setfit ``` You can then run inference as follows: ```python from setfit import SetFitModel # Download from Hub and run inference model = SetFitModel.from_pretrained("cvapict/yhi-message-type-paraphrase-mpnet-base-v2") # Run inference preds = model(["i loved the spiderman movie!", "pineapple on pizza is the worst 🤮"]) ``` ## BibTeX entry and citation info ```bibtex @article{https://doi.org/10.48550/arxiv.2209.11055, doi = {10.48550/ARXIV.2209.11055}, url = {https://arxiv.org/abs/2209.11055}, author = {Tunstall, Lewis and Reimers, Nils and Jo, Unso Eun Seo and Bates, Luke and Korat, Daniel and Wasserblat, Moshe and Pereg, Oren}, keywords = {Computation and Language (cs.CL), FOS: Computer and information sciences, FOS: Computer and information sciences}, title = {Efficient Few-Shot Learning Without Prompts}, publisher = {arXiv}, year = {2022}, copyright = {Creative Commons Attribution 4.0 International} } ```

vinayaksodar/ppo-LunarLander-v2

vinayaksodar

2023-08-30T12:10:20Z

0

stable-baselines3

[ "stable-baselines3", "LunarLander-v2", "deep-reinforcement-learning", "reinforcement-learning", "model-index", "region:us" ]

reinforcement-learning

2023-08-30T12:10:00Z

--- library_name: stable-baselines3 tags: - LunarLander-v2 - deep-reinforcement-learning - reinforcement-learning - stable-baselines3 model-index: - name: PPO results: - task: type: reinforcement-learning name: reinforcement-learning dataset: name: LunarLander-v2 type: LunarLander-v2 metrics: - type: mean_reward value: 267.79 +/- 9.72 name: mean_reward verified: false --- # **PPO** Agent playing **LunarLander-v2** This is a trained model of a **PPO** agent playing **LunarLander-v2** using the [stable-baselines3 library](https://github.com/DLR-RM/stable-baselines3). ## Usage (with Stable-baselines3) TODO: Add your code ```python from stable_baselines3 import ... from huggingface_sb3 import load_from_hub ... ```

bigscience/bloom-3b-intermediate

bigscience

2023-08-30T12:06:50Z

20

1

transformers

[ "transformers", "pytorch", "safetensors", "bloom", "text-generation", "ak", "ar", "as", "bm", "bn", "ca", "code", "en", "es", "eu", "fon", "fr", "gu", "hi", "id", "ig", "ki", "kn", "lg", "ln", "ml", "mr", "ne", "nso", "ny", "or", "pa", "pt", "rn", "rw", "sn", "st", "sw", "ta", "te", "tn", "ts", "tum", "tw", "ur", "vi", "wo", "xh", "yo", "zh", "zhs", "zht", "zu", "arxiv:1909.08053", "arxiv:2110.02861", "arxiv:2108.12409", "license:bigscience-bloom-rail-1.0", "autotrain_compatible", "text-generation-inference", "endpoints_compatible", "region:us" ]

text-generation

2022-08-07T17:47:50Z

--- license: bigscience-bloom-rail-1.0 language: - ak - ar - as - bm - bn - ca - code - en - es - eu - fon - fr - gu - hi - id - ig - ki - kn - lg - ln - ml - mr - ne - nso - ny - or - pa - pt - rn - rw - sn - st - sw - ta - te - tn - ts - tum - tw - ur - vi - wo - xh - yo - zh - zhs - zht - zu --- # WARNING: The checkpoints on this repo are not fully trained model. Evaluations of intermediary checkpoints and the final model will be added when conducted (see below). # BLOOM LM _BigScience Large Open-science Open-access Multilingual Language Model_ Model Card <img src="https://assets.website-files.com/6139f3cdcbbff3a68486761d/613cd8997b270da063e230c5_Tekengebied%201-p-500.png" alt="BigScience Logo" width="200"/> Version 1.3 / 11.July.2022 - Available intermediary checkpoints - global steps: + `1000`, `10000`, `50000`, `100000`, `150000`, `200000`, `250000`, `300000` You can check the available checkpoints by clicking on the branches section of the repo # How to load a specific version We use `git tags` to load a model in a specific version (eg. `global_step1000`): ```python from transformers import AutoModelForCausalLM model = AutoModelForCausalLM.from_pretrained( "bigscience/bloom-2b5-intermediate", revision="global_step1000", torch_dtype="auto", ) ``` # Table of Contents 1. [Model Details](#model-details) 2. [Uses](#uses) 3. [Training Data](#training-data) 4. [Risks and Limitations](#risks-and-limitations) 5. [Evaluation](#evaluation) 6. [Recommendations](#recommendations) 7. [Glossary and Calculations](#glossary-and-calculations) 8. [More Information](#more-information) 9. [Model Card Authors](#model-card-authors) --- # Model Details BLOOM is a type of language model, which is a probability distribution over sequences of words. Specifically, BLOOM is a Large Language Model (LLM), meaning that it is trained on vast amounts of text data using industrial-scale computational resources. As such, the model is able to capture the statistical tendencies of words, phrases, sentences, and larger spans of text that it is exposed to in the training data. ## Basics *This section provides information about the model type, version, license, funders, release date, developers, and contact information.* *It is useful for anyone who wants to reference the model.* <details> <summary>Click to expand</summary> **Developed by:** BigScience ([website](https://bigscience.huggingface.co)) *All collaborators are either volunteers or have an agreement with their employer. (Further breakdown of participants forthcoming.)* **Model Type:** Transformer-based Language Model **Version:** 1.0.0 **Languages:** Multiple; see [training data](#training-data) **License:** RAIL License v1.0 ([link](https://huggingface.co/spaces/bigscience/license)) **Release Date Estimate:** Monday, 11.July.2022 **Send Questions to:** [email protected] **Cite as:** BigScience, _BigScience Language Open-science Open-access Multilingual (BLOOM) Language Model_. International, May 2021-May 2022 **Funded by:** * The French government. * Hugging Face ([website](https://huggingface.co)). * Organizations of contributors. *(Further breakdown of organizations forthcoming.)* </details> ## Technical Specifications *This section includes details about the model objective and architecture, and the compute infrastructure.* *It is useful for people interested in model development.* <details> <summary>Click to expand</summary> Please see [the BLOOM training README](https://github.com/bigscience-workshop/bigscience/tree/master/train/tr11-176B-ml#readme) for full details on replicating training. ### Model Architecture and Objective * Modified from Megatron-LM GPT2 (see [paper](https://arxiv.org/abs/1909.08053), [BLOOM Megatron code](https://github.com/bigscience-workshop/Megatron-DeepSpeed)): * Decoder-only architecture * Layer normalization applied to word embeddings layer (`StableEmbedding`; see [code](https://github.com/facebookresearch/bitsandbytes), [paper](https://arxiv.org/pdf/2110.02861.pdf)) * ALiBI positional encodings (see [paper](https://arxiv.org/pdf/2108.12409.pdf)), with GeLU activation functions * 176 billion parameters: * 70 layers, 112 attention heads * Hidden layers are 14336-dimensional * Sequence length of 2048 tokens used (see [BLOOM tokenizer](https://huggingface.co/bigscience/tokenizer), [tokenizer description](#tokenization)) **Objective Function:** Cross Entropy with mean reduction (see [API documentation](https://pytorch.org/docs/stable/generated/torch.nn.CrossEntropyLoss.html#torch.nn.CrossEntropyLoss)). ### Compute infrastructure Jean Zay Public Supercomputer, provided by the French government (see [announcement](https://www.enseignementsup-recherche.gouv.fr/fr/signature-du-marche-d-acquisition-de-l-un-des-supercalculateurs-les-plus-puissants-d-europe-46733)). #### Hardware * 384 A100 80GB GPUs (48 nodes) * Additional 32 A100 80GB GPUs (4 nodes) in reserve * 8 GPUs per node Using NVLink 4 inter-gpu connects, 4 OmniPath links * CPU: AMD * CPU memory: 512GB per node * GPU memory: 640GB per node * Inter-node connect: Omni-Path Architecture (OPA) * NCCL-communications network: a fully dedicated subnet * Disc IO network: shared network with other types of nodes #### Software * Megatron-DeepSpeed ([Github link](https://github.com/bigscience-workshop/Megatron-DeepSpeed)) * DeepSpeed ([Github link](https://github.com/microsoft/DeepSpeed)) * PyTorch (pytorch-1.11 w/ CUDA-11.5; see [Github link](https://github.com/pytorch/pytorch)) * apex ([Github link](https://github.com/NVIDIA/apex)) </details> --- # Training *This section provides information about the training data, the speed and size of training elements, and the environmental impact of training.* *It is useful for people who want to learn more about the model inputs and training footprint.* <details> <summary>Click to expand</summary> ## Training Data *This section provides a high-level overview of the training data. It is relevant for anyone who wants to know the basics of what the model is learning.* Details for each dataset are provided in individual [Data Cards](https://huggingface.co/spaces/bigscience/BigScienceCorpus). Training data includes: - 45 natural languages - 12 programming languages - In 1.5TB of pre-processed text, converted into 350B unique tokens (see [the tokenizer section](#tokenization) for more.) ### Languages The pie chart shows the distribution of languages in training data. ![pie chart showing the distribution of languages in training data](https://github.com/bigscience-workshop/model_card/blob/main/assets/data/pie_chart.svg?raw=true) The following tables shows the further distribution of Niger-Congo & Indic languages and programming languages in the training data. Distribution of Niger Congo and Indic languages. | Niger Congo | Percentage | | Indic | Percentage | |----------------|------------ |------ |-----------|------------| | Chi Tumbuka | 0.00002 | | Assamese | 0.01 | | Kikuyu | 0.00004 | | Odia | 0.04 | | Bambara | 0.00004 | | Gujarati | 0.04 | | Akan | 0.00007 | | Marathi | 0.05 | | Xitsonga | 0.00007 | | Punjabi | 0.05 | | Sesotho | 0.00007 | | Kannada | 0.06 | | Chi Chewa | 0.0001 | | Nepali | 0.07 | | Setswana | 0.0002 | | Telugu | 0.09 | | Northern Sotho | 0.0002 | | Malayalam | 0.10 | | Fon | 0.0002 | | Urdu | 0.10 | | Kirundi | 0.0003 | | Tamil | 0.20 | | Wolof | 0.0004 | | Bengali | 0.50 | | Kuganda | 0.0004 | | Hindi | 0.70 | | Chi Shona | 0.001 | | Isi Zulu | 0.001 | | Igbo | 0.001 | | Xhosa | 0.001 | | Kinyarwanda | 0.003 | | Yoruba | 0.006 | | Swahili | 0.02 | Distribution of programming languages. | Extension | Language | Number of files | |----------------|------------|-----------------| | java | Java | 5,407,724 | | php | PHP | 4,942,186 | | cpp | C++ | 2,503,930 | | py | Python | 2,435,072 | | js | JavaScript | 1,905,518 | | cs | C# | 1,577,347 | | rb | Ruby | 6,78,413 | | cc | C++ | 443,054 | | hpp | C++ | 391,048 | | lua | Lua | 352,317 | | go | GO | 227,763 | | ts | TypeScript | 195,254 | | C | C | 134,537 | | scala | Scala | 92,052 | | hh | C++ | 67,161 | | H | C++ | 55,899 | | tsx | TypeScript | 33,107 | | rs | Rust | 29,693 | | phpt | PHP | 9,702 | | c++ | C++ | 1,342 | | h++ | C++ | 791 | | php3 | PHP | 540 | | phps | PHP | 270 | | php5 | PHP | 166 | | php4 | PHP | 29 | ### Preprocessing **Tokenization:** The BLOOM tokenizer ([link](https://huggingface.co/bigscience/tokenizer)), a learned subword tokenizer trained using: - A byte-level Byte Pair Encoding (BPE) algorithm - A simple pre-tokenization rule, no normalization - A vocabulary size of 250,680 It was trained on a subset of a preliminary version of the corpus using alpha-weighting per language. ## Speeds, Sizes, Times Training logs: [Tensorboard link](https://huggingface.co/tensorboard/bigscience/tr11-176B-ml-logs/) - Dates: - Started 11th March, 2022 11:42am PST - Estimated end: 5th July, 2022 - Checkpoint size: - Bf16 weights: 329GB - Full checkpoint with optimizer states: 2.3TB - Training throughput: About 150 TFLOP per GPU per second - Number of epochs: 1 - Estimated cost of training: Equivalent of $2-5M in cloud computing (including preliminary experiments) - Server training location: Île-de-France, France ## Environmental Impact The training supercomputer, Jean Zay ([website](http://www.idris.fr/eng/jean-zay/jean-zay-presentation-eng.html)), uses mostly nuclear energy. The heat generated by it is reused for heating campus housing. **Estimated carbon emissions:** *(Forthcoming.)* **Estimated electricity usage:** *(Forthcoming.)* </details> --- # Uses *This section addresses questions around how the model is intended to be used, discusses the foreseeable users of the model (including those affected by the model), and describes uses that are considered out of scope or misuse of the model.* *It is useful for anyone considering using the model or who is affected by the model.* <details> <summary>Click to expand</summary> ## Intended Use This model is being created in order to enable public research on large language models (LLMs). LLMs are intended to be used for language generation or as a pretrained base model that can be further fine-tuned for specific tasks. Use cases below are not exhaustive. ### Direct Use - Text generation - Exploring characteristics of language generated by a language model - Examples: Cloze tests, counterfactuals, generations with reframings ### Downstream Use - Tasks that leverage language models include: Information Extraction, Question Answering, Summarization ### Misuse and Out-of-scope Use *This section addresses what users ought not do with the model.* See the [BLOOM License](https://huggingface.co/spaces/bigscience/license), Attachment A, for detailed usage restrictions. The below list is non-exhaustive, but lists some easily foreseeable problematic use cases. #### Out-of-scope Uses Using the model in [high-stakes](#high-stakes) settings is out of scope for this model. The model is not designed for [critical decisions](#critical-decisions) nor uses with any material consequences on an individual's livelihood or wellbeing. The model outputs content that appears factual but is not correct. Out-of-scope Uses Include: - Usage in biomedical domains, political and legal domains, or finance domains - Usage for evaluating or scoring individuals, such as for employment, education, or credit - Applying the model for critical automatic decisions, generating factual content, creating reliable summaries, or generating predictions that must be correct #### Misuse Intentionally using the model for harm, violating [human rights](#human-rights), or other kinds of malicious activities, is a misuse of this model. This includes: - Spam generation - Disinformation and influence operations - Disparagement and defamation - Harassment and abuse - [Deception](#deception) - Unconsented impersonation and imitation - Unconsented surveillance - Generating content without attribution to the model, as specified in the [RAIL License, Use Restrictions](https://huggingface.co/spaces/bigscience/license) ## Intended Users ### Direct Users - General Public - Researchers - Students - Educators - Engineers/developers - Non-commercial entities - Community advocates, including human and civil rights groups ### Indirect Users - Users of derivatives created by Direct Users, such as those using software with an [intended use](#intended-use) - Users of [Derivatives of the Model, as described in the License](https://huggingface.co/spaces/bigscience/license) ### Others Affected (Parties Prenantes) - People and groups referred to by the LLM - People and groups exposed to outputs of, or decisions based on, the LLM - People and groups whose original work is included in the LLM </details> --- # Risks and Limitations *This section identifies foreseeable harms and misunderstandings.* <details> <summary>Click to expand</summary> Model may: - Overrepresent some viewpoints and underrepresent others - Contain stereotypes - Contain [personal information](#personal-data-and-information) - Generate: - Hateful, abusive, or violent language - Discriminatory or prejudicial language - Content that may not be appropriate for all settings, including sexual content - Make errors, including producing incorrect information as if it were factual - Generate irrelevant or repetitive outputs </details> --- # Evaluation *This section describes the evaluation protocols and provides the results.* <details> <summary>Click to expand</summary> ## Metrics *This section describes the different ways performance is calculated and why.* Includes: | Metric | Why chosen | |--------------------|--------------------------------------------------------------------| | [Perplexity](#perplexity) | Standard metric for quantifying model improvements during training | | Cross Entropy [Loss](#loss) | Standard objective for language models. | And multiple different metrics for specific tasks. _(More evaluation metrics forthcoming upon completion of evaluation protocol.)_ ## Factors *This section lists some different aspects of what BLOOM models. Its focus is on those aspects that are likely to give rise to high variance in model behavior.* - Language, such as English or Yoruba - Domain, such as newswire or stories - Demographic characteristics, such as gender or nationality ## Results *Results are based on the [Factors](#factors) and [Metrics](#metrics).* **Train-time Evaluation:** As of 25.May.2022, 15:00 PST: - Training Loss: 2.0 - Validation Loss: 2.2 - Perplexity: 8.9 (More evaluation scores forthcoming.) </details> --- # Recommendations *This section provides information on warnings and potential mitigations.* <details> <summary>Click to expand</summary> - Indirect users should be made aware when the content they're working with is created by the LLM. - Users should be aware of [Risks and Limitations](#risks-and-limitations), and include an appropriate age disclaimer or blocking interface as necessary. - Models trained or finetuned downstream of BLOOM LM should include an updated Model Card. - Users of the model should provide mechanisms for those affected to provide feedback, such as an email address for comments. </details> --- # Glossary and Calculations *This section defines common terms and how metrics are calculated.* <details> <summary>Click to expand</summary> - <a name="loss">**Loss:**</a> A calculation of the difference between what the model has learned and what the data shows ("groundtruth"). The lower the loss, the better. The training process aims to minimize the loss. - <a name="perplexity">**Perplexity:**</a> This is based on what the model estimates the probability of new data is. The lower the perplexity, the better. If the model is 100% correct at predicting the next token it will see, then the perplexity is 1. Mathematically this is calculated using entropy. - <a name="high-stakes">**High-stakes settings:**</a> Such as those identified as "high-risk AI systems" and "unacceptable risk AI systems" in the European Union's proposed [Artificial Intelligence (AI) Act](https://artificialintelligenceact.eu/annexes/). - <a name="critical-decisions">**Critical decisions:**</a> Such as those defined in [the United States' proposed Algorithmic Accountability Act](https://www.congress.gov/117/bills/s3572/BILLS-117s3572is.pdf). - <a name="human-rights">**Human rights:**</a> Includes those rights defined in the [Universal Declaration of Human Rights](https://www.un.org/sites/un2.un.org/files/2021/03/udhr.pdf). - <a name="personal-data-and-information">**Personal Data and Personal Information:**</a> Personal data and information is defined in multiple data protection regulations, such as "[personal data](https://gdpr-info.eu/issues/personal-data/)" in the [European Union's General Data Protection Regulation](https://gdpr-info.eu); and "personal information" in the Republic of South Africa's [Protection of Personal Information Act](https://www.gov.za/sites/default/files/gcis_document/201409/3706726-11act4of2013popi.pdf), The People's Republic of China's [Personal information protection law](http://en.npc.gov.cn.cdurl.cn/2021-12/29/c_694559.htm). - <a name="sensitive-characteristics">**Sensitive characteristics:**</a> This includes specifically protected categories in human rights (see [UHDR, Article 2](https://www.un.org/sites/un2.un.org/files/2021/03/udhr.pdf)) and personal information regulation (see GDPR, [Article 9; Protection of Personal Information Act, Chapter 1](https://www.gov.za/sites/default/files/gcis_document/201409/3706726-11act4of2013popi.pdf)) - <a name="deception">**Deception:**</a> Doing something to intentionally mislead individuals to believe something that is false, such as by creating deadbots or chatbots on social media posing as real people, or generating text documents without making consumers aware that the text is machine generated. </details> --- # More Information *This section provides links to writing on dataset creation, technical specifications, lessons learned, and initial results.* <details> <summary>Click to expand</summary> ## Dataset Creation Blog post detailing the design choices during the dataset creation: https://bigscience.huggingface.co/blog/building-a-tb-scale-multilingual-dataset-for-language-modeling ## Technical Specifications Blog post summarizing how the architecture, size, shape, and pre-training duration where selected: https://bigscience.huggingface.co/blog/what-language-model-to-train-if-you-have-two-million-gpu-hours More details on the architecture/optimizer: https://github.com/bigscience-workshop/bigscience/tree/master/train/tr11-176B-ml Blog post on the hardware/engineering side: https://bigscience.huggingface.co/blog/which-hardware-to-train-a-176b-parameters-model Details on the distributed setup used for the training: https://github.com/bigscience-workshop/bigscience/tree/master/train/tr11-176B-ml Tensorboard updated during the training: https://huggingface.co/bigscience/tr11-176B-ml-logs/tensorboard#scalars&tagFilter=loss ## Lessons Insights on how to approach training, negative results: https://github.com/bigscience-workshop/bigscience/blob/master/train/lessons-learned.md Details on the obstacles overcome during the preparation on the engineering side (instabilities, optimization of training throughput, so many technical tricks and questions): https://github.com/bigscience-workshop/bigscience/blob/master/train/tr11-176B-ml/chronicles.md ## Initial Results Initial prompting experiments using interim checkpoints: https://huggingface.co/spaces/bigscience/bloom-book </details> --- # Model Card Authors *Ordered roughly chronologically and by amount of time spent.* Margaret Mitchell, Giada Pistilli, Yacine Jernite, Ezinwanne Ozoani, Marissa Gerchick, Nazneen Rajani, Sasha Luccioni, Irene Solaiman, Maraim Masoud, Somaieh Nikpoor, Carlos Muñoz Ferrandis, Stas Bekman, Christopher Akiki, Danish Contractor, David Lansky, Angelina McMillan-Major, Tristan Thrush, Suzana Ilić, Gérard Dupont, Shayne Longpre, Manan Dey, Stella Biderman, Douwe Kiela, Emi Baylor, Teven Le Scao, Aaron Gokaslan, Julien Launay, Niklas Muennighoff

SHENMU007/neunit_BASE_V13.5.6

SHENMU007

2023-08-30T11:59:02Z

75

0

transformers

[ "transformers", "pytorch", "tensorboard", "speecht5", "text-to-audio", "1.1.0", "generated_from_trainer", "zh", "dataset:facebook/voxpopuli", "base_model:microsoft/speecht5_tts", "base_model:finetune:microsoft/speecht5_tts", "license:mit", "endpoints_compatible", "region:us" ]

text-to-audio

2023-08-30T07:34:28Z

--- language: - zh license: mit base_model: microsoft/speecht5_tts tags: - 1.1.0 - generated_from_trainer datasets: - facebook/voxpopuli model-index: - name: SpeechT5 TTS Dutch neunit results: [] ---  # SpeechT5 TTS Dutch neunit This model is a fine-tuned version of [microsoft/speecht5_tts](https://huggingface.co/microsoft/speecht5_tts) on the VoxPopuli dataset. ## Model description More information needed ## Intended uses & limitations More information needed ## Training and evaluation data More information needed ## Training procedure ### Training hyperparameters The following hyperparameters were used during training: - learning_rate: 1e-05 - train_batch_size: 8 - eval_batch_size: 8 - seed: 42 - gradient_accumulation_steps: 4 - total_train_batch_size: 32 - optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08 - lr_scheduler_type: linear - lr_scheduler_warmup_steps: 500 - training_steps: 4000 ### Training results ### Framework versions - Transformers 4.31.0.dev0 - Pytorch 2.0.1+cu117 - Datasets 2.12.0 - Tokenizers 0.13.3

cvapict/yhi-message-type-all-distilroberta-v1

cvapict

2023-08-30T11:45:48Z

5

0

sentence-transformers

[ "sentence-transformers", "pytorch", "roberta", "setfit", "text-classification", "arxiv:2209.11055", "license:apache-2.0", "region:us" ]

text-classification

2023-08-30T11:44:20Z

--- license: apache-2.0 tags: - setfit - sentence-transformers - text-classification pipeline_tag: text-classification --- # cvapict/yhi-message-type-all-distilroberta-v1 {'accuracy': 0.8373983739837398} This is a [SetFit model](https://github.com/huggingface/setfit) that can be used for text classification. The model has been trained using an efficient few-shot learning technique that involves: 1. Fine-tuning a [Sentence Transformer](https://www.sbert.net) with contrastive learning. 2. Training a classification head with features from the fine-tuned Sentence Transformer. ## Usage To use this model for inference, first install the SetFit library: ```bash python -m pip install setfit ``` You can then run inference as follows: ```python from setfit import SetFitModel # Download from Hub and run inference model = SetFitModel.from_pretrained("cvapict/yhi-message-type-all-distilroberta-v1") # Run inference preds = model(["i loved the spiderman movie!", "pineapple on pizza is the worst 🤮"]) ``` ## BibTeX entry and citation info ```bibtex @article{https://doi.org/10.48550/arxiv.2209.11055, doi = {10.48550/ARXIV.2209.11055}, url = {https://arxiv.org/abs/2209.11055}, author = {Tunstall, Lewis and Reimers, Nils and Jo, Unso Eun Seo and Bates, Luke and Korat, Daniel and Wasserblat, Moshe and Pereg, Oren}, keywords = {Computation and Language (cs.CL), FOS: Computer and information sciences, FOS: Computer and information sciences}, title = {Efficient Few-Shot Learning Without Prompts}, publisher = {arXiv}, year = {2022}, copyright = {Creative Commons Attribution 4.0 International} } ```

Shivam098/my_awesome_opus_books_model

Shivam098

2023-08-30T11:40:07Z

12

0

transformers

[ "transformers", "pytorch", "tensorboard", "safetensors", "t5", "text2text-generation", "generated_from_trainer", "dataset:opus100", "base_model:google-t5/t5-small", "base_model:finetune:google-t5/t5-small", "license:apache-2.0", "model-index", "autotrain_compatible", "text-generation-inference", "endpoints_compatible", "region:us" ]

text2text-generation

2023-08-24T11:36:29Z

--- license: apache-2.0 base_model: t5-small tags: - generated_from_trainer datasets: - opus100 metrics: - bleu model-index: - name: my_awesome_opus_books_model results: - task: name: Sequence-to-sequence Language Modeling type: text2text-generation dataset: name: opus100 type: opus100 config: en-ps split: train args: en-ps metrics: - name: Bleu type: bleu value: 8.2239 ---  # my_awesome_opus_books_model This model is a fine-tuned version of [t5-small](https://huggingface.co/t5-small) on the opus100 dataset. It achieves the following results on the evaluation set: - Loss: 3.1766 - Bleu: 8.2239 - Gen Len: 7.6785 ## Model description More information needed ## Intended uses & limitations More information needed ## Training and evaluation data More information needed ## Training procedure ### Training hyperparameters The following hyperparameters were used during training: - learning_rate: 2e-05 - train_batch_size: 16 - eval_batch_size: 16 - seed: 42 - optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08 - lr_scheduler_type: linear - num_epochs: 2 ### Training results | Training Loss | Epoch | Step | Validation Loss | Bleu | Gen Len | |:-------------:|:-----:|:----:|:---------------:|:------:|:-------:| | 3.5442 | 1.0 | 3957 | 3.2616 | 9.5837 | 6.9205 | | 3.3951 | 2.0 | 7914 | 3.1766 | 8.2239 | 7.6785 | ### Framework versions - Transformers 4.32.0 - Pytorch 2.0.1+cu118 - Datasets 2.14.4 - Tokenizers 0.13.3

cvapict/yhi-message-type-all-mpnet-base-v2

cvapict

2023-08-30T11:39:14Z

4

0

sentence-transformers

[ "sentence-transformers", "pytorch", "mpnet", "setfit", "text-classification", "arxiv:2209.11055", "license:apache-2.0", "region:us" ]

text-classification

2023-08-30T11:37:48Z

--- license: apache-2.0 tags: - setfit - sentence-transformers - text-classification pipeline_tag: text-classification --- # cvapict/yhi-message-type-all-mpnet-base-v2 {'accuracy': 0.8292682926829268} This is a [SetFit model](https://github.com/huggingface/setfit) that can be used for text classification. The model has been trained using an efficient few-shot learning technique that involves: 1. Fine-tuning a [Sentence Transformer](https://www.sbert.net) with contrastive learning. 2. Training a classification head with features from the fine-tuned Sentence Transformer. ## Usage To use this model for inference, first install the SetFit library: ```bash python -m pip install setfit ``` You can then run inference as follows: ```python from setfit import SetFitModel # Download from Hub and run inference model = SetFitModel.from_pretrained("cvapict/yhi-message-type-all-mpnet-base-v2") # Run inference preds = model(["i loved the spiderman movie!", "pineapple on pizza is the worst 🤮"]) ``` ## BibTeX entry and citation info ```bibtex @article{https://doi.org/10.48550/arxiv.2209.11055, doi = {10.48550/ARXIV.2209.11055}, url = {https://arxiv.org/abs/2209.11055}, author = {Tunstall, Lewis and Reimers, Nils and Jo, Unso Eun Seo and Bates, Luke and Korat, Daniel and Wasserblat, Moshe and Pereg, Oren}, keywords = {Computation and Language (cs.CL), FOS: Computer and information sciences, FOS: Computer and information sciences}, title = {Efficient Few-Shot Learning Without Prompts}, publisher = {arXiv}, year = {2022}, copyright = {Creative Commons Attribution 4.0 International} } ```

dkqjrm/20230830152959

dkqjrm

2023-08-30T11:34:30Z

105

0

transformers

[ "transformers", "pytorch", "tensorboard", "bert", "text-classification", "generated_from_trainer", "dataset:super_glue", "license:apache-2.0", "autotrain_compatible", "endpoints_compatible", "region:us" ]

text-classification

2023-08-30T06:30:26Z

--- license: apache-2.0 tags: - generated_from_trainer datasets: - super_glue metrics: - accuracy model-index: - name: '20230830152959' results: [] ---  # 20230830152959 This model is a fine-tuned version of [bert-large-cased](https://huggingface.co/bert-large-cased) on the super_glue dataset. It achieves the following results on the evaluation set: - Loss: 0.7271 - Accuracy: 0.5 ## Model description More information needed ## Intended uses & limitations More information needed ## Training and evaluation data More information needed ## Training procedure ### Training hyperparameters The following hyperparameters were used during training: - learning_rate: 0.0007 - train_batch_size: 16 - eval_batch_size: 8 - seed: 11 - optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08 - lr_scheduler_type: linear - num_epochs: 80.0 ### Training results | Training Loss | Epoch | Step | Validation Loss | Accuracy | |:-------------:|:-----:|:-----:|:---------------:|:--------:| | No log | 1.0 | 340 | 0.7372 | 0.5 | | 0.7583 | 2.0 | 680 | 0.7264 | 0.5 | | 0.7541 | 3.0 | 1020 | 0.7485 | 0.5 | | 0.7541 | 4.0 | 1360 | 0.7304 | 0.5 | | 0.7509 | 5.0 | 1700 | 0.7363 | 0.5 | | 0.7455 | 6.0 | 2040 | 0.7360 | 0.5 | | 0.7455 | 7.0 | 2380 | 0.7240 | 0.5 | | 0.7493 | 8.0 | 2720 | 0.7806 | 0.5 | | 0.7515 | 9.0 | 3060 | 0.7284 | 0.5 | | 0.7515 | 10.0 | 3400 | 0.7375 | 0.5 | | 0.7489 | 11.0 | 3740 | 0.7340 | 0.5 | | 0.7452 | 12.0 | 4080 | 0.7255 | 0.5 | | 0.7452 | 13.0 | 4420 | 0.8158 | 0.5 | | 0.7463 | 14.0 | 4760 | 0.7256 | 0.5 | | 0.75 | 15.0 | 5100 | 0.7364 | 0.5 | | 0.75 | 16.0 | 5440 | 0.7260 | 0.5 | | 0.7411 | 17.0 | 5780 | 0.7320 | 0.5 | | 0.7482 | 18.0 | 6120 | 0.7296 | 0.5 | | 0.7482 | 19.0 | 6460 | 0.7312 | 0.5 | | 0.7424 | 20.0 | 6800 | 0.7405 | 0.5 | | 0.7374 | 21.0 | 7140 | 0.7258 | 0.5 | | 0.7374 | 22.0 | 7480 | 0.7609 | 0.5 | | 0.7388 | 23.0 | 7820 | 0.7608 | 0.5 | | 0.7382 | 24.0 | 8160 | 0.7239 | 0.5 | | 0.7385 | 25.0 | 8500 | 0.7315 | 0.5 | | 0.7385 | 26.0 | 8840 | 0.7472 | 0.5 | | 0.7392 | 27.0 | 9180 | 0.7863 | 0.5 | | 0.737 | 28.0 | 9520 | 0.7261 | 0.5 | | 0.737 | 29.0 | 9860 | 0.7403 | 0.5 | | 0.7322 | 30.0 | 10200 | 0.7245 | 0.5 | | 0.7359 | 31.0 | 10540 | 0.7239 | 0.5 | | 0.7359 | 32.0 | 10880 | 0.7555 | 0.5 | | 0.7368 | 33.0 | 11220 | 0.7239 | 0.5 | | 0.7349 | 34.0 | 11560 | 0.7380 | 0.5 | | 0.7349 | 35.0 | 11900 | 0.7279 | 0.5 | | 0.7367 | 36.0 | 12240 | 0.7263 | 0.5 | | 0.7343 | 37.0 | 12580 | 0.7252 | 0.5 | | 0.7343 | 38.0 | 12920 | 0.7299 | 0.5 | | 0.7347 | 39.0 | 13260 | 0.7344 | 0.5 | | 0.7315 | 40.0 | 13600 | 0.7247 | 0.5 | | 0.7315 | 41.0 | 13940 | 0.7277 | 0.5 | | 0.7324 | 42.0 | 14280 | 0.7246 | 0.5 | | 0.7319 | 43.0 | 14620 | 0.7289 | 0.5 | | 0.7319 | 44.0 | 14960 | 0.7297 | 0.5 | | 0.7321 | 45.0 | 15300 | 0.7389 | 0.5 | | 0.7304 | 46.0 | 15640 | 0.7245 | 0.5 | | 0.7304 | 47.0 | 15980 | 0.7306 | 0.5 | | 0.7316 | 48.0 | 16320 | 0.7239 | 0.5 | | 0.7325 | 49.0 | 16660 | 0.7312 | 0.5 | | 0.7311 | 50.0 | 17000 | 0.7241 | 0.5 | | 0.7311 | 51.0 | 17340 | 0.7239 | 0.5 | | 0.7292 | 52.0 | 17680 | 0.7244 | 0.5 | | 0.732 | 53.0 | 18020 | 0.7250 | 0.5 | | 0.732 | 54.0 | 18360 | 0.7239 | 0.5 | | 0.727 | 55.0 | 18700 | 0.7257 | 0.5 | | 0.7277 | 56.0 | 19040 | 0.7247 | 0.5 | | 0.7277 | 57.0 | 19380 | 0.7256 | 0.5 | | 0.7296 | 58.0 | 19720 | 0.7269 | 0.5 | | 0.7284 | 59.0 | 20060 | 0.7340 | 0.5 | | 0.7284 | 60.0 | 20400 | 0.7257 | 0.5 | | 0.7269 | 61.0 | 20740 | 0.7254 | 0.5 | | 0.7266 | 62.0 | 21080 | 0.7240 | 0.5 | | 0.7266 | 63.0 | 21420 | 0.7245 | 0.5 | | 0.7274 | 64.0 | 21760 | 0.7261 | 0.5 | | 0.7278 | 65.0 | 22100 | 0.7240 | 0.5 | | 0.7278 | 66.0 | 22440 | 0.7492 | 0.5 | | 0.7252 | 67.0 | 22780 | 0.7254 | 0.5 | | 0.727 | 68.0 | 23120 | 0.7241 | 0.5 | | 0.727 | 69.0 | 23460 | 0.7244 | 0.5 | | 0.7253 | 70.0 | 23800 | 0.7242 | 0.5 | | 0.7249 | 71.0 | 24140 | 0.7247 | 0.5 | | 0.7249 | 72.0 | 24480 | 0.7280 | 0.5 | | 0.7244 | 73.0 | 24820 | 0.7267 | 0.5 | | 0.7232 | 74.0 | 25160 | 0.7284 | 0.5 | | 0.7234 | 75.0 | 25500 | 0.7272 | 0.5 | | 0.7234 | 76.0 | 25840 | 0.7294 | 0.5 | | 0.7233 | 77.0 | 26180 | 0.7257 | 0.5 | | 0.7222 | 78.0 | 26520 | 0.7284 | 0.5 | | 0.7222 | 79.0 | 26860 | 0.7276 | 0.5 | | 0.7225 | 80.0 | 27200 | 0.7271 | 0.5 | ### Framework versions - Transformers 4.26.1 - Pytorch 2.0.1+cu118 - Datasets 2.12.0 - Tokenizers 0.13.3

eliept1/ppo-Pyramids

eliept1

2023-08-30T11:24:47Z

0

ml-agents

[ "ml-agents", "tensorboard", "onnx", "Pyramids", "deep-reinforcement-learning", "reinforcement-learning", "ML-Agents-Pyramids", "region:us" ]

reinforcement-learning

2023-08-30T11:24:21Z

--- library_name: ml-agents tags: - Pyramids - deep-reinforcement-learning - reinforcement-learning - ML-Agents-Pyramids --- # **ppo** Agent playing **Pyramids** This is a trained model of a **ppo** agent playing **Pyramids** using the [Unity ML-Agents Library](https://github.com/Unity-Technologies/ml-agents). ## Usage (with ML-Agents) The Documentation: https://unity-technologies.github.io/ml-agents/ML-Agents-Toolkit-Documentation/ We wrote a complete tutorial to learn to train your first agent using ML-Agents and publish it to the Hub: - A *short tutorial* where you teach Huggy the Dog 🐶 to fetch the stick and then play with him directly in your browser: https://huggingface.co/learn/deep-rl-course/unitbonus1/introduction - A *longer tutorial* to understand how works ML-Agents: https://huggingface.co/learn/deep-rl-course/unit5/introduction ### Resume the training ```bash mlagents-learn <your_configuration_file_path.yaml> --run-id=<run_id> --resume ``` ### Watch your Agent play You can watch your agent **playing directly in your browser** 1. If the environment is part of ML-Agents official environments, go to https://huggingface.co/unity 2. Step 1: Find your model_id: eliept1/ppo-Pyramids 3. Step 2: Select your *.nn /*.onnx file 4. Click on Watch the agent play 👀

pumaML/turkishReviews-ds-mini

pumaML

2023-08-30T11:21:06Z

61

0

transformers

[ "transformers", "tf", "gpt2", "text-generation", "generated_from_keras_callback", "base_model:openai-community/gpt2", "base_model:finetune:openai-community/gpt2", "license:mit", "autotrain_compatible", "endpoints_compatible", "region:us" ]

text-generation

2023-08-30T11:15:23Z

--- license: mit base_model: gpt2 tags: - generated_from_keras_callback model-index: - name: pumaML/turkishReviews-ds-mini results: [] ---  # pumaML/turkishReviews-ds-mini This model is a fine-tuned version of [gpt2](https://huggingface.co/gpt2) on an unknown dataset. It achieves the following results on the evaluation set: - Train Loss: 9.1942 - Validation Loss: 9.2669 - Epoch: 2 ## Model description More information needed ## Intended uses & limitations More information needed ## Training and evaluation data More information needed ## Training procedure ### Training hyperparameters The following hyperparameters were used during training: - optimizer: {'name': 'AdamWeightDecay', 'learning_rate': {'class_name': 'WarmUp', 'config': {'initial_learning_rate': 5e-05, 'decay_schedule_fn': {'class_name': 'PolynomialDecay', 'config': {'initial_learning_rate': 5e-05, 'decay_steps': -896, 'end_learning_rate': 0.0, 'power': 1.0, 'cycle': False, 'name': None}, '__passive_serialization__': True}, 'warmup_steps': 1000, 'power': 1.0, 'name': None}}, 'decay': 0.0, 'beta_1': 0.9, 'beta_2': 0.999, 'epsilon': 1e-08, 'amsgrad': False, 'weight_decay_rate': 0.01} - training_precision: mixed_float16 ### Training results | Train Loss | Validation Loss | Epoch | |:----------:|:---------------:|:-----:| | 10.3292 | 9.9955 | 0 | | 9.6809 | 9.6477 | 1 | | 9.1942 | 9.2669 | 2 | ### Framework versions - Transformers 4.32.1 - TensorFlow 2.12.0 - Datasets 2.14.4 - Tokenizers 0.13.3

anshulsc/q-Taxi-v3

anshulsc

2023-08-30T11:17:54Z

0

null

[ "Taxi-v3", "q-learning", "reinforcement-learning", "custom-implementation", "model-index", "region:us" ]

reinforcement-learning

2023-08-30T11:17:52Z

--- tags: - Taxi-v3 - q-learning - reinforcement-learning - custom-implementation model-index: - name: q-Taxi-v3 results: - task: type: reinforcement-learning name: reinforcement-learning dataset: name: Taxi-v3 type: Taxi-v3 metrics: - type: mean_reward value: 7.56 +/- 2.71 name: mean_reward verified: false --- # **Q-Learning** Agent playing1 **Taxi-v3** This is a trained model of a **Q-Learning** agent playing **Taxi-v3** . ## Usage ```python model = load_from_hub(repo_id="anshulsc/q-Taxi-v3", filename="q-learning.pkl") # Don't forget to check if you need to add additional attributes (is_slippery=False etc) env = gym.make(model["env_id"]) ```

anshulsc/q-FrozenLake-v1-4x4-noSlippery

anshulsc

2023-08-30T11:11:13Z

0

null

[ "FrozenLake-v1-4x4-no_slippery", "q-learning", "reinforcement-learning", "custom-implementation", "model-index", "region:us" ]

reinforcement-learning

2023-08-30T11:06:45Z

--- tags: - FrozenLake-v1-4x4-no_slippery - q-learning - reinforcement-learning - custom-implementation model-index: - name: q-FrozenLake-v1-4x4-noSlippery results: - task: type: reinforcement-learning name: reinforcement-learning dataset: name: FrozenLake-v1-4x4-no_slippery type: FrozenLake-v1-4x4-no_slippery metrics: - type: mean_reward value: 1.00 +/- 0.00 name: mean_reward verified: false --- # **Q-Learning** Agent playing1 **FrozenLake-v1** This is a trained model of a **Q-Learning** agent playing **FrozenLake-v1** . ## Usage ```python model = load_from_hub(repo_id="anshulsc/q-FrozenLake-v1-4x4-noSlippery", filename="q-learning.pkl") # Don't forget to check if you need to add additional attributes (is_slippery=False etc) env = gym.make(model["env_id"]) ```

unrealheinrich/vegetables

unrealheinrich

2023-08-30T11:07:16Z

0

null

[ "onnx", "en", "license:mit", "region:us" ]

null

2023-08-30T11:02:02Z

--- license: mit language: - en metrics: - f1 ---

dkqjrm/20230830151725

dkqjrm

2023-08-30T11:06:41Z

104

0

transformers

[ "transformers", "pytorch", "tensorboard", "bert", "text-classification", "generated_from_trainer", "dataset:super_glue", "license:apache-2.0", "autotrain_compatible", "endpoints_compatible", "region:us" ]

text-classification

2023-08-30T06:17:54Z

--- license: apache-2.0 tags: - generated_from_trainer datasets: - super_glue metrics: - accuracy model-index: - name: '20230830151725' results: [] ---  # 20230830151725 This model is a fine-tuned version of [bert-large-cased](https://huggingface.co/bert-large-cased) on the super_glue dataset. It achieves the following results on the evaluation set: - Loss: 0.7242 - Accuracy: 0.5 ## Model description More information needed ## Intended uses & limitations More information needed ## Training and evaluation data More information needed ## Training procedure ### Training hyperparameters The following hyperparameters were used during training: - learning_rate: 0.0005 - train_batch_size: 16 - eval_batch_size: 8 - seed: 11 - optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08 - lr_scheduler_type: linear - num_epochs: 80.0 ### Training results | Training Loss | Epoch | Step | Validation Loss | Accuracy | |:-------------:|:-----:|:-----:|:---------------:|:--------:| | No log | 1.0 | 340 | 0.7462 | 0.4984 | | 0.7527 | 2.0 | 680 | 0.7272 | 0.5047 | | 0.7493 | 3.0 | 1020 | 0.7466 | 0.5 | | 0.7493 | 4.0 | 1360 | 0.7254 | 0.5 | | 0.7484 | 5.0 | 1700 | 0.7484 | 0.5 | | 0.7446 | 6.0 | 2040 | 0.7319 | 0.5 | | 0.7446 | 7.0 | 2380 | 0.7327 | 0.5 | | 0.7385 | 8.0 | 2720 | 0.7423 | 0.5 | | 0.7404 | 9.0 | 3060 | 0.7351 | 0.4875 | | 0.7404 | 10.0 | 3400 | 0.7567 | 0.5 | | 0.7576 | 11.0 | 3740 | 0.7305 | 0.5 | | 0.745 | 12.0 | 4080 | 0.7374 | 0.5 | | 0.745 | 13.0 | 4420 | 0.7735 | 0.5 | | 0.7346 | 14.0 | 4760 | 0.7443 | 0.5 | | 0.7407 | 15.0 | 5100 | 0.7278 | 0.5 | | 0.7407 | 16.0 | 5440 | 0.7490 | 0.5 | | 0.7367 | 17.0 | 5780 | 0.7443 | 0.5 | | 0.7377 | 18.0 | 6120 | 0.7330 | 0.5 | | 0.7377 | 19.0 | 6460 | 0.7249 | 0.5 | | 0.7325 | 20.0 | 6800 | 0.7362 | 0.5 | | 0.7357 | 21.0 | 7140 | 0.7240 | 0.5 | | 0.7357 | 22.0 | 7480 | 0.7470 | 0.5 | | 0.7356 | 23.0 | 7820 | 0.7296 | 0.5 | | 0.734 | 24.0 | 8160 | 0.7255 | 0.5 | | 0.7327 | 25.0 | 8500 | 0.7249 | 0.5 | | 0.7327 | 26.0 | 8840 | 0.7262 | 0.5 | | 0.7345 | 27.0 | 9180 | 0.7743 | 0.5 | | 0.735 | 28.0 | 9520 | 0.7276 | 0.5 | | 0.735 | 29.0 | 9860 | 0.7271 | 0.5 | | 0.7309 | 30.0 | 10200 | 0.7239 | 0.5 | | 0.7345 | 31.0 | 10540 | 0.7241 | 0.5 | | 0.7345 | 32.0 | 10880 | 0.7599 | 0.5 | | 0.733 | 33.0 | 11220 | 0.7245 | 0.5 | | 0.7325 | 34.0 | 11560 | 0.7385 | 0.5 | | 0.7325 | 35.0 | 11900 | 0.7245 | 0.5 | | 0.7318 | 36.0 | 12240 | 0.7242 | 0.5 | | 0.7308 | 37.0 | 12580 | 0.7239 | 0.5 | | 0.7308 | 38.0 | 12920 | 0.7241 | 0.5 | | 0.73 | 39.0 | 13260 | 0.7317 | 0.5 | | 0.7288 | 40.0 | 13600 | 0.7258 | 0.5 | | 0.7288 | 41.0 | 13940 | 0.7241 | 0.5 | | 0.7311 | 42.0 | 14280 | 0.7241 | 0.5 | | 0.7284 | 43.0 | 14620 | 0.7344 | 0.5 | | 0.7284 | 44.0 | 14960 | 0.7297 | 0.5 | | 0.73 | 45.0 | 15300 | 0.7393 | 0.5 | | 0.7269 | 46.0 | 15640 | 0.7239 | 0.5 | | 0.7269 | 47.0 | 15980 | 0.7282 | 0.5 | | 0.7284 | 48.0 | 16320 | 0.7240 | 0.5 | | 0.729 | 49.0 | 16660 | 0.7343 | 0.5 | | 0.7281 | 50.0 | 17000 | 0.7240 | 0.5 | | 0.7281 | 51.0 | 17340 | 0.7245 | 0.5 | | 0.7264 | 52.0 | 17680 | 0.7291 | 0.5 | | 0.7294 | 53.0 | 18020 | 0.7255 | 0.5 | | 0.7294 | 54.0 | 18360 | 0.7251 | 0.5 | | 0.7263 | 55.0 | 18700 | 0.7256 | 0.5 | | 0.7255 | 56.0 | 19040 | 0.7294 | 0.5 | | 0.7255 | 57.0 | 19380 | 0.7242 | 0.5 | | 0.7261 | 58.0 | 19720 | 0.7243 | 0.5 | | 0.7265 | 59.0 | 20060 | 0.7315 | 0.5 | | 0.7265 | 60.0 | 20400 | 0.7239 | 0.5 | | 0.7253 | 61.0 | 20740 | 0.7246 | 0.5 | | 0.7265 | 62.0 | 21080 | 0.7248 | 0.5 | | 0.7265 | 63.0 | 21420 | 0.7244 | 0.5 | | 0.7247 | 64.0 | 21760 | 0.7250 | 0.5 | | 0.7273 | 65.0 | 22100 | 0.7240 | 0.5 | | 0.7273 | 66.0 | 22440 | 0.7251 | 0.5 | | 0.7233 | 67.0 | 22780 | 0.7239 | 0.5 | | 0.7268 | 68.0 | 23120 | 0.7239 | 0.5 | | 0.7268 | 69.0 | 23460 | 0.7251 | 0.5 | | 0.7246 | 70.0 | 23800 | 0.7241 | 0.5 | | 0.7249 | 71.0 | 24140 | 0.7242 | 0.5 | | 0.7249 | 72.0 | 24480 | 0.7259 | 0.5 | | 0.7241 | 73.0 | 24820 | 0.7247 | 0.5 | | 0.7237 | 74.0 | 25160 | 0.7257 | 0.5 | | 0.7244 | 75.0 | 25500 | 0.7245 | 0.5 | | 0.7244 | 76.0 | 25840 | 0.7251 | 0.5 | | 0.7234 | 77.0 | 26180 | 0.7240 | 0.5 | | 0.7231 | 78.0 | 26520 | 0.7243 | 0.5 | | 0.7231 | 79.0 | 26860 | 0.7243 | 0.5 | | 0.7235 | 80.0 | 27200 | 0.7242 | 0.5 | ### Framework versions - Transformers 4.26.1 - Pytorch 2.0.1+cu118 - Datasets 2.12.0 - Tokenizers 0.13.3

Pemaled/FraudGuard

Pemaled

2023-08-30T11:04:27Z

0

null

[ "license:creativeml-openrail-m", "region:us" ]

null

2023-08-30T11:04:27Z

--- license: creativeml-openrail-m ---

dhiruHF/llama2-docqa-v2-downloaded

dhiruHF

2023-08-30T11:02:44Z

12

0

peft

[ "peft", "autotrain", "text-generation", "region:us" ]

text-generation

2023-08-30T11:02:42Z

--- tags: - autotrain - text-generation widget: - text: "I love AutoTrain because " library_name: peft --- # Model Trained Using AutoTrain## Training procedure The following `bitsandbytes` quantization config was used during training: - load_in_8bit: False - load_in_4bit: True - llm_int8_threshold: 6.0 - llm_int8_skip_modules: None - llm_int8_enable_fp32_cpu_offload: False - llm_int8_has_fp16_weight: False - bnb_4bit_quant_type: nf4 - bnb_4bit_use_double_quant: True - bnb_4bit_compute_dtype: bfloat16 ### Framework versions - PEFT 0.4.0

aggie/edos-alpaca-option-llama-demo

aggie

2023-08-30T10:59:20Z

1

0

peft

[ "peft", "region:us" ]

null

2023-08-18T16:51:22Z

--- library_name: peft --- ## Training procedure ### Framework versions - PEFT 0.5.0.dev0

aggie/edos-alpaca-label-llama-demo

aggie

2023-08-30T10:58:07Z

0

peft

[ "peft", "region:us" ]

null

2023-08-30T10:48:38Z

--- library_name: peft --- ## Training procedure ### Framework versions - PEFT 0.5.0.dev0

aggie/edos-alpaca-label-llama

aggie

2023-08-30T10:57:38Z

2

0

peft

[ "peft", "region:us" ]

null

2023-08-30T10:48:11Z

--- library_name: peft --- ## Training procedure ### Framework versions - PEFT 0.5.0.dev0

ShekDass/donut-base-cord-hifi-100

ShekDass

2023-08-30T10:53:16Z

3

0

transformers

[ "transformers", "pytorch", "vision-encoder-decoder", "image-text-to-text", "generated_from_trainer", "dataset:imagefolder", "base_model:naver-clova-ix/donut-base-finetuned-cord-v2", "base_model:finetune:naver-clova-ix/donut-base-finetuned-cord-v2", "license:mit", "endpoints_compatible", "region:us" ]

image-text-to-text

2023-08-29T14:01:24Z

--- license: mit base_model: naver-clova-ix/donut-base-finetuned-cord-v2 tags: - generated_from_trainer datasets: - imagefolder model-index: - name: donut-base-cord-hifi-100 results: [] ---  # donut-base-cord-hifi-100 This model is a fine-tuned version of [naver-clova-ix/donut-base-finetuned-cord-v2](https://huggingface.co/naver-clova-ix/donut-base-finetuned-cord-v2) on the imagefolder dataset. ## Model description More information needed ## Intended uses & limitations More information needed ## Training and evaluation data More information needed ## Training procedure ### Training hyperparameters The following hyperparameters were used during training: - learning_rate: 2e-05 - train_batch_size: 1 - eval_batch_size: 8 - seed: 42 - optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08 - lr_scheduler_type: linear - num_epochs: 20 ### Training results ### Framework versions - Transformers 4.32.1 - Pytorch 2.0.1+cu118 - Datasets 2.14.4 - Tokenizers 0.13.3

Snowad/French-Tortoise

Snowad

2023-08-30T10:42:57Z

0

28

null

[ "TTS", "text-to-speech", "fr", "license:apache-2.0", "region:us" ]

text-to-speech

2023-03-04T18:48:14Z

--- license: apache-2.0 language: - fr pipeline_tag: text-to-speech tags: - TTS - text-to-speech --- **V2.5 Model :** Fine tune of my V2 model on all CommonVoice dataset (517k sample) on 2.5k step (batch size 200), Voice cloning has improved a bit but is still not great. However, if you fine tune this model on your own personality dataset then you can get pretty good results. A good V3 model would be to fine tune for like 50k steps on this dataset and I think there would be a way to get good results but I won't try **V2 Model :** Tortoise base model Fine tuned on a custom multispeaker French dataset of 120k samples (SIWIS + Common Voice subset + M-AILABS) on 10k step with a RTX 3090 (~= 21 hours of training), with Text LR Weight at 1 Result : The model can speak French much better without an English accent but the voice clone hardly works **V1 Model :** Tortoise base model Fine tuned on a custom multispeaker French dataset of 24k samples (SIWIS + Common Voice subset) on 8850 step with a RTX 3090 (~= 19 hours of training) **Inference :** * You can use the model by downloading the "V2_9750_gpt.pth" model and use it in the tortoise-tts optimized forks (git.ecker.tech/mrq/ai-voice-cloning | 152334H/tortoise-tts-fast) **Fine tuning :** * I used 152334H/DL-Art-School for training, if you want to resume training from my epoch, follow its documentation and download "V2_9750.state"

jankovicsandras/Llama-2-13b-chat-norwegian-Q5_K_M-GGUF

jankovicsandras

2023-08-30T10:42:29Z

4

0

null

[ "gguf", "no", "nb", "nn", "norsk", "norwegian", "llama2", "q5_k_m", "en", "license:other", "endpoints_compatible", "region:us" ]

null

2023-08-30T07:06:18Z

--- language: - no - nb - nn - en tags: - no - nb - nn - norsk - norwegian - llama2 - gguf - q5_k_m license: "other" --- # This is the Q5_K_M quantized GGUF format ( see https://github.com/ggerganov/llama.cpp ) version of this model: https://huggingface.co/RuterNorway/Llama-2-13b-chat-norwegian Description (copied and edited from RuterNorway/Llama-2-13b-chat-norwegian) : # Llama 2 13b Chat Norwegian Llama-2-13b-chat-norwegian is a variant of Meta´s Llama 2 13b Chat model, finetuned on a mix of norwegian datasets created in Ruter AI Lab the summer of 2023. The model is tuned to understand and generate text in Norwegian. It's trained for one epoch on norwegian-alpaca + 15000 samples of machine-translated data from OpenOrca. A small subset of custom-made instructional data is also included. For other versions of this model see: https://huggingface.co/RuterNorway/Llama-2-13b-chat-norwegian ## Data - Norwegian alpaca - 15k Norwegian OpenOrcra (to be released) - Small subset of custom made instructional data ## Limitations - This is an LLM, not a knowledge model. It can not be expected to have more information about Norway than the basemodel. - It will generally preform better on tasks that involves summarization, question answering and chat, than on tasks that requires more knowledge about Norway, specific domains, or tasks where the model can answer freely. - The data used for training is machine translated, and may contain grammatical errors and other errors. - The model is released as is, and would in most cases need prompt tuning to achieve optimal results. ## License Llama 2 is licensed under the LLAMA 2 Community License, Copyright © Meta Platforms, Inc. All Rights Reserved. See the original model card for more information. From norwegian-alpaca we also note that "the current version uses OpenAI's gpt-3.5-turbo; hence, this dataset cannot be used to create models that compete in any way against OpenAI." Disclaimer ## The model is available "as is". Ruter As takes no responsibility for further use. ## During testing, it seems that the safeguards implemented by Meta, still work as expected in this model. However, we want to point to the Ethical Considerations and Limitations from the origenal model card: Llama 2 is a new technology that carries risks with use. Testing conducted to date has been in English, and has not covered, nor could it cover all scenarios. For these reasons, as with all LLMs, Llama 2’s potential outputs cannot be predicted in advance, and the model may in some instances produce inaccurate, biased or other objectionable responses to user prompts. Therefore, before deploying any applications of Llama 2, developers should perform safety testing and tuning tailored to their specific applications of the model. Please see the Responsible Use Guide available at https://ai.meta.com/llama/responsible-use-guide/ ## Credits This model was made at Ruters AI Lab which is a part of Ruters Data & AI division.

1Poireau/Doully_Millet

1Poireau

2023-08-30T10:23:36Z

0

null

[ "license:openrail++", "region:us" ]

null

2023-08-30T10:07:32Z

--- license: openrail++ --- La voix de Doully Millet du Groland, créé par 1Poireau le 30/08/2023. Entrainé pendant 8H sur 31min d'audio que j'ai trouvé sur la chaine Youtube officielle du Groland, avec 300 epochs The voice of Doully Millet from Groland, created by 1Poireau on 30/08/2023. Trained for 8H on 31min audio that i found on Groland's official Youtube channel, with 300 epochs

vishnuhaasan/ppo-lunarlanderv2

vishnuhaasan

2023-08-30T10:20:21Z

0

stable-baselines3

[ "stable-baselines3", "LunarLander-v2", "deep-reinforcement-learning", "reinforcement-learning", "model-index", "region:us" ]

reinforcement-learning

2023-08-30T10:19:45Z

--- library_name: stable-baselines3 tags: - LunarLander-v2 - deep-reinforcement-learning - reinforcement-learning - stable-baselines3 model-index: - name: PPO results: - task: type: reinforcement-learning name: reinforcement-learning dataset: name: LunarLander-v2 type: LunarLander-v2 metrics: - type: mean_reward value: 206.33 +/- 94.63 name: mean_reward verified: false --- # **PPO** Agent playing **LunarLander-v2** This is a trained model of a **PPO** agent playing **LunarLander-v2** using the [stable-baselines3 library](https://github.com/DLR-RM/stable-baselines3). ## Usage (with Stable-baselines3) TODO: Add your code ```python from stable_baselines3 import ... from huggingface_sb3 import load_from_hub ... ```

hoang14/chatbot_29_8

hoang14

2023-08-30T10:10:02Z

0

null

[ "region:us" ]

null

2023-08-29T02:57:40Z

ATASET = "task-focus + sample from remain datasets" DATASET_FORMAT = 'input-output' PER_DEVICE_TRAIN_BATCH_SIZE = 2 GRADIENT_ACCUMULATION_STEPS = 4 LEARNING_RATE = 0.0003 LR_SCHEDULER_TYPE = 'cosine' WARMUP_RATIO = 0.03 LORA_R = 192 LORA_ALPHA = 64 LORA_DROPOUT = 0.1 TRAIN_ON_SOURCE = False SOURCE_MAX_LENGTH = 1024 TARGET_MAX_LENGTH = 1024 LOGGING_STEPS = 20 SAVE_STEPS = 100 SAVE_TOTAL_LIMIT = 4

eliept1/ppo-SnowballTarget

eliept1

2023-08-30T10:07:23Z

0

ml-agents

[ "ml-agents", "tensorboard", "onnx", "SnowballTarget", "deep-reinforcement-learning", "reinforcement-learning", "ML-Agents-SnowballTarget", "region:us" ]

reinforcement-learning

2023-08-30T10:07:19Z

--- library_name: ml-agents tags: - SnowballTarget - deep-reinforcement-learning - reinforcement-learning - ML-Agents-SnowballTarget --- # **ppo** Agent playing **SnowballTarget** This is a trained model of a **ppo** agent playing **SnowballTarget** using the [Unity ML-Agents Library](https://github.com/Unity-Technologies/ml-agents). ## Usage (with ML-Agents) The Documentation: https://unity-technologies.github.io/ml-agents/ML-Agents-Toolkit-Documentation/ We wrote a complete tutorial to learn to train your first agent using ML-Agents and publish it to the Hub: - A *short tutorial* where you teach Huggy the Dog 🐶 to fetch the stick and then play with him directly in your browser: https://huggingface.co/learn/deep-rl-course/unitbonus1/introduction - A *longer tutorial* to understand how works ML-Agents: https://huggingface.co/learn/deep-rl-course/unit5/introduction ### Resume the training ```bash mlagents-learn <your_configuration_file_path.yaml> --run-id=<run_id> --resume ``` ### Watch your Agent play You can watch your agent **playing directly in your browser** 1. If the environment is part of ML-Agents official environments, go to https://huggingface.co/unity 2. Step 1: Find your model_id: eliept1/ppo-SnowballTarget 3. Step 2: Select your *.nn /*.onnx file 4. Click on Watch the agent play 👀

julio030/ppo-Huggy

julio030

2023-08-30T10:02:14Z

3

0

ml-agents

[ "ml-agents", "tensorboard", "onnx", "Huggy", "deep-reinforcement-learning", "reinforcement-learning", "ML-Agents-Huggy", "region:us" ]

reinforcement-learning

2023-08-30T10:02:04Z

--- library_name: ml-agents tags: - Huggy - deep-reinforcement-learning - reinforcement-learning - ML-Agents-Huggy --- # **ppo** Agent playing **Huggy** This is a trained model of a **ppo** agent playing **Huggy** using the [Unity ML-Agents Library](https://github.com/Unity-Technologies/ml-agents). ## Usage (with ML-Agents) The Documentation: https://unity-technologies.github.io/ml-agents/ML-Agents-Toolkit-Documentation/ We wrote a complete tutorial to learn to train your first agent using ML-Agents and publish it to the Hub: - A *short tutorial* where you teach Huggy the Dog 🐶 to fetch the stick and then play with him directly in your browser: https://huggingface.co/learn/deep-rl-course/unitbonus1/introduction - A *longer tutorial* to understand how works ML-Agents: https://huggingface.co/learn/deep-rl-course/unit5/introduction ### Resume the training ```bash mlagents-learn <your_configuration_file_path.yaml> --run-id=<run_id> --resume ``` ### Watch your Agent play You can watch your agent **playing directly in your browser** 1. If the environment is part of ML-Agents official environments, go to https://huggingface.co/unity 2. Step 1: Find your model_id: julio030/ppo-Huggy 3. Step 2: Select your *.nn /*.onnx file 4. Click on Watch the agent play 👀

ibrahimciko/q-taxi-v3

ibrahimciko

2023-08-30T09:44:46Z

0

null

[ "Taxi-v3", "q-learning", "reinforcement-learning", "custom-implementation", "model-index", "region:us" ]

reinforcement-learning

2023-08-30T09:44:41Z

--- tags: - Taxi-v3 - q-learning - reinforcement-learning - custom-implementation model-index: - name: q-taxi-v3 results: - task: type: reinforcement-learning name: reinforcement-learning dataset: name: Taxi-v3 type: Taxi-v3 metrics: - type: mean_reward value: 7.56 +/- 2.71 name: mean_reward verified: false --- # **Q-Learning** Agent playing1 **Taxi-v3** This is a trained model of a **Q-Learning** agent playing **Taxi-v3** . ## Usage ```python model = load_from_hub(repo_id="ibrahimciko/q-taxi-v3", filename="q-learning.pkl") # Don't forget to check if you need to add additional attributes (is_slippery=False etc) env = gym.make(model["env_id"]) ```

vnktrmnb/MBERT_FT-TyDiQA_S311

vnktrmnb

2023-08-30T09:43:41Z

61

0

transformers

[ "transformers", "tf", "tensorboard", "bert", "question-answering", "generated_from_keras_callback", "base_model:google-bert/bert-base-multilingual-cased", "base_model:finetune:google-bert/bert-base-multilingual-cased", "license:apache-2.0", "endpoints_compatible", "region:us" ]

question-answering

2023-08-30T08:49:08Z

--- license: apache-2.0 base_model: bert-base-multilingual-cased tags: - generated_from_keras_callback model-index: - name: vnktrmnb/MBERT_FT-TyDiQA_S311 results: [] ---  # vnktrmnb/MBERT_FT-TyDiQA_S311 This model is a fine-tuned version of [bert-base-multilingual-cased](https://huggingface.co/bert-base-multilingual-cased) on an unknown dataset. It achieves the following results on the evaluation set: - Train Loss: 0.6000 - Train End Logits Accuracy: 0.8451 - Train Start Logits Accuracy: 0.8711 - Validation Loss: 0.4689 - Validation End Logits Accuracy: 0.8686 - Validation Start Logits Accuracy: 0.9137 - Epoch: 2 ## Model description More information needed ## Intended uses & limitations More information needed ## Training and evaluation data More information needed ## Training procedure ### Training hyperparameters The following hyperparameters were used during training: - optimizer: {'name': 'Adam', 'weight_decay': None, 'clipnorm': None, 'global_clipnorm': None, 'clipvalue': None, 'use_ema': False, 'ema_momentum': 0.99, 'ema_overwrite_frequency': None, 'jit_compile': True, 'is_legacy_optimizer': False, 'learning_rate': {'class_name': 'PolynomialDecay', 'config': {'initial_learning_rate': 2e-05, 'decay_steps': 2412, 'end_learning_rate': 0.0, 'power': 1.0, 'cycle': False, 'name': None}}, 'beta_1': 0.9, 'beta_2': 0.999, 'epsilon': 1e-08, 'amsgrad': False} - training_precision: float32 ### Training results | Train Loss | Train End Logits Accuracy | Train Start Logits Accuracy | Validation Loss | Validation End Logits Accuracy | Validation Start Logits Accuracy | Epoch | |:----------:|:-------------------------:|:---------------------------:|:---------------:|:------------------------------:|:--------------------------------:|:-----:| | 1.4308 | 0.6683 | 0.6988 | 0.4811 | 0.8647 | 0.9072 | 0 | | 0.8301 | 0.7904 | 0.8263 | 0.4455 | 0.8698 | 0.9111 | 1 | | 0.6000 | 0.8451 | 0.8711 | 0.4689 | 0.8686 | 0.9137 | 2 | ### Framework versions - Transformers 4.32.1 - TensorFlow 2.12.0 - Datasets 2.14.4 - Tokenizers 0.13.3

ThuyNT03/xlm-roberta-base-Final_VietNam-aug_replace_synonym

ThuyNT03

2023-08-30T09:39:50Z

104

0

transformers

[ "transformers", "pytorch", "xlm-roberta", "text-classification", "generated_from_trainer", "base_model:FacebookAI/xlm-roberta-base", "base_model:finetune:FacebookAI/xlm-roberta-base", "license:mit", "autotrain_compatible", "endpoints_compatible", "region:us" ]

text-classification

2023-08-30T09:32:37Z

--- license: mit base_model: xlm-roberta-base tags: - generated_from_trainer metrics: - accuracy - f1 model-index: - name: xlm-roberta-base-Final_VietNam-aug_replace_synonym results: [] ---  # xlm-roberta-base-Final_VietNam-aug_replace_synonym This model is a fine-tuned version of [xlm-roberta-base](https://huggingface.co/xlm-roberta-base) on the None dataset. It achieves the following results on the evaluation set: - Loss: 1.1196 - Accuracy: 0.74 - F1: 0.7476 ## Model description More information needed ## Intended uses & limitations More information needed ## Training and evaluation data More information needed ## Training procedure ### Training hyperparameters The following hyperparameters were used during training: - learning_rate: 2e-05 - train_batch_size: 16 - eval_batch_size: 16 - seed: 42 - optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08 - lr_scheduler_type: linear - num_epochs: 8 ### Training results | Training Loss | Epoch | Step | Validation Loss | Accuracy | F1 | |:-------------:|:-----:|:----:|:---------------:|:--------:|:------:| | 1.0236 | 1.0 | 86 | 0.8388 | 0.66 | 0.6569 | | 0.698 | 2.0 | 172 | 0.7581 | 0.71 | 0.7064 | | 0.5136 | 3.0 | 258 | 0.7604 | 0.74 | 0.7348 | | 0.3853 | 4.0 | 344 | 0.8044 | 0.75 | 0.7563 | | 0.2299 | 5.0 | 430 | 0.7855 | 0.74 | 0.7484 | | 0.2054 | 6.0 | 516 | 0.9726 | 0.74 | 0.7459 | | 0.1438 | 7.0 | 602 | 1.1100 | 0.73 | 0.7386 | | 0.108 | 8.0 | 688 | 1.1196 | 0.74 | 0.7476 | ### Framework versions - Transformers 4.32.1 - Pytorch 2.0.0 - Datasets 2.14.4 - Tokenizers 0.13.3

AdanLee/poca-SoccerTwos

AdanLee

2023-08-30T09:27:16Z

1

0

ml-agents

[ "ml-agents", "tensorboard", "onnx", "SoccerTwos", "deep-reinforcement-learning", "reinforcement-learning", "ML-Agents-SoccerTwos", "region:us" ]

reinforcement-learning

2023-08-30T09:26:09Z

--- library_name: ml-agents tags: - SoccerTwos - deep-reinforcement-learning - reinforcement-learning - ML-Agents-SoccerTwos --- # **poca** Agent playing **SoccerTwos** This is a trained model of a **poca** agent playing **SoccerTwos** using the [Unity ML-Agents Library](https://github.com/Unity-Technologies/ml-agents). ## Usage (with ML-Agents) The Documentation: https://unity-technologies.github.io/ml-agents/ML-Agents-Toolkit-Documentation/ We wrote a complete tutorial to learn to train your first agent using ML-Agents and publish it to the Hub: - A *short tutorial* where you teach Huggy the Dog 🐶 to fetch the stick and then play with him directly in your browser: https://huggingface.co/learn/deep-rl-course/unitbonus1/introduction - A *longer tutorial* to understand how works ML-Agents: https://huggingface.co/learn/deep-rl-course/unit5/introduction ### Resume the training ```bash mlagents-learn <your_configuration_file_path.yaml> --run-id=<run_id> --resume ``` ### Watch your Agent play You can watch your agent **playing directly in your browser** 1. If the environment is part of ML-Agents official environments, go to https://huggingface.co/unity 2. Step 1: Find your model_id: AdanLee/poca-SoccerTwos 3. Step 2: Select your *.nn /*.onnx file 4. Click on Watch the agent play 👀

ThuyNT03/xlm-roberta-base-Final_VietNam-aug_insert_tfidf

ThuyNT03

2023-08-30T09:25:39Z

124

0

transformers

[ "transformers", "pytorch", "xlm-roberta", "text-classification", "generated_from_trainer", "base_model:FacebookAI/xlm-roberta-base", "base_model:finetune:FacebookAI/xlm-roberta-base", "license:mit", "autotrain_compatible", "endpoints_compatible", "region:us" ]

text-classification

2023-08-30T08:45:51Z

--- license: mit base_model: xlm-roberta-base tags: - generated_from_trainer metrics: - accuracy - f1 model-index: - name: xlm-roberta-base-Final_VietNam-aug_insert_tfidf results: [] ---  # xlm-roberta-base-Final_VietNam-aug_insert_tfidf This model is a fine-tuned version of [xlm-roberta-base](https://huggingface.co/xlm-roberta-base) on the None dataset. It achieves the following results on the evaluation set: - Loss: 1.2753 - Accuracy: 0.71 - F1: 0.7161 ## Model description More information needed ## Intended uses & limitations More information needed ## Training and evaluation data More information needed ## Training procedure ### Training hyperparameters The following hyperparameters were used during training: - learning_rate: 2e-05 - train_batch_size: 16 - eval_batch_size: 16 - seed: 42 - optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08 - lr_scheduler_type: linear - num_epochs: 8 ### Training results | Training Loss | Epoch | Step | Validation Loss | Accuracy | F1 | |:-------------:|:-----:|:----:|:---------------:|:--------:|:------:| | 1.0036 | 1.0 | 87 | 0.7585 | 0.63 | 0.6038 | | 0.6784 | 2.0 | 174 | 0.6901 | 0.7 | 0.6939 | | 0.5132 | 3.0 | 261 | 0.6510 | 0.76 | 0.7658 | | 0.3868 | 4.0 | 348 | 0.7266 | 0.74 | 0.7436 | | 0.2694 | 5.0 | 435 | 0.8702 | 0.72 | 0.7264 | | 0.1805 | 6.0 | 522 | 1.1744 | 0.72 | 0.7207 | | 0.1813 | 7.0 | 609 | 1.2328 | 0.72 | 0.7256 | | 0.1258 | 8.0 | 696 | 1.2753 | 0.71 | 0.7161 | ### Framework versions - Transformers 4.32.1 - Pytorch 2.0.0 - Datasets 2.14.4 - Tokenizers 0.13.3

aviroes/elderly_whisper-medium-LoRA

aviroes

2023-08-30T09:11:57Z

0

null

[ "generated_from_trainer", "base_model:openai/whisper-medium", "base_model:finetune:openai/whisper-medium", "license:apache-2.0", "region:us" ]

null

2023-08-30T07:18:13Z

--- license: apache-2.0 base_model: openai/whisper-medium tags: - generated_from_trainer model-index: - name: elderly_whisper-medium-LoRA results: [] ---  # elderly_whisper-medium-LoRA This model is a fine-tuned version of [openai/whisper-medium](https://huggingface.co/openai/whisper-medium) on the None dataset. ## Model description More information needed ## Intended uses & limitations More information needed ## Training and evaluation data More information needed ## Training procedure ### Training hyperparameters The following hyperparameters were used during training: - learning_rate: 0.001 - train_batch_size: 8 - eval_batch_size: 8 - seed: 42 - optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08 - lr_scheduler_type: linear - lr_scheduler_warmup_steps: 200 - num_epochs: 3 ### Framework versions - Transformers 4.33.0.dev0 - Pytorch 2.0.1+cu118 - Datasets 2.14.4 - Tokenizers 0.13.3

ldhldh/12.8b_lora

ldhldh

2023-08-30T09:10:10Z

0

peft

[ "peft", "region:us" ]

null

2023-08-28T10:05:47Z

--- library_name: peft --- ## Training procedure The following `bitsandbytes` quantization config was used during training: - quant_method: bitsandbytes - load_in_8bit: False - load_in_4bit: True - llm_int8_threshold: 6.0 - llm_int8_skip_modules: None - llm_int8_enable_fp32_cpu_offload: False - llm_int8_has_fp16_weight: False - bnb_4bit_quant_type: nf4 - bnb_4bit_use_double_quant: False - bnb_4bit_compute_dtype: bfloat16 ### Framework versions - PEFT 0.6.0.dev0

AbdelKarim95/ppo-Huggy

AbdelKarim95

2023-08-30T08:39:48Z

0

ml-agents

[ "ml-agents", "tensorboard", "onnx", "Huggy", "deep-reinforcement-learning", "reinforcement-learning", "ML-Agents-Huggy", "region:us" ]

reinforcement-learning

2023-08-30T07:53:50Z

--- library_name: ml-agents tags: - Huggy - deep-reinforcement-learning - reinforcement-learning - ML-Agents-Huggy --- # **ppo** Agent playing **Huggy** This is a trained model of a **ppo** agent playing **Huggy** using the [Unity ML-Agents Library](https://github.com/Unity-Technologies/ml-agents). ## Usage (with ML-Agents) The Documentation: https://unity-technologies.github.io/ml-agents/ML-Agents-Toolkit-Documentation/ We wrote a complete tutorial to learn to train your first agent using ML-Agents and publish it to the Hub: - A *short tutorial* where you teach Huggy the Dog 🐶 to fetch the stick and then play with him directly in your browser: https://huggingface.co/learn/deep-rl-course/unitbonus1/introduction - A *longer tutorial* to understand how works ML-Agents: https://huggingface.co/learn/deep-rl-course/unit5/introduction ### Resume the training ```bash mlagents-learn <your_configuration_file_path.yaml> --run-id=<run_id> --resume ``` ### Watch your Agent play You can watch your agent **playing directly in your browser** 1. If the environment is part of ML-Agents official environments, go to https://huggingface.co/unity 2. Step 1: Find your model_id: AbdelKarim95/ppo-Huggy 3. Step 2: Select your *.nn /*.onnx file 4. Click on Watch the agent play 👀

jasanfa/distilbert-base-uncased-finetuned-emotion

jasanfa

2023-08-30T08:37:10Z

106

0

transformers

[ "transformers", "pytorch", "distilbert", "text-classification", "generated_from_trainer", "base_model:distilbert/distilbert-base-uncased", "base_model:finetune:distilbert/distilbert-base-uncased", "license:apache-2.0", "autotrain_compatible", "endpoints_compatible", "region:us" ]

text-classification

2023-08-29T14:43:00Z

--- license: apache-2.0 base_model: distilbert-base-uncased tags: - generated_from_trainer metrics: - accuracy - f1 model-index: - name: distilbert-base-uncased-finetuned-emotion results: [] ---  # distilbert-base-uncased-finetuned-emotion This model is a fine-tuned version of [distilbert-base-uncased](https://huggingface.co/distilbert-base-uncased) on an unknown dataset. It achieves the following results on the evaluation set: - Loss: 0.2165 - Accuracy: 0.9195 - F1: 0.9197 ## Model description More information needed ## Intended uses & limitations More information needed ## Training and evaluation data More information needed ## Training procedure ### Training hyperparameters The following hyperparameters were used during training: - learning_rate: 2e-05 - train_batch_size: 64 - eval_batch_size: 64 - seed: 42 - optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08 - lr_scheduler_type: linear - num_epochs: 2 ### Training results | Training Loss | Epoch | Step | Validation Loss | Accuracy | F1 | |:-------------:|:-----:|:----:|:---------------:|:--------:|:------:| | 0.8388 | 1.0 | 250 | 0.3212 | 0.906 | 0.9049 | | 0.251 | 2.0 | 500 | 0.2165 | 0.9195 | 0.9197 | ### Framework versions - Transformers 4.32.1 - Pytorch 2.0.1+cu118 - Tokenizers 0.13.3

yuanxing/Inuyasha

yuanxing

2023-08-30T08:36:02Z

0

null

[ "region:us" ]

null

2023-08-30T08:26:57Z

Inuyasha_犬夜叉 ![image/png](https://cdn-uploads.huggingface.co/production/uploads/64326351b0dee98d1779e08b/4kM2o96wf3D337MtdAVG0.png) ![image/png](https://cdn-uploads.huggingface.co/production/uploads/64326351b0dee98d1779e08b/C-u9z8AMIZKwaEMRIJQWp.png)

rasha-salim/autotrain-dreambooth-512

rasha-salim

2023-08-30T08:34:29Z

2

1

diffusers

[ "diffusers", "text-to-image", "autotrain", "base_model:stabilityai/stable-diffusion-xl-base-1.0", "base_model:finetune:stabilityai/stable-diffusion-xl-base-1.0", "region:us" ]

text-to-image

2023-08-30T05:34:35Z

--- base_model: stabilityai/stable-diffusion-xl-base-1.0 instance_prompt: photo of a sks adam tags: - text-to-image - diffusers - autotrain inference: true --- # DreamBooth trained by AutoTrain Text encoder was not trained.

sunkai4u/sd-class-butterflies-32

sunkai4u

2023-08-30T08:26:49Z

46

0

diffusers

[ "diffusers", "safetensors", "pytorch", "unconditional-image-generation", "diffusion-models-class", "license:mit", "diffusers:DDPMPipeline", "region:us" ]

unconditional-image-generation

2023-08-30T08:25:57Z

--- license: mit tags: - pytorch - diffusers - unconditional-image-generation - diffusion-models-class --- # Model Card for Unit 1 of the [Diffusion Models Class 🧨](https://github.com/huggingface/diffusion-models-class) This model is a diffusion model for unconditional image generation of cute 🦋. ## Usage ```python from diffusers import DDPMPipeline pipeline = DDPMPipeline.from_pretrained('sunkai4u/sd-class-butterflies-32') image = pipeline().images[0] image ```

RyyyT/ppo-Huggy

RyyyT

2023-08-30T08:22:39Z

0

ml-agents

[ "ml-agents", "tensorboard", "onnx", "Huggy", "deep-reinforcement-learning", "reinforcement-learning", "ML-Agents-Huggy", "region:us" ]

reinforcement-learning

2023-08-30T08:22:27Z

--- library_name: ml-agents tags: - Huggy - deep-reinforcement-learning - reinforcement-learning - ML-Agents-Huggy --- # **ppo** Agent playing **Huggy** This is a trained model of a **ppo** agent playing **Huggy** using the [Unity ML-Agents Library](https://github.com/Unity-Technologies/ml-agents). ## Usage (with ML-Agents) The Documentation: https://unity-technologies.github.io/ml-agents/ML-Agents-Toolkit-Documentation/ We wrote a complete tutorial to learn to train your first agent using ML-Agents and publish it to the Hub: - A *short tutorial* where you teach Huggy the Dog 🐶 to fetch the stick and then play with him directly in your browser: https://huggingface.co/learn/deep-rl-course/unitbonus1/introduction - A *longer tutorial* to understand how works ML-Agents: https://huggingface.co/learn/deep-rl-course/unit5/introduction ### Resume the training ```bash mlagents-learn <your_configuration_file_path.yaml> --run-id=<run_id> --resume ``` ### Watch your Agent play You can watch your agent **playing directly in your browser** 1. If the environment is part of ML-Agents official environments, go to https://huggingface.co/unity 2. Step 1: Find your model_id: RyyyT/ppo-Huggy 3. Step 2: Select your *.nn /*.onnx file 4. Click on Watch the agent play 👀

r0llingclouds/distilbert-base-uncased-finetuned-clinc

r0llingclouds

2023-08-30T08:21:12Z

105

0

transformers

[ "transformers", "pytorch", "tensorboard", "distilbert", "text-classification", "generated_from_trainer", "dataset:clinc_oos", "license:apache-2.0", "model-index", "autotrain_compatible", "endpoints_compatible", "region:us" ]

text-classification

2023-08-30T08:17:41Z

--- license: apache-2.0 tags: - generated_from_trainer datasets: - clinc_oos metrics: - accuracy model-index: - name: distilbert-base-uncased-finetuned-clinc results: - task: name: Text Classification type: text-classification dataset: name: clinc_oos type: clinc_oos args: plus metrics: - name: Accuracy type: accuracy value: 0.9161290322580645 ---  # distilbert-base-uncased-finetuned-clinc This model is a fine-tuned version of [distilbert-base-uncased](https://huggingface.co/distilbert-base-uncased) on the clinc_oos dataset. It achieves the following results on the evaluation set: - Loss: 0.7754 - Accuracy: 0.9161 ## Model description More information needed ## Intended uses & limitations More information needed ## Training and evaluation data More information needed ## Training procedure ### Training hyperparameters The following hyperparameters were used during training: - learning_rate: 2e-05 - train_batch_size: 48 - eval_batch_size: 48 - seed: 42 - optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08 - lr_scheduler_type: linear - num_epochs: 5 ### Training results | Training Loss | Epoch | Step | Validation Loss | Accuracy | |:-------------:|:-----:|:----:|:---------------:|:--------:| | 4.2893 | 1.0 | 318 | 3.2831 | 0.7397 | | 2.6289 | 2.0 | 636 | 1.8731 | 0.8345 | | 1.5481 | 3.0 | 954 | 1.1580 | 0.89 | | 1.0137 | 4.0 | 1272 | 0.8584 | 0.9077 | | 0.7969 | 5.0 | 1590 | 0.7754 | 0.9161 | ### Framework versions - Transformers 4.16.2 - Pytorch 2.0.1+cu118 - Datasets 1.16.1 - Tokenizers 0.13.3

nishant-glance/model-sd-1-4-priorp-unet-2000-lr2e-ab

nishant-glance

2023-08-30T08:09:13Z

0

diffusers

[ "diffusers", "tensorboard", "safetensors", "stable-diffusion", "stable-diffusion-diffusers", "text-to-image", "dreambooth", "base_model:CompVis/stable-diffusion-v1-4", "base_model:finetune:CompVis/stable-diffusion-v1-4", "license:creativeml-openrail-m", "autotrain_compatible", "endpoints_compatible", "diffusers:StableDiffusionPipeline", "region:us" ]

text-to-image

2023-08-29T07:24:54Z

--- license: creativeml-openrail-m base_model: CompVis/stable-diffusion-v1-4 instance_prompt: a photo of yqg person tags: - stable-diffusion - stable-diffusion-diffusers - text-to-image - diffusers - dreambooth inference: true --- # DreamBooth - nishant-glance/model-sd-1-4-priorp-unet-2000-lr2e-ab This is a dreambooth model derived from CompVis/stable-diffusion-v1-4. The weights were trained on a photo of yqg person using [DreamBooth](https://dreambooth.github.io/). You can find some example images in the following. DreamBooth for the text encoder was enabled: True.

IA-hf/sd-pokemon-model-lora-sdxl

IA-hf

2023-08-30T08:04:08Z

1

2

diffusers

[ "diffusers", "stable-diffusion-xl", "stable-diffusion-xl-diffusers", "text-to-image", "lora", "base_model:stabilityai/stable-diffusion-xl-base-1.0", "base_model:adapter:stabilityai/stable-diffusion-xl-base-1.0", "license:creativeml-openrail-m", "region:us" ]

text-to-image

2023-08-30T05:17:25Z

--- license: creativeml-openrail-m base_model: stabilityai/stable-diffusion-xl-base-1.0 dataset: lambdalabs/pokemon-blip-captions tags: - stable-diffusion-xl - stable-diffusion-xl-diffusers - text-to-image - diffusers - lora inference: true --- # LoRA text2image fine-tuning - IA-hf/sd-pokemon-model-lora-sdxl These are LoRA adaption weights for stabilityai/stable-diffusion-xl-base-1.0. The weights were fine-tuned on the lambdalabs/pokemon-blip-captions dataset. You can find some example images in the following. ![img_0](./image_0.png) ![img_1](./image_1.png) ![img_2](./image_2.png) ![img_3](./image_3.png) LoRA for the text encoder was enabled: False. Special VAE used for training: madebyollin/sdxl-vae-fp16-fix.

raruidol/SchemeClassifier-ENG

raruidol

2023-08-30T08:02:50Z

114

0

transformers

[ "transformers", "pytorch", "safetensors", "roberta", "text-classification", "license:cc-by-nc-sa-4.0", "autotrain_compatible", "endpoints_compatible", "region:us" ]

text-classification

2023-08-24T18:37:43Z

--- license: cc-by-nc-sa-4.0 ---

TFMC/ELYZA-japanese-Llama-2-7b-instruct-GPTQ-4bit-64g

TFMC

2023-08-30T07:53:28Z

96

4

transformers

[ "transformers", "safetensors", "llama", "text-generation", "license:llama2", "autotrain_compatible", "text-generation-inference", "endpoints_compatible", "region:us" ]

text-generation

2023-08-30T01:30:19Z

--- license: llama2 --- # ELYZA-japanese-Llama-2-7b-instruct-GPTQ-4bit-64g GPTQ model for ["elyza/ELYZA-japanese-Llama-2-7b-instruct"](https://huggingface.co/elyza/ELYZA-japanese-Llama-2-7b-instruct) : 4bits, gr64, desc_act=True

Idriska/my_awesome_model

Idriska

2023-08-30T07:46:17Z

105

0

transformers

[ "transformers", "pytorch", "distilbert", "text-classification", "generated_from_trainer", "dataset:imdb", "base_model:distilbert/distilbert-base-uncased", "base_model:finetune:distilbert/distilbert-base-uncased", "license:apache-2.0", "model-index", "autotrain_compatible", "endpoints_compatible", "region:us" ]

text-classification

2023-08-29T14:48:43Z

--- license: apache-2.0 base_model: distilbert-base-uncased tags: - generated_from_trainer datasets: - imdb metrics: - accuracy model-index: - name: my_awesome_model results: - task: name: Text Classification type: text-classification dataset: name: imdb type: imdb config: plain_text split: test args: plain_text metrics: - name: Accuracy type: accuracy value: 0.93348 ---  # my_awesome_model This model is a fine-tuned version of [distilbert-base-uncased](https://huggingface.co/distilbert-base-uncased) on the imdb dataset. It achieves the following results on the evaluation set: - Loss: 0.2250 - Accuracy: 0.9335 ## Model description More information needed ## Intended uses & limitations More information needed ## Training and evaluation data More information needed ## Training procedure ### Training hyperparameters The following hyperparameters were used during training: - learning_rate: 2e-05 - train_batch_size: 16 - eval_batch_size: 16 - seed: 42 - optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08 - lr_scheduler_type: linear - num_epochs: 2 ### Training results | Training Loss | Epoch | Step | Validation Loss | Accuracy | |:-------------:|:-----:|:----:|:---------------:|:--------:| | 0.2249 | 1.0 | 1563 | 0.2371 | 0.9148 | | 0.1521 | 2.0 | 3126 | 0.2250 | 0.9335 | ### Framework versions - Transformers 4.32.1 - Pytorch 2.0.1+cu118 - Datasets 2.14.4 - Tokenizers 0.13.3

nikinetrahutama/afx-ai-llama-chat-model-21

nikinetrahutama

2023-08-30T07:42:37Z

0

peft

[ "peft", "region:us" ]

null

2023-08-30T07:42:29Z

--- library_name: peft --- ## Training procedure The following `bitsandbytes` quantization config was used during training: - quant_method: bitsandbytes - load_in_8bit: False - load_in_4bit: True - llm_int8_threshold: 6.0 - llm_int8_skip_modules: None - llm_int8_enable_fp32_cpu_offload: False - llm_int8_has_fp16_weight: False - bnb_4bit_quant_type: nf4 - bnb_4bit_use_double_quant: True - bnb_4bit_compute_dtype: bfloat16 ### Framework versions - PEFT 0.6.0.dev0

Vineetttt/layoutlmv3-base-finetuned-FUNSD

Vineetttt

2023-08-30T07:29:45Z

126

0

transformers

[ "transformers", "pytorch", "layoutlmv3", "token-classification", "generated_from_trainer", "dataset:funsd-layoutlmv3", "base_model:microsoft/layoutlmv3-base", "base_model:finetune:microsoft/layoutlmv3-base", "license:cc-by-nc-sa-4.0", "model-index", "autotrain_compatible", "endpoints_compatible", "region:us" ]

token-classification

2023-08-30T07:15:58Z

--- license: cc-by-nc-sa-4.0 base_model: microsoft/layoutlmv3-base tags: - generated_from_trainer datasets: - funsd-layoutlmv3 metrics: - precision - recall - f1 - accuracy model-index: - name: test results: - task: name: Token Classification type: token-classification dataset: name: funsd-layoutlmv3 type: funsd-layoutlmv3 config: funsd split: test args: funsd metrics: - name: Precision type: precision value: 0.9002457002457003 - name: Recall type: recall value: 0.9100844510680576 - name: F1 type: f1 value: 0.9051383399209486 - name: Accuracy type: accuracy value: 0.8547486033519553 ---  # test This model is a fine-tuned version of [microsoft/layoutlmv3-base](https://huggingface.co/microsoft/layoutlmv3-base) on the funsd-layoutlmv3 dataset. It achieves the following results on the evaluation set: - Loss: 0.6194 - Precision: 0.9002 - Recall: 0.9101 - F1: 0.9051 - Accuracy: 0.8547 ## Model description More information needed ## Intended uses & limitations More information needed ## Training and evaluation data More information needed ## Training procedure ### Training hyperparameters The following hyperparameters were used during training: - learning_rate: 1e-05 - train_batch_size: 2 - eval_batch_size: 2 - seed: 42 - optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08 - lr_scheduler_type: linear - training_steps: 1000 ### Training results | Training Loss | Epoch | Step | Validation Loss | Precision | Recall | F1 | Accuracy | |:-------------:|:-----:|:----:|:---------------:|:---------:|:------:|:------:|:--------:| | No log | 1.33 | 100 | 0.6953 | 0.7761 | 0.8058 | 0.7906 | 0.7680 | | No log | 2.67 | 200 | 0.5117 | 0.8250 | 0.8808 | 0.8520 | 0.8290 | | No log | 4.0 | 300 | 0.5177 | 0.8397 | 0.8897 | 0.8640 | 0.8337 | | No log | 5.33 | 400 | 0.5165 | 0.8642 | 0.9106 | 0.8868 | 0.8509 | | 0.5653 | 6.67 | 500 | 0.5378 | 0.8735 | 0.9091 | 0.8909 | 0.8458 | | 0.5653 | 8.0 | 600 | 0.5698 | 0.8733 | 0.9111 | 0.8918 | 0.8482 | | 0.5653 | 9.33 | 700 | 0.5773 | 0.8934 | 0.9076 | 0.9004 | 0.8557 | | 0.5653 | 10.67 | 800 | 0.6073 | 0.8905 | 0.9006 | 0.8955 | 0.8520 | | 0.5653 | 12.0 | 900 | 0.6090 | 0.8940 | 0.9091 | 0.9015 | 0.8513 | | 0.1357 | 13.33 | 1000 | 0.6194 | 0.9002 | 0.9101 | 0.9051 | 0.8547 | ### Framework versions - Transformers 4.33.0.dev0 - Pytorch 2.0.1+cu118 - Datasets 2.14.4 - Tokenizers 0.13.3

morpheuslord/secllama

morpheuslord

2023-08-30T07:17:51Z

1

2

null

[ "tensorboard", "generated_from_trainer", "region:us" ]

null

2023-07-30T12:22:06Z

--- tags: - generated_from_trainer model-index: - name: secllama results: [] ---  # secllama This model is a fine-tuned version of [meta-llama/Llama-2-7b-hf](https://huggingface.co/meta-llama/Llama-2-7b-hf) on an unknown dataset. ## Model description More information needed ## Intended uses & limitations More information needed ## Training and evaluation data More information needed ## Training procedure ### Training hyperparameters The following hyperparameters were used during training: - learning_rate: 0.0002 - train_batch_size: 1 - eval_batch_size: 8 - seed: 42 - gradient_accumulation_steps: 16 - total_train_batch_size: 16 - optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08 - lr_scheduler_type: linear - training_steps: 1 ### Training results ### Framework versions - Transformers 4.30.2 - Pytorch 2.0.1+cu118 - Datasets 2.14.4 - Tokenizers 0.13.3

prashanth07/falcon7b_updated_samples_D

prashanth07

2023-08-30T07:14:51Z

0

peft

[ "peft", "region:us" ]

null

2023-08-30T07:14:33Z

--- library_name: peft --- ## Training procedure The following `bitsandbytes` quantization config was used during training: - quant_method: bitsandbytes - load_in_8bit: False - load_in_4bit: True - llm_int8_threshold: 6.0 - llm_int8_skip_modules: None - llm_int8_enable_fp32_cpu_offload: False - llm_int8_has_fp16_weight: False - bnb_4bit_quant_type: nf4 - bnb_4bit_use_double_quant: False - bnb_4bit_compute_dtype: float16 ### Framework versions - PEFT 0.6.0.dev0

Raychani1/slovakbert-ner-v2

Raychani1

2023-08-30T06:59:38Z

111

0

transformers

[ "transformers", "pytorch", "roberta", "token-classification", "generated_from_trainer", "sk", "dataset:NBS_sentence", "license:mit", "model-index", "autotrain_compatible", "region:us" ]

token-classification

2023-08-30T05:42:08Z

--- license: mit language: - sk tags: - generated_from_trainer datasets: - NBS_sentence metrics: - precision - recall - f1 - accuracy inference: false model-index: - name: slovakbert-ner-v2 results: - task: name: Token Classification type: token-classification metrics: - name: Precision type: precision value: 0.9715 - name: Recall type: recall value: 0.9433 - name: F1 type: f1 value: 0.9547 - name: Accuracy type: accuracy value: 0.9897 --- # SlovakBERT based Named Entity Recognition Deep Learning model developed for Named Entity Recognition (NER) in Slovak. The [**Gerulata/SlovakBERT**](https://huggingface.co/gerulata/slovakbert) based model is fine-tuned on webscraped Slovak news articles. The finished model supports the following IOB tagged entity categories: **PERSON**, **ORGANIZATION**, **LOCATION**, **DATE**, **TIME**, **MONEY** and **PERCENTAGE** ### **Related Work** [![Thesis][Thesis]][Thesis-url] ## Model usage ### Simple Named Entity Recognition (NER) ```python from transformers import pipeline ner_pipeline = pipeline(task='ner', model='Raychani1/slovakbert-ner-v2') input_sentence = 'Hoci podľa ostatných údajov NBS pre Bratislavský kraj je aktuálna priemerná cena nehnuteľností na úrovni 2 072 eur za štvorcový meter, ceny bytov v hlavnom meste sú podstatne vyššie.' classifications = ner_pipeline(input_sentence) ``` ### Named Entity Recognition (NER) with Visualization For a Visualization Example please refer to the following [Gist](https://gist.github.com/Raychani1/7d4455491f0aa681ed8ea99d8b1d8279). ### Model Prediction Output Example ![prediction_output](https://github.com/Raychani1/Text_Parsing_Methods_Using_NLP/assets/45550552/723ab7f1-4efb-4d03-87d6-b9ac1e40990f) ## Model Training ### Training Hyperparameters | **Parameter** | **Value** | |:---------------------------:|:---------:| | per_device_train_batch_size | 4 | | per_device_eval_batch_size | 4 | | learning_rate | 5e-05 | | adam_beta1 | 0.9 | | adam_beta1 | 0.999 | | adam_epsilon | 1e-08 | | num_train_epochs | 15 | | lr_scheduler_type | linear | | seed | 42 | ### Training results Best model results are reached in the 8th training epoch. | Training Loss | Epoch | Step | Validation Loss | Precision | Recall | F1 | Accuracy | |:-------------:|:-----:|:----:|:---------------:|:---------:|:------:|:------:|:--------:| | 0.6721 | 1.0 | 70 | 0.2214 | 0.6972 | 0.7308 | 0.7136 | 0.9324 | | 0.1849 | 2.0 | 140 | 0.1697 | 0.8056 | 0.8365 | 0.8208 | 0.952 | | 0.0968 | 3.0 | 210 | 0.1213 | 0.882 | 0.8622 | 0.872 | 0.9728 | | 0.0468 | 4.0 | 280 | 0.1107 | 0.8372 | 0.907 | 0.8708 | 0.9684 | | 0.0415 | 5.0 | 350 | 0.1644 | 0.8059 | 0.8782 | 0.8405 | 0.9615 | | 0.0233 | 6.0 | 420 | 0.1255 | 0.8576 | 0.8878 | 0.8724 | 0.9716 | | 0.0198 | 7.0 | 490 | 0.1383 | 0.8545 | 0.8846 | 0.8693 | 0.9703 | | 0.0133 | 8.0 | 560 | 0.1241 | 0.884 | 0.9038 | 0.8938 | 0.9735 | ## Model Evaluation ### Evaluation Dataset Distribution | **NER Tag** | **Number of Tokens** | |:-----------------:|:--------------------:| | **0** | 6568 | | **B-Person** | 96 | | **I-Person** | 83 | | **B-Organizaton** | 583 | | **I-Organizaton** | 585 | | **B-Location** | 59 | | **I-Location** | 15 | | **B-Date** | 113 | | **I-Date** | 87 | | **Time** | 5 | | **B-Money** | 44 | | **I-Money** | 74 | | **B-Percentage** | 57 | | **I-Percentage** | 54 | ### Evaluation Confusion Matrix ![image](https://github.com/Raychani1/Text_Parsing_Methods_Using_NLP/assets/45550552/e6d1a1c6-e02f-4de9-9684-5882a405d31f) ### Evaluation Model Metrics | **Precision** | **Macro-Precision** | **Recall** | **Macro-Recall** | **F1** | **Macro-F1** | **Accuracy** | |:-------------:|:-------------------:|:----------:|:----------------:|:------:|:------------:|:------------:| | 0.9897 | 0.9715 | 0.9897 | 0.9433 | 0.9895 | 0.9547 | 0.9897 | ## Framework Versions - Transformers 4.26.1 - PyTorch 1.13.1 - Tokenizers 0.13.2  [Thesis]: https://img.shields.io/badge/%F0%9F%93%9C-Masters%20Thesis-blue?style=for-the-badge [Thesis-url]: https://opac.crzp.sk/?fn=detailBiblioForm&sid=C0DEB8E07572332BA2230915805F

susmitabhatt/my_awesome_mind_model

susmitabhatt

2023-08-30T06:49:32Z

14

0

transformers

[ "transformers", "pytorch", "tensorboard", "wav2vec2", "audio-classification", "generated_from_trainer", "dataset:marsyas/gtzan", "license:apache-2.0", "endpoints_compatible", "region:us" ]

audio-classification

2023-08-28T04:50:07Z

--- license: apache-2.0 tags: - generated_from_trainer datasets: - marsyas/gtzan metrics: - accuracy model-index: - name: my_awesome_mind_model results: [] ---  # my_awesome_mind_model This model is a fine-tuned version of [facebook/wav2vec2-base](https://huggingface.co/facebook/wav2vec2-base) on the GTZAN dataset. It achieves the following results on the evaluation set: - Loss: 1.7894 - Accuracy: 0.46 ## Model description More information needed ## Intended uses & limitations More information needed ## Training and evaluation data More information needed ## Training procedure ### Training hyperparameters The following hyperparameters were used during training: - learning_rate: 7e-05 - train_batch_size: 32 - eval_batch_size: 32 - seed: 42 - gradient_accumulation_steps: 4 - total_train_batch_size: 128 - optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08 - lr_scheduler_type: linear - lr_scheduler_warmup_ratio: 0.1 - num_epochs: 10 ### Training results | Training Loss | Epoch | Step | Validation Loss | Accuracy | |:-------------:|:-----:|:----:|:---------------:|:--------:| | No log | 0.96 | 6 | 2.2804 | 0.22 | | 2.2742 | 1.92 | 12 | 2.1887 | 0.25 | | 2.2742 | 2.88 | 18 | 2.1220 | 0.33 | | 2.1487 | 4.0 | 25 | 2.0666 | 0.355 | | 2.0354 | 4.96 | 31 | 1.9963 | 0.325 | | 2.0354 | 5.92 | 37 | 1.9090 | 0.375 | | 1.9246 | 6.88 | 43 | 1.8679 | 0.405 | | 1.8416 | 8.0 | 50 | 1.8185 | 0.44 | | 1.8416 | 8.96 | 56 | 1.7970 | 0.43 | | 1.7858 | 9.6 | 60 | 1.7894 | 0.46 | ### Framework versions - Transformers 4.30.2 - Pytorch 2.0.0 - Datasets 2.1.0 - Tokenizers 0.13.3

dkqjrm/20230830102643

dkqjrm

2023-08-30T06:29:45Z

105

0

transformers

[ "transformers", "pytorch", "tensorboard", "bert", "text-classification", "generated_from_trainer", "dataset:super_glue", "license:apache-2.0", "autotrain_compatible", "endpoints_compatible", "region:us" ]

text-classification

2023-08-30T01:27:07Z

--- license: apache-2.0 tags: - generated_from_trainer datasets: - super_glue metrics: - accuracy model-index: - name: '20230830102643' results: [] ---  # 20230830102643 This model is a fine-tuned version of [bert-large-cased](https://huggingface.co/bert-large-cased) on the super_glue dataset. It achieves the following results on the evaluation set: - Loss: 0.6732 - Accuracy: 0.5 ## Model description More information needed ## Intended uses & limitations More information needed ## Training and evaluation data More information needed ## Training procedure ### Training hyperparameters The following hyperparameters were used during training: - learning_rate: 0.0007 - train_batch_size: 16 - eval_batch_size: 8 - seed: 11 - optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08 - lr_scheduler_type: linear - num_epochs: 80.0 ### Training results | Training Loss | Epoch | Step | Validation Loss | Accuracy | |:-------------:|:-----:|:-----:|:---------------:|:--------:| | No log | 1.0 | 340 | 0.6910 | 0.5 | | 0.7049 | 2.0 | 680 | 0.6698 | 0.5078 | | 0.6976 | 3.0 | 1020 | 0.6913 | 0.5 | | 0.6976 | 4.0 | 1360 | 0.7490 | 0.5 | | 0.6898 | 5.0 | 1700 | 0.8592 | 0.5078 | | 0.6423 | 6.0 | 2040 | 0.7080 | 0.5987 | | 0.6423 | 7.0 | 2380 | 0.6940 | 0.5 | | 0.6463 | 8.0 | 2720 | 0.6703 | 0.5 | | 0.7054 | 9.0 | 3060 | 0.6741 | 0.5047 | | 0.7054 | 10.0 | 3400 | 0.6784 | 0.5 | | 0.7011 | 11.0 | 3740 | 0.6707 | 0.5 | | 0.6895 | 12.0 | 4080 | 0.6941 | 0.5 | | 0.6895 | 13.0 | 4420 | 0.7550 | 0.5 | | 0.6898 | 14.0 | 4760 | 0.7095 | 0.5 | | 0.6916 | 15.0 | 5100 | 0.6711 | 0.5 | | 0.6916 | 16.0 | 5440 | 0.7051 | 0.5 | | 0.6891 | 17.0 | 5780 | 0.6831 | 0.5 | | 0.6875 | 18.0 | 6120 | 0.6733 | 0.5172 | | 0.6875 | 19.0 | 6460 | 0.6707 | 0.5 | | 0.6854 | 20.0 | 6800 | 0.6851 | 0.5 | | 0.686 | 21.0 | 7140 | 0.6704 | 0.5 | | 0.686 | 22.0 | 7480 | 0.6938 | 0.5 | | 0.6874 | 23.0 | 7820 | 0.6848 | 0.5 | | 0.6848 | 24.0 | 8160 | 0.6710 | 0.5 | | 0.6811 | 25.0 | 8500 | 0.6783 | 0.4953 | | 0.6811 | 26.0 | 8840 | 0.6886 | 0.5 | | 0.6837 | 27.0 | 9180 | 0.7146 | 0.5 | | 0.6851 | 28.0 | 9520 | 0.6703 | 0.5 | | 0.6851 | 29.0 | 9860 | 0.6884 | 0.5 | | 0.6813 | 30.0 | 10200 | 0.6704 | 0.5 | | 0.6826 | 31.0 | 10540 | 0.6704 | 0.5 | | 0.6826 | 32.0 | 10880 | 0.6970 | 0.5 | | 0.6813 | 33.0 | 11220 | 0.6707 | 0.5 | | 0.6786 | 34.0 | 11560 | 0.6840 | 0.5 | | 0.6786 | 35.0 | 11900 | 0.6722 | 0.5 | | 0.6821 | 36.0 | 12240 | 0.6706 | 0.5 | | 0.6799 | 37.0 | 12580 | 0.6707 | 0.5 | | 0.6799 | 38.0 | 12920 | 0.6824 | 0.4953 | | 0.6803 | 39.0 | 13260 | 0.6995 | 0.5 | | 0.6775 | 40.0 | 13600 | 0.6728 | 0.5 | | 0.6775 | 41.0 | 13940 | 0.6711 | 0.4984 | | 0.679 | 42.0 | 14280 | 0.6743 | 0.5 | | 0.6775 | 43.0 | 14620 | 0.6742 | 0.5016 | | 0.6775 | 44.0 | 14960 | 0.6724 | 0.5016 | | 0.6766 | 45.0 | 15300 | 0.6884 | 0.5 | | 0.6755 | 46.0 | 15640 | 0.6709 | 0.5 | | 0.6755 | 47.0 | 15980 | 0.6811 | 0.5 | | 0.6785 | 48.0 | 16320 | 0.6709 | 0.5 | | 0.6765 | 49.0 | 16660 | 0.6813 | 0.5 | | 0.6761 | 50.0 | 17000 | 0.6724 | 0.5 | | 0.6761 | 51.0 | 17340 | 0.6713 | 0.5 | | 0.6764 | 52.0 | 17680 | 0.6715 | 0.5016 | | 0.6774 | 53.0 | 18020 | 0.6730 | 0.5 | | 0.6774 | 54.0 | 18360 | 0.6730 | 0.5 | | 0.673 | 55.0 | 18700 | 0.6716 | 0.5016 | | 0.6718 | 56.0 | 19040 | 0.6714 | 0.5 | | 0.6718 | 57.0 | 19380 | 0.6714 | 0.4984 | | 0.6745 | 58.0 | 19720 | 0.6715 | 0.5016 | | 0.6735 | 59.0 | 20060 | 0.6863 | 0.5 | | 0.6735 | 60.0 | 20400 | 0.6710 | 0.4984 | | 0.6725 | 61.0 | 20740 | 0.6718 | 0.5063 | | 0.6734 | 62.0 | 21080 | 0.6717 | 0.5 | | 0.6734 | 63.0 | 21420 | 0.6714 | 0.4984 | | 0.6725 | 64.0 | 21760 | 0.6749 | 0.5 | | 0.6719 | 65.0 | 22100 | 0.6723 | 0.5 | | 0.6719 | 66.0 | 22440 | 0.6776 | 0.5 | | 0.6714 | 67.0 | 22780 | 0.6716 | 0.5 | | 0.6724 | 68.0 | 23120 | 0.6723 | 0.5 | | 0.6724 | 69.0 | 23460 | 0.6717 | 0.5 | | 0.6705 | 70.0 | 23800 | 0.6712 | 0.4984 | | 0.6722 | 71.0 | 24140 | 0.6725 | 0.5 | | 0.6722 | 72.0 | 24480 | 0.6775 | 0.5 | | 0.6715 | 73.0 | 24820 | 0.6744 | 0.5 | | 0.6708 | 74.0 | 25160 | 0.6762 | 0.5 | | 0.6705 | 75.0 | 25500 | 0.6737 | 0.5 | | 0.6705 | 76.0 | 25840 | 0.6745 | 0.5 | | 0.6698 | 77.0 | 26180 | 0.6717 | 0.5 | | 0.6691 | 78.0 | 26520 | 0.6735 | 0.5 | | 0.6691 | 79.0 | 26860 | 0.6732 | 0.5 | | 0.6693 | 80.0 | 27200 | 0.6732 | 0.5 | ### Framework versions - Transformers 4.26.1 - Pytorch 2.0.1+cu118 - Datasets 2.12.0 - Tokenizers 0.13.3

dkqjrm/20230830102630

dkqjrm

2023-08-30T06:17:08Z

114

0

transformers

[ "transformers", "pytorch", "tensorboard", "bert", "text-classification", "generated_from_trainer", "dataset:super_glue", "license:apache-2.0", "autotrain_compatible", "endpoints_compatible", "region:us" ]

text-classification

2023-08-30T01:26:53Z

--- license: apache-2.0 tags: - generated_from_trainer datasets: - super_glue metrics: - accuracy model-index: - name: '20230830102630' results: [] ---  # 20230830102630 This model is a fine-tuned version of [bert-large-cased](https://huggingface.co/bert-large-cased) on the super_glue dataset. It achieves the following results on the evaluation set: - Loss: 0.6751 - Accuracy: 0.4984 ## Model description More information needed ## Intended uses & limitations More information needed ## Training and evaluation data More information needed ## Training procedure ### Training hyperparameters The following hyperparameters were used during training: - learning_rate: 0.0005 - train_batch_size: 16 - eval_batch_size: 8 - seed: 11 - optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08 - lr_scheduler_type: linear - num_epochs: 80.0 ### Training results | Training Loss | Epoch | Step | Validation Loss | Accuracy | |:-------------:|:-----:|:-----:|:---------------:|:--------:| | No log | 1.0 | 340 | 0.6962 | 0.5 | | 0.7003 | 2.0 | 680 | 0.6657 | 0.5345 | | 0.6959 | 3.0 | 1020 | 0.6703 | 0.5 | | 0.6959 | 4.0 | 1360 | 0.6845 | 0.5 | | 0.6978 | 5.0 | 1700 | 0.6767 | 0.5 | | 0.6854 | 6.0 | 2040 | 0.6876 | 0.5 | | 0.6854 | 7.0 | 2380 | 0.6705 | 0.5 | | 0.6851 | 8.0 | 2720 | 0.6806 | 0.5 | | 0.6845 | 9.0 | 3060 | 0.6881 | 0.5 | | 0.6845 | 10.0 | 3400 | 0.6737 | 0.5 | | 0.6835 | 11.0 | 3740 | 0.6734 | 0.5 | | 0.6821 | 12.0 | 4080 | 0.7058 | 0.5 | | 0.6821 | 13.0 | 4420 | 0.7057 | 0.5 | | 0.682 | 14.0 | 4760 | 0.7057 | 0.5 | | 0.6827 | 15.0 | 5100 | 0.6771 | 0.5 | | 0.6827 | 16.0 | 5440 | 0.6848 | 0.5 | | 0.6803 | 17.0 | 5780 | 0.7044 | 0.5 | | 0.6821 | 18.0 | 6120 | 0.6720 | 0.4984 | | 0.6821 | 19.0 | 6460 | 0.6716 | 0.5 | | 0.6784 | 20.0 | 6800 | 0.6855 | 0.5 | | 0.6821 | 21.0 | 7140 | 0.6705 | 0.5 | | 0.6821 | 22.0 | 7480 | 0.6753 | 0.5 | | 0.6888 | 23.0 | 7820 | 0.6745 | 0.4953 | | 0.6821 | 24.0 | 8160 | 0.6716 | 0.5 | | 0.682 | 25.0 | 8500 | 0.6702 | 0.5 | | 0.682 | 26.0 | 8840 | 0.6791 | 0.5 | | 0.6829 | 27.0 | 9180 | 0.6771 | 0.5 | | 0.6807 | 28.0 | 9520 | 0.6719 | 0.5 | | 0.6807 | 29.0 | 9860 | 0.6739 | 0.5 | | 0.6783 | 30.0 | 10200 | 0.6716 | 0.5 | | 0.6789 | 31.0 | 10540 | 0.6706 | 0.5 | | 0.6789 | 32.0 | 10880 | 0.7163 | 0.5 | | 0.6798 | 33.0 | 11220 | 0.6703 | 0.5 | | 0.6785 | 34.0 | 11560 | 0.6822 | 0.5 | | 0.6785 | 35.0 | 11900 | 0.6715 | 0.5 | | 0.6783 | 36.0 | 12240 | 0.6720 | 0.5 | | 0.6781 | 37.0 | 12580 | 0.6733 | 0.5 | | 0.6781 | 38.0 | 12920 | 0.6707 | 0.5 | | 0.6798 | 39.0 | 13260 | 0.6950 | 0.5 | | 0.6755 | 40.0 | 13600 | 0.6705 | 0.5 | | 0.6755 | 41.0 | 13940 | 0.6715 | 0.5 | | 0.6776 | 42.0 | 14280 | 0.6704 | 0.5 | | 0.6772 | 43.0 | 14620 | 0.6789 | 0.5 | | 0.6772 | 44.0 | 14960 | 0.6707 | 0.5 | | 0.6755 | 45.0 | 15300 | 0.6925 | 0.5 | | 0.6748 | 46.0 | 15640 | 0.6727 | 0.5 | | 0.6748 | 47.0 | 15980 | 0.6801 | 0.5 | | 0.6754 | 48.0 | 16320 | 0.6714 | 0.5 | | 0.6762 | 49.0 | 16660 | 0.6882 | 0.5 | | 0.6753 | 50.0 | 17000 | 0.6710 | 0.5 | | 0.6753 | 51.0 | 17340 | 0.6707 | 0.5 | | 0.6734 | 52.0 | 17680 | 0.6726 | 0.5063 | | 0.678 | 53.0 | 18020 | 0.6727 | 0.5 | | 0.678 | 54.0 | 18360 | 0.6751 | 0.5 | | 0.6719 | 55.0 | 18700 | 0.6712 | 0.5 | | 0.6726 | 56.0 | 19040 | 0.6721 | 0.5 | | 0.6726 | 57.0 | 19380 | 0.6715 | 0.5 | | 0.6732 | 58.0 | 19720 | 0.6717 | 0.5016 | | 0.6736 | 59.0 | 20060 | 0.6819 | 0.5 | | 0.6736 | 60.0 | 20400 | 0.6728 | 0.5141 | | 0.6732 | 61.0 | 20740 | 0.6716 | 0.5016 | | 0.6727 | 62.0 | 21080 | 0.6747 | 0.5 | | 0.6727 | 63.0 | 21420 | 0.6715 | 0.4984 | | 0.6726 | 64.0 | 21760 | 0.6737 | 0.5 | | 0.6721 | 65.0 | 22100 | 0.6724 | 0.5 | | 0.6721 | 66.0 | 22440 | 0.6744 | 0.5 | | 0.6711 | 67.0 | 22780 | 0.6720 | 0.5 | | 0.6725 | 68.0 | 23120 | 0.6722 | 0.4984 | | 0.6725 | 69.0 | 23460 | 0.6722 | 0.4984 | | 0.6713 | 70.0 | 23800 | 0.6722 | 0.4984 | | 0.6708 | 71.0 | 24140 | 0.6743 | 0.5 | | 0.6708 | 72.0 | 24480 | 0.6794 | 0.5 | | 0.6703 | 73.0 | 24820 | 0.6756 | 0.5 | | 0.6702 | 74.0 | 25160 | 0.6760 | 0.5 | | 0.6688 | 75.0 | 25500 | 0.6741 | 0.4984 | | 0.6688 | 76.0 | 25840 | 0.6753 | 0.5 | | 0.67 | 77.0 | 26180 | 0.6730 | 0.4984 | | 0.6688 | 78.0 | 26520 | 0.6751 | 0.4984 | | 0.6688 | 79.0 | 26860 | 0.6750 | 0.4984 | | 0.6685 | 80.0 | 27200 | 0.6751 | 0.4984 | ### Framework versions - Transformers 4.26.1 - Pytorch 2.0.1+cu118 - Datasets 2.12.0 - Tokenizers 0.13.3

fuguizhen/dqn-SpaceInvadersNoFrameskip-v4

fuguizhen

2023-08-30T06:16:35Z

0

stable-baselines3

[ "stable-baselines3", "SpaceInvadersNoFrameskip-v4", "deep-reinforcement-learning", "reinforcement-learning", "model-index", "region:us" ]

reinforcement-learning

2023-08-30T06:16:12Z

--- library_name: stable-baselines3 tags: - SpaceInvadersNoFrameskip-v4 - deep-reinforcement-learning - reinforcement-learning - stable-baselines3 model-index: - name: DQN results: - task: type: reinforcement-learning name: reinforcement-learning dataset: name: SpaceInvadersNoFrameskip-v4 type: SpaceInvadersNoFrameskip-v4 metrics: - type: mean_reward value: 6.50 +/- 10.74 name: mean_reward verified: false --- # **DQN** Agent playing **SpaceInvadersNoFrameskip-v4** This is a trained model of a **DQN** agent playing **SpaceInvadersNoFrameskip-v4** using the [stable-baselines3 library](https://github.com/DLR-RM/stable-baselines3) and the [RL Zoo](https://github.com/DLR-RM/rl-baselines3-zoo). The RL Zoo is a training framework for Stable Baselines3 reinforcement learning agents, with hyperparameter optimization and pre-trained agents included. ## Usage (with SB3 RL Zoo) RL Zoo: https://github.com/DLR-RM/rl-baselines3-zoo SB3: https://github.com/DLR-RM/stable-baselines3 SB3 Contrib: https://github.com/Stable-Baselines-Team/stable-baselines3-contrib Install the RL Zoo (with SB3 and SB3-Contrib): ```bash pip install rl_zoo3 ``` ``` # Download model and save it into the logs/ folder python -m rl_zoo3.load_from_hub --algo dqn --env SpaceInvadersNoFrameskip-v4 -orga fuguizhen -f logs/ python -m rl_zoo3.enjoy --algo dqn --env SpaceInvadersNoFrameskip-v4 -f logs/ ``` If you installed the RL Zoo3 via pip (`pip install rl_zoo3`), from anywhere you can do: ``` python -m rl_zoo3.load_from_hub --algo dqn --env SpaceInvadersNoFrameskip-v4 -orga fuguizhen -f logs/ python -m rl_zoo3.enjoy --algo dqn --env SpaceInvadersNoFrameskip-v4 -f logs/ ``` ## Training (with the RL Zoo) ``` python -m rl_zoo3.train --algo dqn --env SpaceInvadersNoFrameskip-v4 -f logs/ # Upload the model and generate video (when possible) python -m rl_zoo3.push_to_hub --algo dqn --env SpaceInvadersNoFrameskip-v4 -f logs/ -orga fuguizhen ``` ## Hyperparameters ```python OrderedDict([('batch_size', 32), ('buffer_size', 100000), ('env_wrapper', ['stable_baselines3.common.atari_wrappers.AtariWrapper']), ('exploration_final_eps', 0.01), ('exploration_fraction', 0.1), ('frame_stack', 4), ('gradient_steps', 1), ('learning_rate', 0.0001), ('learning_starts', 100000), ('n_timesteps', 100000.0), ('optimize_memory_usage', False), ('policy', 'CnnPolicy'), ('target_update_interval', 1000), ('train_freq', 4), ('normalize', False)]) ``` # Environment Arguments ```python {'render_mode': 'rgb_array'} ```

hellomyoh/nmt-s12000-kullm-polyglot-5.8b-v1

hellomyoh

2023-08-30T06:15:48Z

8

0

transformers

[ "transformers", "pytorch", "gpt_neox", "text-generation", "autotrain_compatible", "text-generation-inference", "endpoints_compatible", "region:us" ]

text-generation

2023-08-30T03:41:44Z

|no|english|korean| --|--|--| |1 | Do you know who i am?| '제가 누군지 아시겠어요?'| |2 |Tired of scrolling through the same posts? When you create an account you’ll always come back to where you left off. With an account you can also be notified of new replies, save bookmarks, and use likes to thank others. We can all work together to make this community great.. | 똑같은 글을 스크롤하는 것에 지쳤나요? 계정을 만들면 항상 처음으로 돌아오게 됩니다. 계정을 만들면 새로운 댓글을 확인할 수 있고 북마크를 저장할 수 있으며 다른 사람들에게 감사할 수 있습니다. 우리는 함께 이 커뮤니티를 멋지게 만들어 나갈 수 있습니다. | |3 | As technology continues to advance, new vulnerabilities emerge and the importance of security becomes increasingly crucial. In this regard, Cho Hong Ki, the Information Security Specialist at the 2bytes, shared valuable knowledge on the significance of game security and the solutions it offers.|기술이 계속 발전함에 따라 새로운 취약점이 발생하고 보안의 중요성이 더욱 커지고 있습니다. 이와 관련해 2bytes의 정보보안 전문가인 조홍기 씨는 게임 보안의 중요성과 그 해결책에 대한 귀중한 지식을 공유했습니다.| |4 |They are gifts to my Queen from the goddess Tiamat herself. They reside in the great city of Tu'narath, awaiting the privilege of battle | 신들의 여왕인 티아맛의 선물입니다. 이들은 투나랏의 거대한 도시에 살고 있으며, 전투의 영광을 누리고 있습니다.| |5|I showed my masterpiece to the grown-ups, and asked them whether the drawing frightened them. But they answered: 'Frighten? Why should any one be frightened by a hat?' | "나는 어른들에게 내 작품을 보여주었고, 그들에게 모자가 무섭지 않으냐고 물었다. 하지만 그들은 대답했다. '무섭다고? 왜 누군가가 무서워하지? 모자가 왜 무서워?"| |6 |New York. CNN. The Federal Aviation Administration has certified for testing a vehicle that a California startup describes as a flying car — the first fully electric vehicle that can both fly and travel on roads to receive US government approval. Alef Automotive said that its vehicle/aircraft, dubbed the “Model A,” is the first flying vehicle that is drivable on public roads and able to park like a normal car. It also has vertical takeoff and landing capabilities. It apparently will be able to carry one or two occupants and will have a road-range of 200 miles and a flying range of 110 miles. The company expects to sell the vehicle for $300,000 each with the first delivery by projected for the end of 2025. The FAA confirmed that it has issued the company a special airworthiness certificate, allowing for limited purposes that include exhibition, research and development. Numerous companies are working on all-electric VTOLs, which stands for vehicle takeoff and landing aircraft. The FAA said that Alef is “not the first aircraft of its kind” to get a special airworthiness certificate. However, Alef noted that its vehicle is different because of its ability to function both on roads and in the air, to appear like a normal car and to park in a normal parking space. | Alef Automotive는 캘리포니아의 한 신생 기업이 묘사한 것처럼 비행할 수 있는 최초의 완전 전기 자동차를 테스트하기 위해 뉴욕의 CNN과 연방 항공국에 의해 인증을 받았다고 말했습니다. Alef Automotive는 이 차량을 "모델 A"라고 부르며 도로에서도 날 수 있고 일반 자동차처럼 주차할 수 있는 최초의 비행 가능한 차량이라고 말했습니다. 수직 이착륙과 도로 주행이 가능하며 한 두 명 정도의 승객을 태울 수 있고 주행 거리는 200마일과 110마일의 비행이 가능할 것으로 예상하고 있습니다. FAA는 이 회사가 2025년 말까지 첫 번째 배송을 완료할 것으로 예상되는 가운데 이 차량을 3억 달러에 판매할 수 있을 것으로 기대하고 있다고 밝혔습니다. FAA는 이 회사에 특별한 인증을 발급하여 전시, 연구 및 개발을 위한 제한된 목적에 한해서만 사용할 수 있도록 했습니다. 많은 회사들이 모든 전기식 비행체를 개발하고 있으며 이는 비행기에 착륙하고 이륙하는 차량을 의미합니다. FAA는 Alef가 도로에서나 공중에서나 정상적인 차량처럼 보이고 일반적인 주차 공간에 주차할 수 있는 차량이라는 점에서 다른 차량과 다르다고 말했습니다.|

trieudemo11/llama_7b_attrb_cate_batch8_len320_gpu8_10

trieudemo11

2023-08-30T06:13:44Z

1

0

peft

[ "peft", "region:us" ]

null

2023-08-30T06:13:28Z

--- library_name: peft --- ## Training procedure ### Framework versions - PEFT 0.6.0.dev0 - PEFT 0.6.0.dev0 - PEFT 0.6.0.dev0 - PEFT 0.6.0.dev0 - PEFT 0.6.0.dev0 - PEFT 0.6.0.dev0 - PEFT 0.6.0.dev0 - PEFT 0.6.0.dev0 - PEFT 0.6.0.dev0 - PEFT 0.6.0.dev0 - PEFT 0.6.0.dev0 - PEFT 0.6.0.dev0 - PEFT 0.6.0.dev0 - PEFT 0.6.0.dev0 - PEFT 0.6.0.dev0 - PEFT 0.6.0.dev0 - PEFT 0.6.0.dev0 - PEFT 0.6.0.dev0 - PEFT 0.6.0.dev0 - PEFT 0.6.0.dev0 - PEFT 0.6.0.dev0 - PEFT 0.6.0.dev0 - PEFT 0.6.0.dev0 - PEFT 0.6.0.dev0 - PEFT 0.6.0.dev0 - PEFT 0.6.0.dev0 - PEFT 0.6.0.dev0 - PEFT 0.6.0.dev0 - PEFT 0.6.0.dev0 - PEFT 0.6.0.dev0 - PEFT 0.6.0.dev0 - PEFT 0.6.0.dev0 - PEFT 0.6.0.dev0 - PEFT 0.6.0.dev0 - PEFT 0.6.0.dev0 - PEFT 0.6.0.dev0 - PEFT 0.6.0.dev0 - PEFT 0.6.0.dev0 - PEFT 0.6.0.dev0 - PEFT 0.6.0.dev0 - PEFT 0.6.0.dev0 - PEFT 0.6.0.dev0 - PEFT 0.6.0.dev0 - PEFT 0.6.0.dev0 - PEFT 0.6.0.dev0 - PEFT 0.6.0.dev0 - PEFT 0.6.0.dev0 - PEFT 0.6.0.dev0 - PEFT 0.6.0.dev0 - PEFT 0.6.0.dev0 - PEFT 0.6.0.dev0 - PEFT 0.6.0.dev0 - PEFT 0.6.0.dev0 - PEFT 0.6.0.dev0 - PEFT 0.6.0.dev0 - PEFT 0.6.0.dev0 - PEFT 0.6.0.dev0 - PEFT 0.6.0.dev0 - PEFT 0.6.0.dev0 - PEFT 0.6.0.dev0 - PEFT 0.6.0.dev0 - PEFT 0.6.0.dev0 - PEFT 0.6.0.dev0 - PEFT 0.6.0.dev0 - PEFT 0.6.0.dev0 - PEFT 0.6.0.dev0 - PEFT 0.6.0.dev0 - PEFT 0.6.0.dev0 - PEFT 0.6.0.dev0

LarryAIDraw/yukina_kiritani_imocho

LarryAIDraw

2023-08-30T05:56:58Z

0

null

[ "license:creativeml-openrail-m", "region:us" ]

null

2023-08-30T05:41:01Z

--- license: creativeml-openrail-m --- https://civitai.com/models/137304/yukina-kiritani-recently-my-sister-is-unusual

LarryAIDraw/Katori-10

LarryAIDraw

2023-08-30T05:54:54Z

0

null

[ "license:creativeml-openrail-m", "region:us" ]

null

2023-08-30T05:42:27Z

--- license: creativeml-openrail-m --- https://civitai.com/models/135972/katori-batsuunsai-bleach

LarryAIDraw/August

LarryAIDraw

2023-08-30T05:53:12Z

0

null

[ "license:creativeml-openrail-m", "region:us" ]

null

2023-08-30T05:36:19Z

--- license: creativeml-openrail-m --- https://civitai.com/models/137145/augustvonparseval-azur-lane

LarryAIDraw/Modernia-FirstAffection-LoRA-v1

LarryAIDraw

2023-08-30T05:50:06Z

0

null

[ "license:creativeml-openrail-m", "region:us" ]

null

2023-08-30T05:34:44Z

--- license: creativeml-openrail-m --- https://civitai.com/models/136485/modernia-first-affection-nikke-goddess-of-victory

LarryAIDraw/InoueOrihime-2k

LarryAIDraw

2023-08-30T05:49:25Z

0

null

[ "license:creativeml-openrail-m", "region:us" ]

null

2023-08-30T05:33:36Z

--- license: creativeml-openrail-m --- https://civitai.com/models/136699/inoue-orihime-from-bleach

LarryAIDraw/djeeta_20230722

LarryAIDraw

2023-08-30T05:40:07Z

0

null

[ "license:creativeml-openrail-m", "region:us" ]

null

2023-08-30T05:40:07Z

--- license: creativeml-openrail-m ---

BM-K/NewsKoT5-small

BM-K

2023-08-30T05:37:14Z

109

0

transformers

[ "transformers", "pytorch", "safetensors", "t5", "text2text-generation", "ko", "autotrain_compatible", "text-generation-inference", "endpoints_compatible", "region:us" ]

text2text-generation

2023-06-01T07:36:27Z

--- language: - ko --- # NewsKoT5 The training data for this T5 model consists of Korean news articles (29GB). However, the performance has not been fine-tuned through the use of small batches and a limited number of training steps, so it may not be fully optimized. ## Quick tour ```python from transformers import AutoTokenizer, T5ForConditionalGeneration tokenizer = AutoTokenizer.from_pretrained("BM-K/NewsKoT5-small") model = T5ForConditionalGeneration.from_pretrained("BM-K/NewsKoT5-small") input_ids = tokenizer("한국형발사체 누리호가 실용급 <extra_id_0> 발사체로서 ‘데뷔’를 성공적으로 <extra_id_1>", return_tensors="pt").input_ids labels = tokenizer("<extra_id_0> 위성 <extra_id_1> 마쳤다 <extra_id_2>", return_tensors="pt").input_ids outputs = model(input_ids=input_ids, labels=labels) ``` ## News Summarization Performance (F1-score) After restoring the model's tokenized output to the original text, Rouge performance was evaluated by comparing it to the reference and hypothesis tokenized using [mecab](https://konlpy.org/ko/v0.4.0/). - Dacon 한국어 문서 생성요약 AI 경진대회 [Dataset](https://dacon.io/competitions/official/235673/overview/description) - Training: 29,432 - Validation: 7,358 - Test: 9,182 | | #Param | rouge-1 |rouge-2|rouge-l| |-------|--------:|--------:|--------:|--------:| | pko-t5-small | 95M | 51.48 | 33.18 | 44.96 | | NewsT5-small | 61M | 52.15 | 33.59 | 45.41 | - AI-Hub 문서요약 텍스트 [Dataset](https://www.aihub.or.kr/aihubdata/data/view.do?currMenu=115&topMenu=100&aihubDataSe=realm&dataSetSn=97) - Training: 245,626 - Validation: 20,296 - Test: 9,931 | | #Param | rouge-1 |rouge-2|rouge-l| |-------|--------:|--------:|--------:|--------:| | pko-t5-small | 95M | 53.44 | 34.03 | 45.36 | | NewsT5-small | 61M | 53.74 | 34.27 | 45.52 | - [pko-t5-small](https://github.com/paust-team/pko-t5)

BM-K/KoMiniLM

BM-K

2023-08-30T05:37:05Z

134

2

transformers

[ "transformers", "pytorch", "safetensors", "bert", "feature-extraction", "arxiv:2002.10957", "text-embeddings-inference", "endpoints_compatible", "region:us" ]

feature-extraction

2022-05-23T04:26:31Z

# KoMiniLM 🐣 Korean mini language model ## Overview Current language models usually consist of hundreds of millions of parameters which brings challenges for fine-tuning and online serving in real-life applications due to latency and capacity constraints. In this project, we release a light weight korean language model to address the aforementioned shortcomings of existing language models. ## Quick tour ```python from transformers import AutoTokenizer, AutoModel tokenizer = AutoTokenizer.from_pretrained("BM-K/KoMiniLM") # 23M model model = AutoModel.from_pretrained("BM-K/KoMiniLM") inputs = tokenizer("안녕 세상아!", return_tensors="pt") outputs = model(**inputs) ``` ## Update history ** Updates on 2022.06.20 ** - Release KoMiniLM-bert-68M ** Updates on 2022.05.24 ** - Release KoMiniLM-bert-23M ## Pre-training `Teacher Model`: [KLUE-BERT(base)](https://github.com/KLUE-benchmark/KLUE) ### Object Self-Attention Distribution and Self-Attention Value-Relation [[Wang et al., 2020]](https://arxiv.org/abs/2002.10957) were distilled from each discrete layer of the teacher model to the student model. Wang et al. distilled in the last layer of the transformer, but that was not the case in this project. ### Data sets |Data|News comments|News article| |:----:|:----:|:----:| |size|10G|10G| ### Config - **KoMiniLM-23M** ```json { "architectures": [ "BertForPreTraining" ], "attention_probs_dropout_prob": 0.1, "classifier_dropout": null, "hidden_act": "gelu", "hidden_dropout_prob": 0.1, "hidden_size": 384, "initializer_range": 0.02, "intermediate_size": 1536, "layer_norm_eps": 1e-12, "max_position_embeddings": 512, "model_type": "bert", "num_attention_heads": 12, "num_hidden_layers": 6, "output_attentions": true, "pad_token_id": 0, "position_embedding_type": "absolute", "return_dict": false, "torch_dtype": "float32", "transformers_version": "4.13.0", "type_vocab_size": 2, "use_cache": true, "vocab_size": 32000 } ``` ### Performance on subtasks - The results of our fine-tuning experiments are an average of 3 runs for each task. ``` cd KoMiniLM-Finetune bash scripts/run_all_kominilm.sh ``` || #Param | Average | NSMC (Acc) | Naver NER (F1) | PAWS (Acc) | KorNLI (Acc) | KorSTS (Spearman) | Question Pair (Acc) | KorQuaD (Dev) (EM/F1) | |:----:|:----:|:----:|:----:|:----:|:----:|:----:|:----:|:----:|:----:| |KoBERT(KLUE)| 110M | 86.84 | 90.20±0.07 | 87.11±0.05 | 81.36±0.21 | 81.06±0.33 | 82.47±0.14 | 95.03±0.44 | 84.43±0.18 / 93.05±0.04 | |KcBERT| 108M | 78.94 | 89.60±0.10 | 84.34±0.13 | 67.02±0.42| 74.17±0.52 | 76.57±0.51 | 93.97±0.27 | 60.87±0.27 / 85.01±0.14 | |KoBERT(SKT)| 92M | 79.73 | 89.28±0.42 | 87.54±0.04 | 80.93±0.91 | 78.18±0.45 | 75.98±2.81 | 94.37±0.31 | 51.94±0.60 / 79.69±0.66 | |DistilKoBERT| 28M | 74.73 | 88.39±0.08 | 84.22±0.01 | 61.74±0.45 | 70.22±0.14 | 72.11±0.27 | 92.65±0.16 | 52.52±0.48 / 76.00±0.71 | | | | | | | | | | | |**KoMiniLM†**| **68M** | 85.90 | 89.84±0.02 | 85.98±0.09 | 80.78±0.30 | 79.28±0.17 | 81.00±0.07 | 94.89±0.37 | 83.27±0.08 / 92.08±0.06 | |**KoMiniLM†**| **23M** | 84.79 | 89.67±0.03 | 84.79±0.09 | 78.67±0.45 | 78.10±0.07 | 78.90±0.11 | 94.81±0.12 | 82.11±0.42 / 91.21±0.29 | - [NSMC](https://github.com/e9t/nsmc) (Naver Sentiment Movie Corpus) - [Naver NER](https://github.com/naver/nlp-challenge) (NER task on Naver NLP Challenge 2018) - [PAWS](https://github.com/google-research-datasets/paws) (Korean Paraphrase Adversaries from Word Scrambling) - [KorNLI/KorSTS](https://github.com/kakaobrain/KorNLUDatasets) (Korean Natural Language Understanding) - [Question Pair](https://github.com/songys/Question_pair) (Paired Question) - [KorQuAD](https://korquad.github.io/) (The Korean Question Answering Dataset) <img src = "https://user-images.githubusercontent.com/55969260/174229747-279122dc-9d27-4da9-a6e7-f9f1fe1651f7.png"> ### User Contributed Examples - ## Reference - [KLUE BERT](https://github.com/KLUE-benchmark/KLUE) - [KcBERT](https://github.com/Beomi/KcBERT) - [SKT KoBERT](https://github.com/SKTBrain/KoBERT) - [DistilKoBERT](https://github.com/monologg/DistilKoBERT) - [lassl](https://github.com/lassl/lassl)

BM-K/KoDiffCSE-RoBERTa

BM-K

2023-08-30T05:36:55Z

27,758

4

transformers

[ "transformers", "pytorch", "safetensors", "roberta", "feature-extraction", "arxiv:2204.10298", "arxiv:2004.03289", "arxiv:2105.09680", "text-embeddings-inference", "endpoints_compatible", "region:us" ]

feature-extraction

2023-02-28T04:17:36Z

# KoDiffCSE Difference-based Contrastive Learning for Korean Sentence Embeddings - [DiffCSE-[NAACL 2022]](https://arxiv.org/abs/2204.10298) - [[Github]](https://github.com/voidism/DiffCSE) Official implementation of DiffCSE <img src=https://user-images.githubusercontent.com/55969260/201829550-9674a3ac-cb9b-4e17-b777-7d96fdf5c633.png> ## Quick tour ```python import torch from transformers import AutoModel, AutoTokenizer def cal_score(a, b): if len(a.shape) == 1: a = a.unsqueeze(0) if len(b.shape) == 1: b = b.unsqueeze(0) a_norm = a / a.norm(dim=1)[:, None] b_norm = b / b.norm(dim=1)[:, None] return torch.mm(a_norm, b_norm.transpose(0, 1)) * 100 model = AutoModel.from_pretrained('BM-K/KoDiffCSE-RoBERTa') tokenizer = AutoTokenizer.from_pretrained('BM-K/KoDiffCSE-RoBERTa') sentences = ['치타가 들판을 가로 질러 먹이를 쫓는다.', '치타 한 마리가 먹이 뒤에서 달리고 있다.', '원숭이 한 마리가 드럼을 연주한다.'] inputs = tokenizer(sentences, padding=True, truncation=True, return_tensors="pt") embeddings, _ = model(**inputs, return_dict=False) score01 = cal_score(embeddings[0][0], embeddings[1][0]) # 84.56 # '치타가 들판을 가로 질러 먹이를 쫓는다.' @ '치타 한 마리가 먹이 뒤에서 달리고 있다.' score02 = cal_score(embeddings[0][0], embeddings[2][0]) # 48.06 # '치타가 들판을 가로 질러 먹이를 쫓는다.' @ '원숭이 한 마리가 드럼을 연주한다.' ``` ## Setups [![Python](https://img.shields.io/badge/python-3.8.5-blue?logo=python&logoColor=FED643)](https://www.python.org/downloads/release/python-385/) [![Pytorch](https://img.shields.io/badge/pytorch-1.7.1-red?logo=pytorch)](https://pytorch.org/get-started/previous-versions/) ## Encoder Models Baseline encoders used for korean sentence embedding - [KLUE-PLMs](https://github.com/KLUE-benchmark/KLUE/blob/main/README.md) | Model | Embedding size | Hidden size | # Layers | # Heads | |----------------------|----------------|-------------|----------|---------| | KLUE-BERT-base | 768 | 768 | 12 | 12 | | KLUE-RoBERTa-base | 768 | 768 | 12 | 12 | > **Warning** > Large pre-trained models need a lot of GPU memory to train ## Datasets The data must exist in the "--path_to_data" folder - [wiki-corpus](https://github.com/jeongukjae/korean-wikipedia-corpus) (Unsupervised Training) - [KorSTS](https://github.com/kakaobrain/KorNLUDatasets) (Validation & Testing) ## Training - unsupervised ``` python main.py \ --model klue/roberta-base \ --generator_name klue/roberta-small \ --multi_gpu True \ --train True \ --test False \ --max_len 64 \ --batch_size 256 \ --epochs 1 \ --eval_steps 125 \ --lr 0.00005 \ --masking_ratio 0.15 \ --lambda_weight 0.005 \ --warmup_ratio 0.05 \ --temperature 0.05 \ --path_to_data Dataset/ \ --train_data wiki_corpus_examples.txt \ --valid_data valid_sts.tsv \ --ckpt best_checkpoint.pt ``` ``` bash run_diff.sh ``` > **Note** > Using roberta as an encoder is beneficial for training because the KoBERT model cannot build a small-sized generator. ## Evaluation ``` python main.py \ --model klue/roberta-base \ --generator klue/roberta-small \ --train False \ --test True \ --max_len 64 \ --batch_size 256 \ --path_to_data Dataset/ \ --test_data test_sts.tsv \ --path_to_saved_model output/best_checkpoint.pt ``` ## Performance - unsupervised | Model | Average | Cosine Pearson | Cosine Spearman | Euclidean Pearson | Euclidean Spearman | Manhattan Pearson | Manhattan Spearman | Dot Pearson | Dot Spearman | |------------------------|:----:|:----:|:----:|:----:|:----:|:----:|:----:|:----:|:----:| | KoSRoBERTa-base† | N/A | N/A | 48.96 | N/A | N/A | N/A | N/A | N/A | N/A | | KoSRoBERTa-large† | N/A | N/A | 51.35 | N/A | N/A | N/A | N/A | N/A | N/A | | | | | | | | | | | | | KoSimCSE-BERT | 74.08 | 74.92 | 73.98 | 74.15 | 74.22 | 74.07 | 74.07 | 74.15 | 73.14 | | KoSimCSE-RoBERTa | 75.27 | 75.93 | 75.00 | 75.28 | 75.01 | 75.17 | 74.83 | 75.95 | 75.01 | | | | | | | | | | | | | KoDiffCSE-RoBERTa | 77.17 | 77.73 | 76.96 | 77.21 | 76.89 | 77.11 | 76.81 | 77.74 | 76.97 | - [Korean-SRoBERTa†](https://arxiv.org/abs/2004.03289) ## License This work is licensed under a <a rel="license" href="http://creativecommons.org/licenses/by-sa/4.0/">Creative Commons Attribution-ShareAlike 4.0 International License</a>. <a rel="license" href="http://creativecommons.org/licenses/by-sa/4.0/"><img alt="Creative Commons License" style="border-width:0" src="https://i.creativecommons.org/l/by-sa/4.0/88x31.png" /></a> ## References ```bibtex @inproceedings{chuang2022diffcse, title={{DiffCSE}: Difference-based Contrastive Learning for Sentence Embeddings}, author={Chuang, Yung-Sung and Dangovski, Rumen and Luo, Hongyin and Zhang, Yang and Chang, Shiyu and Soljacic, Marin and Li, Shang-Wen and Yih, Wen-tau and Kim, Yoon and Glass, James}, booktitle={Annual Conference of the North American Chapter of the Association for Computational Linguistics (NAACL)}, year={2022} } @misc{park2021klue, title={KLUE: Korean Language Understanding Evaluation}, author={Sungjoon Park and Jihyung Moon and Sungdong Kim and Won Ik Cho and Jiyoon Han and Jangwon Park and Chisung Song and Junseong Kim and Yongsook Song and Taehwan Oh and Joohong Lee and Juhyun Oh and Sungwon Lyu and Younghoon Jeong and Inkwon Lee and Sangwoo Seo and Dongjun Lee and Hyunwoo Kim and Myeonghwa Lee and Seongbo Jang and Seungwon Do and Sunkyoung Kim and Kyungtae Lim and Jongwon Lee and Kyumin Park and Jamin Shin and Seonghyun Kim and Lucy Park and Alice Oh and Jungwoo Ha and Kyunghyun Cho}, year={2021}, eprint={2105.09680}, archivePrefix={arXiv}, primaryClass={cs.CL} } @article{ham2020kornli, title={KorNLI and KorSTS: New Benchmark Datasets for Korean Natural Language Understanding}, author={Ham, Jiyeon and Choe, Yo Joong and Park, Kyubyong and Choi, Ilji and Soh, Hyungjoon}, journal={arXiv preprint arXiv:2004.03289}, year={2020} } ```

Shivam098/opt-Translator

Shivam098

2023-08-30T05:16:01Z

79

0

transformers

[ "transformers", "pytorch", "mbart", "text2text-generation", "generated_from_trainer", "dataset:opus100", "base_model:Shivam098/opt-translation", "base_model:quantized:Shivam098/opt-translation", "autotrain_compatible", "endpoints_compatible", "4-bit", "gptq", "region:us" ]

text2text-generation

2023-08-29T13:17:01Z

--- base_model: Shivam098/opt-translation tags: - generated_from_trainer datasets: - opus100 model-index: - name: opt-Translator results: [] ---  # opt-Translator This model is a fine-tuned version of [Shivam098/opt-translation](https://huggingface.co/Shivam098/opt-translation) on the opus100 dataset. ## Model description More information needed ## Intended uses & limitations More information needed ## Training and evaluation data More information needed ## Training procedure ### Training hyperparameters The following hyperparameters were used during training: - learning_rate: 5e-05 - train_batch_size: 2 - eval_batch_size: 8 - seed: 42 - optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08 - lr_scheduler_type: linear - training_steps: 1000 ### Training results ### Framework versions - Transformers 4.33.0.dev0 - Pytorch 2.0.1+cu118 - Datasets 2.14.4 - Tokenizers 0.13.3

AmlanSamanta/ludwig-webinar

AmlanSamanta

2023-08-30T04:40:12Z

1

0

peft

[ "peft", "region:us" ]

null

2023-08-30T04:40:06Z

--- library_name: peft --- ## Training procedure The following `bitsandbytes` quantization config was used during training: - quant_method: bitsandbytes - load_in_8bit: False - load_in_4bit: True - llm_int8_threshold: 6.0 - llm_int8_skip_modules: None - llm_int8_enable_fp32_cpu_offload: False - llm_int8_has_fp16_weight: False - bnb_4bit_quant_type: nf4 - bnb_4bit_use_double_quant: True - bnb_4bit_compute_dtype: float16 The following `bitsandbytes` quantization config was used during training: - quant_method: bitsandbytes - load_in_8bit: False - load_in_4bit: True - llm_int8_threshold: 6.0 - llm_int8_skip_modules: None - llm_int8_enable_fp32_cpu_offload: False - llm_int8_has_fp16_weight: False - bnb_4bit_quant_type: nf4 - bnb_4bit_use_double_quant: True - bnb_4bit_compute_dtype: float16 ### Framework versions - PEFT 0.5.0 - PEFT 0.5.0

MohanaSri/ppo-Pendulum-v2

MohanaSri

2023-08-30T03:51:27Z

1

0

stable-baselines3

[ "stable-baselines3", "Pendulum-v1", "deep-reinforcement-learning", "reinforcement-learning", "model-index", "region:us" ]

reinforcement-learning

2023-08-30T03:51:08Z

--- library_name: stable-baselines3 tags: - Pendulum-v1 - deep-reinforcement-learning - reinforcement-learning - stable-baselines3 model-index: - name: PPO results: - task: type: reinforcement-learning name: reinforcement-learning dataset: name: Pendulum-v1 type: Pendulum-v1 metrics: - type: mean_reward value: -1348.57 +/- 247.55 name: mean_reward verified: false --- # **PPO** Agent playing **Pendulum-v1** This is a trained model of a **PPO** agent playing **Pendulum-v1** using the [stable-baselines3 library](https://github.com/DLR-RM/stable-baselines3). ## Usage (with Stable-baselines3) TODO: Add your code ```python from stable_baselines3 import ... from huggingface_sb3 import load_from_hub ... ```

RajkNakka/a2c-PandaReachDense-v3

RajkNakka

2023-08-30T03:41:38Z

3

0

stable-baselines3

[ "stable-baselines3", "PandaReachDense-v3", "deep-reinforcement-learning", "reinforcement-learning", "model-index", "region:us" ]

reinforcement-learning

2023-08-30T03:36:11Z

--- library_name: stable-baselines3 tags: - PandaReachDense-v3 - deep-reinforcement-learning - reinforcement-learning - stable-baselines3 model-index: - name: A2C results: - task: type: reinforcement-learning name: reinforcement-learning dataset: name: PandaReachDense-v3 type: PandaReachDense-v3 metrics: - type: mean_reward value: -0.27 +/- 0.14 name: mean_reward verified: false --- # **A2C** Agent playing **PandaReachDense-v3** This is a trained model of a **A2C** agent playing **PandaReachDense-v3** using the [stable-baselines3 library](https://github.com/DLR-RM/stable-baselines3). ## Usage (with Stable-baselines3) TODO: Add your code ```python from stable_baselines3 import ... from huggingface_sb3 import load_from_hub ... ```

nightdude/config_611

nightdude

2023-08-30T03:40:54Z

0

peft

[ "peft", "region:us" ]

null

2023-08-30T03:40:29Z

--- library_name: peft --- ## Training procedure The following `bitsandbytes` quantization config was used during training: - quant_method: bitsandbytes - load_in_8bit: False - load_in_4bit: True - llm_int8_threshold: 6.0 - llm_int8_skip_modules: None - llm_int8_enable_fp32_cpu_offload: False - llm_int8_has_fp16_weight: False - bnb_4bit_quant_type: nf4 - bnb_4bit_use_double_quant: True - bnb_4bit_compute_dtype: bfloat16 ### Framework versions - PEFT 0.5.0.dev0

sang-kyung/train_monster_toy1

sang-kyung

2023-08-30T03:38:19Z

1

0

diffusers

[ "diffusers", "stable-diffusion", "stable-diffusion-diffusers", "text-to-image", "lora", "base_model:stabilityai/stable-diffusion-2-1-base", "base_model:adapter:stabilityai/stable-diffusion-2-1-base", "license:creativeml-openrail-m", "region:us" ]

text-to-image

2023-08-30T03:29:39Z

--- license: creativeml-openrail-m base_model: stabilityai/stable-diffusion-2-1-base instance_prompt: a photo of sks monster toy tags: - stable-diffusion - stable-diffusion-diffusers - text-to-image - diffusers - lora inference: true --- # LoRA DreamBooth - sang-kyung/train_monster_toy1 These are LoRA adaption weights for stabilityai/stable-diffusion-2-1-base. The weights were trained on a photo of sks monster toy using [DreamBooth](https://dreambooth.github.io/). You can find some example images in the following. ![img_0](./image_0.png) ![img_1](./image_1.png) ![img_2](./image_2.png) ![img_3](./image_3.png) LoRA for the text encoder was enabled: False.

nightdude/config_601

nightdude

2023-08-30T03:35:59Z

0

peft

[ "peft", "region:us" ]

null

2023-08-30T03:34:59Z

--- library_name: peft --- ## Training procedure The following `bitsandbytes` quantization config was used during training: - quant_method: bitsandbytes - load_in_8bit: False - load_in_4bit: True - llm_int8_threshold: 6.0 - llm_int8_skip_modules: None - llm_int8_enable_fp32_cpu_offload: False - llm_int8_has_fp16_weight: False - bnb_4bit_quant_type: nf4 - bnb_4bit_use_double_quant: True - bnb_4bit_compute_dtype: bfloat16 ### Framework versions - PEFT 0.5.0.dev0

gigant/distilhubert-audio-course-finetuned-gtzan-v5

gigant

2023-08-30T03:32:09Z

159

0

transformers

[ "transformers", "pytorch", "hubert", "audio-classification", "generated_from_trainer", "dataset:marsyas/gtzan", "base_model:ntu-spml/distilhubert", "base_model:finetune:ntu-spml/distilhubert", "license:apache-2.0", "model-index", "endpoints_compatible", "region:us" ]

audio-classification

2023-08-29T23:41:25Z

--- license: apache-2.0 base_model: ntu-spml/distilhubert tags: - generated_from_trainer datasets: - marsyas/gtzan metrics: - accuracy model-index: - name: distilhubert-audio-course-finetuned-gtzan-v5 results: - task: name: Audio Classification type: audio-classification dataset: name: GTZAN type: marsyas/gtzan config: all split: train args: all metrics: - name: Accuracy type: accuracy value: 0.87 ---  # distilhubert-audio-course-finetuned-gtzan-v5 This model is a fine-tuned version of [ntu-spml/distilhubert](https://huggingface.co/ntu-spml/distilhubert) on the GTZAN dataset. It achieves the following results on the evaluation set: - Loss: 0.9236 - Accuracy: 0.87 ## Model description More information needed ## Intended uses & limitations More information needed ## Training and evaluation data More information needed ## Training procedure ### Training hyperparameters The following hyperparameters were used during training: - learning_rate: 0.0001 - train_batch_size: 8 - eval_batch_size: 8 - seed: 42 - gradient_accumulation_steps: 2 - total_train_batch_size: 16 - optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08 - lr_scheduler_type: linear - lr_scheduler_warmup_ratio: 0.7 - num_epochs: 50 ### Training results | Training Loss | Epoch | Step | Validation Loss | Accuracy | |:-------------:|:-----:|:----:|:---------------:|:--------:| | 2.2989 | 0.99 | 56 | 2.2882 | 0.11 | | 2.2716 | 2.0 | 113 | 2.2469 | 0.31 | | 2.1919 | 2.99 | 169 | 2.1317 | 0.4 | | 2.0117 | 4.0 | 226 | 1.9244 | 0.53 | | 1.7966 | 4.99 | 282 | 1.7315 | 0.65 | | 1.6379 | 6.0 | 339 | 1.5920 | 0.59 | | 1.4496 | 6.99 | 395 | 1.3539 | 0.71 | | 1.3264 | 8.0 | 452 | 1.1879 | 0.7 | | 1.0601 | 8.99 | 508 | 1.1342 | 0.7 | | 0.9737 | 10.0 | 565 | 0.9209 | 0.79 | | 0.7915 | 10.99 | 621 | 0.8768 | 0.74 | | 0.6432 | 12.0 | 678 | 0.8060 | 0.8 | | 0.5217 | 12.99 | 734 | 0.6562 | 0.85 | | 0.3335 | 14.0 | 791 | 0.7744 | 0.76 | | 0.2866 | 14.99 | 847 | 0.6969 | 0.82 | | 0.1425 | 16.0 | 904 | 0.6378 | 0.82 | | 0.1278 | 16.99 | 960 | 0.6972 | 0.82 | | 0.0706 | 18.0 | 1017 | 0.7328 | 0.84 | | 0.0301 | 18.99 | 1073 | 0.9245 | 0.76 | | 0.0379 | 20.0 | 1130 | 0.8437 | 0.85 | | 0.0147 | 20.99 | 1186 | 0.7489 | 0.83 | | 0.0067 | 22.0 | 1243 | 0.8975 | 0.83 | | 0.0049 | 22.99 | 1299 | 1.1788 | 0.81 | | 0.0038 | 24.0 | 1356 | 1.1146 | 0.81 | | 0.0028 | 24.99 | 1412 | 1.0270 | 0.85 | | 0.0027 | 26.0 | 1469 | 1.0634 | 0.83 | | 0.0024 | 26.99 | 1525 | 1.0220 | 0.84 | | 0.0023 | 28.0 | 1582 | 1.0282 | 0.83 | | 0.0487 | 28.99 | 1638 | 1.0735 | 0.82 | | 0.0458 | 30.0 | 1695 | 1.1198 | 0.82 | | 0.2453 | 30.99 | 1751 | 1.1154 | 0.81 | | 0.0552 | 32.0 | 1808 | 1.1630 | 0.79 | | 0.1202 | 32.99 | 1864 | 1.2746 | 0.81 | | 0.2709 | 34.0 | 1921 | 1.3797 | 0.79 | | 0.275 | 34.99 | 1977 | 1.5372 | 0.75 | | 0.1268 | 36.0 | 2034 | 0.8140 | 0.86 | | 0.1582 | 36.99 | 2090 | 1.4153 | 0.77 | | 0.0054 | 38.0 | 2147 | 1.3796 | 0.79 | | 0.0299 | 38.99 | 2203 | 1.3653 | 0.78 | | 0.0199 | 40.0 | 2260 | 0.9987 | 0.87 | | 0.0021 | 40.99 | 2316 | 1.0689 | 0.84 | | 0.0007 | 42.0 | 2373 | 1.0383 | 0.85 | | 0.0006 | 42.99 | 2429 | 1.0493 | 0.84 | | 0.0006 | 44.0 | 2486 | 1.0744 | 0.85 | | 0.0005 | 44.99 | 2542 | 0.9151 | 0.86 | | 0.0004 | 46.0 | 2599 | 0.8946 | 0.87 | | 0.01 | 46.99 | 2655 | 0.8960 | 0.88 | | 0.0073 | 48.0 | 2712 | 0.9485 | 0.87 | | 0.0004 | 48.99 | 2768 | 0.9247 | 0.87 | | 0.0004 | 49.56 | 2800 | 0.9236 | 0.87 | ### Framework versions - Transformers 4.32.1 - Pytorch 2.0.1+cu117 - Datasets 2.14.4 - Tokenizers 0.13.3

abeiler/huggingface-goatLora-goatV10-testData

abeiler

2023-08-30T03:19:52Z

0

null

[ "pytorch", "tensorboard", "generated_from_trainer", "region:us" ]

null

2023-08-30T03:07:07Z

--- tags: - generated_from_trainer model-index: - name: huggingface-goatLora-goatV10-testData results: [] ---  # huggingface-goatLora-goatV10-testData This model is a fine-tuned version of [meta-llama/Llama-2-7b-hf](https://huggingface.co/meta-llama/Llama-2-7b-hf) on an unknown dataset. ## Model description More information needed ## Intended uses & limitations More information needed ## Training and evaluation data More information needed ## Training procedure ### Training hyperparameters The following hyperparameters were used during training: - learning_rate: 0.0003 - train_batch_size: 16 - eval_batch_size: 8 - seed: 42 - gradient_accumulation_steps: 8 - total_train_batch_size: 128 - optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08 - lr_scheduler_type: linear - lr_scheduler_warmup_steps: 100 - num_epochs: 1 - mixed_precision_training: Native AMP ### Training results ### Framework versions - Transformers 4.28.1 - Pytorch 2.0.0 - Datasets 2.12.0 - Tokenizers 0.13.3

Adrian101-hnd/ppo-Huggy

Adrian101-hnd

2023-08-30T03:13:37Z

1

0

ml-agents

[ "ml-agents", "tensorboard", "onnx", "Huggy", "deep-reinforcement-learning", "reinforcement-learning", "ML-Agents-Huggy", "region:us" ]

reinforcement-learning

2023-08-30T03:13:32Z

--- library_name: ml-agents tags: - Huggy - deep-reinforcement-learning - reinforcement-learning - ML-Agents-Huggy --- # **ppo** Agent playing **Huggy** This is a trained model of a **ppo** agent playing **Huggy** using the [Unity ML-Agents Library](https://github.com/Unity-Technologies/ml-agents). ## Usage (with ML-Agents) The Documentation: https://unity-technologies.github.io/ml-agents/ML-Agents-Toolkit-Documentation/ We wrote a complete tutorial to learn to train your first agent using ML-Agents and publish it to the Hub: - A *short tutorial* where you teach Huggy the Dog 🐶 to fetch the stick and then play with him directly in your browser: https://huggingface.co/learn/deep-rl-course/unitbonus1/introduction - A *longer tutorial* to understand how works ML-Agents: https://huggingface.co/learn/deep-rl-course/unit5/introduction ### Resume the training ```bash mlagents-learn <your_configuration_file_path.yaml> --run-id=<run_id> --resume ``` ### Watch your Agent play You can watch your agent **playing directly in your browser** 1. If the environment is part of ML-Agents official environments, go to https://huggingface.co/unity 2. Step 1: Find your model_id: Adrian101-hnd/ppo-Huggy 3. Step 2: Select your *.nn /*.onnx file 4. Click on Watch the agent play 👀

hecool108/ct-p1

hecool108

2023-08-30T03:05:55Z

2

0

diffusers

[ "diffusers", "safetensors", "text-generation-inference", "ctee", "text-to-image", "en", "license:openrail", "region:us" ]

text-to-image

2023-08-30T02:58:26Z

--- license: openrail language: - en tags: - text-generation-inference - ctee pipeline_tag: text-to-image ---

uer/chinese_roberta_L-12_H-256

uer

2023-08-30T02:30:45Z

121

0

transformers

[ "transformers", "pytorch", "tf", "jax", "bert", "fill-mask", "zh", "dataset:CLUECorpusSmall", "arxiv:1909.05658", "arxiv:2212.06385", "arxiv:1908.08962", "arxiv:1810.04805", "arxiv:1907.11692", "autotrain_compatible", "endpoints_compatible", "region:us" ]

fill-mask

2022-03-02T23:29:05Z

--- language: zh datasets: CLUECorpusSmall widget: - text: "北京是[MASK]国的首都。" --- # Chinese RoBERTa Miniatures ## Model description This is the set of 24 Chinese RoBERTa models pre-trained by [UER-py](https://github.com/dbiir/UER-py/), which is introduced in [this paper](https://arxiv.org/abs/1909.05658). Besides, the models could also be pre-trained by [TencentPretrain](https://github.com/Tencent/TencentPretrain) introduced in [this paper](https://arxiv.org/abs/2212.06385), which inherits UER-py to support models with parameters above one billion, and extends it to a multimodal pre-training framework. [Turc et al.](https://arxiv.org/abs/1908.08962) have shown that the standard BERT recipe is effective on a wide range of model sizes. Following their paper, we released the 24 Chinese RoBERTa models. In order to facilitate users in reproducing the results, we used a publicly available corpus and provided all training details. You can download the 24 Chinese RoBERTa miniatures either from the [UER-py Modelzoo page](https://github.com/dbiir/UER-py/wiki/Modelzoo), or via HuggingFace from the links below: | | H=128 | H=256 | H=512 | H=768 | | -------- | :-----------------------: | :-----------------------: | :-------------------------: | :-------------------------: | | **L=2** | [**2/128 (Tiny)**][2_128] | [2/256][2_256] | [2/512][2_512] | [2/768][2_768] | | **L=4** | [4/128][4_128] | [**4/256 (Mini)**][4_256] | [**4/512 (Small)**][4_512] | [4/768][4_768] | | **L=6** | [6/128][6_128] | [6/256][6_256] | [6/512][6_512] | [6/768][6_768] | | **L=8** | [8/128][8_128] | [8/256][8_256] | [**8/512 (Medium)**][8_512] | [8/768][8_768] | | **L=10** | [10/128][10_128] | [10/256][10_256] | [10/512][10_512] | [10/768][10_768] | | **L=12** | [12/128][12_128] | [12/256][12_256] | [12/512][12_512] | [**12/768 (Base)**][12_768] | Here are scores on the devlopment set of six Chinese tasks: | Model | Score | book_review | chnsenticorp | lcqmc | tnews(CLUE) | iflytek(CLUE) | ocnli(CLUE) | | -------------- | :---: | :----: | :----------: | :---: | :---------: | :-----------: | :---------: | | RoBERTa-Tiny | 72.3 | 83.4 | 91.4 | 81.8 | 62.0 | 55.0 | 60.3 | | RoBERTa-Mini | 75.9 | 85.7 | 93.7 | 86.1 | 63.9 | 58.3 | 67.4 | | RoBERTa-Small | 76.9 | 87.5 | 93.4 | 86.5 | 65.1 | 59.4 | 69.7 | | RoBERTa-Medium | 78.0 | 88.7 | 94.8 | 88.1 | 65.6 | 59.5 | 71.2 | | RoBERTa-Base | 79.7 | 90.1 | 95.2 | 89.2 | 67.0 | 60.9 | 75.5 | For each task, we selected the best fine-tuning hyperparameters from the lists below, and trained with the sequence length of 128: - epochs: 3, 5, 8 - batch sizes: 32, 64 - learning rates: 3e-5, 1e-4, 3e-4 ## How to use You can use this model directly with a pipeline for masked language modeling (take the case of RoBERTa-Medium): ```python >>> from transformers import pipeline >>> unmasker = pipeline('fill-mask', model='uer/chinese_roberta_L-8_H-512') >>> unmasker("中国的首都是[MASK]京。") [ {'sequence': '[CLS] 中国的首都是北京。 [SEP]', 'score': 0.8701988458633423, 'token': 1266, 'token_str': '北'}, {'sequence': '[CLS] 中国的首都是南京。 [SEP]', 'score': 0.1194809079170227, 'token': 1298, 'token_str': '南'}, {'sequence': '[CLS] 中国的首都是东京。 [SEP]', 'score': 0.0037803512532263994, 'token': 691, 'token_str': '东'}, {'sequence': '[CLS] 中国的首都是普京。 [SEP]', 'score': 0.0017127094324678183, 'token': 3249, 'token_str': '普'}, {'sequence': '[CLS] 中国的首都是望京。 [SEP]', 'score': 0.001687526935711503, 'token': 3307, 'token_str': '望'} ] ``` Here is how to use this model to get the features of a given text in PyTorch: ```python from transformers import BertTokenizer, BertModel tokenizer = BertTokenizer.from_pretrained('uer/chinese_roberta_L-8_H-512') model = BertModel.from_pretrained("uer/chinese_roberta_L-8_H-512") text = "用你喜欢的任何文本替换我。" encoded_input = tokenizer(text, return_tensors='pt') output = model(**encoded_input) ``` and in TensorFlow: ```python from transformers import BertTokenizer, TFBertModel tokenizer = BertTokenizer.from_pretrained('uer/chinese_roberta_L-8_H-512') model = TFBertModel.from_pretrained("uer/chinese_roberta_L-8_H-512") text = "用你喜欢的任何文本替换我。" encoded_input = tokenizer(text, return_tensors='tf') output = model(encoded_input) ``` ## Training data [CLUECorpusSmall](https://github.com/CLUEbenchmark/CLUECorpus2020/) is used as training data. We found that models pre-trained on CLUECorpusSmall outperform those pre-trained on CLUECorpus2020, although CLUECorpus2020 is much larger than CLUECorpusSmall. ## Training procedure Models are pre-trained by [UER-py](https://github.com/dbiir/UER-py/) on [Tencent Cloud](https://cloud.tencent.com/). We pre-train 1,000,000 steps with a sequence length of 128 and then pre-train 250,000 additional steps with a sequence length of 512. We use the same hyper-parameters on different model sizes. Taking the case of RoBERTa-Medium Stage1: ``` python3 preprocess.py --corpus_path corpora/cluecorpussmall.txt \ --vocab_path models/google_zh_vocab.txt \ --dataset_path cluecorpussmall_seq128_dataset.pt \ --processes_num 32 --seq_length 128 \ --dynamic_masking --data_processor mlm ``` ``` python3 pretrain.py --dataset_path cluecorpussmall_seq128_dataset.pt \ --vocab_path models/google_zh_vocab.txt \ --config_path models/bert/medium_config.json \ --output_model_path models/cluecorpussmall_roberta_medium_seq128_model.bin \ --world_size 8 --gpu_ranks 0 1 2 3 4 5 6 7 \ --total_steps 1000000 --save_checkpoint_steps 100000 --report_steps 50000 \ --learning_rate 1e-4 --batch_size 64 \ --data_processor mlm --target mlm ``` Stage2: ``` python3 preprocess.py --corpus_path corpora/cluecorpussmall.txt \ --vocab_path models/google_zh_vocab.txt \ --dataset_path cluecorpussmall_seq512_dataset.pt \ --processes_num 32 --seq_length 512 \ --dynamic_masking --data_processor mlm ``` ``` python3 pretrain.py --dataset_path cluecorpussmall_seq512_dataset.pt \ --vocab_path models/google_zh_vocab.txt \ --pretrained_model_path models/cluecorpussmall_roberta_medium_seq128_model.bin-1000000 \ --config_path models/bert/medium_config.json \ --output_model_path models/cluecorpussmall_roberta_medium_seq512_model.bin \ --world_size 8 --gpu_ranks 0 1 2 3 4 5 6 7 \ --total_steps 250000 --save_checkpoint_steps 50000 --report_steps 10000 \ --learning_rate 5e-5 --batch_size 16 \ --data_processor mlm --target mlm ``` Finally, we convert the pre-trained model into Huggingface's format: ``` python3 scripts/convert_bert_from_uer_to_huggingface.py --input_model_path models/cluecorpussmall_roberta_medium_seq512_model.bin-250000 \ --output_model_path pytorch_model.bin \ --layers_num 8 --type mlm ``` ### BibTeX entry and citation info ``` @article{devlin2018bert, title={Bert: Pre-training of deep bidirectional transformers for language understanding}, author={Devlin, Jacob and Chang, Ming-Wei and Lee, Kenton and Toutanova, Kristina}, journal={arXiv preprint arXiv:1810.04805}, year={2018} } @article{liu2019roberta, title={Roberta: A robustly optimized bert pretraining approach}, author={Liu, Yinhan and Ott, Myle and Goyal, Naman and Du, Jingfei and Joshi, Mandar and Chen, Danqi and Levy, Omer and Lewis, Mike and Zettlemoyer, Luke and Stoyanov, Veselin}, journal={arXiv preprint arXiv:1907.11692}, year={2019} } @article{turc2019, title={Well-Read Students Learn Better: On the Importance of Pre-training Compact Models}, author={Turc, Iulia and Chang, Ming-Wei and Lee, Kenton and Toutanova, Kristina}, journal={arXiv preprint arXiv:1908.08962v2 }, year={2019} } @article{zhao2019uer, title={UER: An Open-Source Toolkit for Pre-training Models}, author={Zhao, Zhe and Chen, Hui and Zhang, Jinbin and Zhao, Xin and Liu, Tao and Lu, Wei and Chen, Xi and Deng, Haotang and Ju, Qi and Du, Xiaoyong}, journal={EMNLP-IJCNLP 2019}, pages={241}, year={2019} } @article{zhao2023tencentpretrain, title={TencentPretrain: A Scalable and Flexible Toolkit for Pre-training Models of Different Modalities}, author={Zhao, Zhe and Li, Yudong and Hou, Cheng and Zhao, Jing and others}, journal={ACL 2023}, pages={217}, year={2023} } ``` [2_128]:https://huggingface.co/uer/chinese_roberta_L-2_H-128 [2_256]:https://huggingface.co/uer/chinese_roberta_L-2_H-256 [2_512]:https://huggingface.co/uer/chinese_roberta_L-2_H-512 [2_768]:https://huggingface.co/uer/chinese_roberta_L-2_H-768 [4_128]:https://huggingface.co/uer/chinese_roberta_L-4_H-128 [4_256]:https://huggingface.co/uer/chinese_roberta_L-4_H-256 [4_512]:https://huggingface.co/uer/chinese_roberta_L-4_H-512 [4_768]:https://huggingface.co/uer/chinese_roberta_L-4_H-768 [6_128]:https://huggingface.co/uer/chinese_roberta_L-6_H-128 [6_256]:https://huggingface.co/uer/chinese_roberta_L-6_H-256 [6_512]:https://huggingface.co/uer/chinese_roberta_L-6_H-512 [6_768]:https://huggingface.co/uer/chinese_roberta_L-6_H-768 [8_128]:https://huggingface.co/uer/chinese_roberta_L-8_H-128 [8_256]:https://huggingface.co/uer/chinese_roberta_L-8_H-256 [8_512]:https://huggingface.co/uer/chinese_roberta_L-8_H-512 [8_768]:https://huggingface.co/uer/chinese_roberta_L-8_H-768 [10_128]:https://huggingface.co/uer/chinese_roberta_L-10_H-128 [10_256]:https://huggingface.co/uer/chinese_roberta_L-10_H-256 [10_512]:https://huggingface.co/uer/chinese_roberta_L-10_H-512 [10_768]:https://huggingface.co/uer/chinese_roberta_L-10_H-768 [12_128]:https://huggingface.co/uer/chinese_roberta_L-12_H-128 [12_256]:https://huggingface.co/uer/chinese_roberta_L-12_H-256 [12_512]:https://huggingface.co/uer/chinese_roberta_L-12_H-512 [12_768]:https://huggingface.co/uer/chinese_roberta_L-12_H-768

uer/chinese_roberta_L-10_H-512

uer

2023-08-30T02:29:34Z

111

0

transformers

[ "transformers", "pytorch", "tf", "jax", "bert", "fill-mask", "zh", "dataset:CLUECorpusSmall", "arxiv:1909.05658", "arxiv:2212.06385", "arxiv:1908.08962", "arxiv:1810.04805", "arxiv:1907.11692", "autotrain_compatible", "endpoints_compatible", "region:us" ]

fill-mask

2022-03-02T23:29:05Z

--- language: zh datasets: CLUECorpusSmall widget: - text: "北京是[MASK]国的首都。" --- # Chinese RoBERTa Miniatures ## Model description This is the set of 24 Chinese RoBERTa models pre-trained by [UER-py](https://github.com/dbiir/UER-py/), which is introduced in [this paper](https://arxiv.org/abs/1909.05658). Besides, the models could also be pre-trained by [TencentPretrain](https://github.com/Tencent/TencentPretrain) introduced in [this paper](https://arxiv.org/abs/2212.06385), which inherits UER-py to support models with parameters above one billion, and extends it to a multimodal pre-training framework. [Turc et al.](https://arxiv.org/abs/1908.08962) have shown that the standard BERT recipe is effective on a wide range of model sizes. Following their paper, we released the 24 Chinese RoBERTa models. In order to facilitate users in reproducing the results, we used a publicly available corpus and provided all training details. You can download the 24 Chinese RoBERTa miniatures either from the [UER-py Modelzoo page](https://github.com/dbiir/UER-py/wiki/Modelzoo), or via HuggingFace from the links below: | | H=128 | H=256 | H=512 | H=768 | | -------- | :-----------------------: | :-----------------------: | :-------------------------: | :-------------------------: | | **L=2** | [**2/128 (Tiny)**][2_128] | [2/256][2_256] | [2/512][2_512] | [2/768][2_768] | | **L=4** | [4/128][4_128] | [**4/256 (Mini)**][4_256] | [**4/512 (Small)**][4_512] | [4/768][4_768] | | **L=6** | [6/128][6_128] | [6/256][6_256] | [6/512][6_512] | [6/768][6_768] | | **L=8** | [8/128][8_128] | [8/256][8_256] | [**8/512 (Medium)**][8_512] | [8/768][8_768] | | **L=10** | [10/128][10_128] | [10/256][10_256] | [10/512][10_512] | [10/768][10_768] | | **L=12** | [12/128][12_128] | [12/256][12_256] | [12/512][12_512] | [**12/768 (Base)**][12_768] | Here are scores on the devlopment set of six Chinese tasks: | Model | Score | book_review | chnsenticorp | lcqmc | tnews(CLUE) | iflytek(CLUE) | ocnli(CLUE) | | -------------- | :---: | :----: | :----------: | :---: | :---------: | :-----------: | :---------: | | RoBERTa-Tiny | 72.3 | 83.4 | 91.4 | 81.8 | 62.0 | 55.0 | 60.3 | | RoBERTa-Mini | 75.9 | 85.7 | 93.7 | 86.1 | 63.9 | 58.3 | 67.4 | | RoBERTa-Small | 76.9 | 87.5 | 93.4 | 86.5 | 65.1 | 59.4 | 69.7 | | RoBERTa-Medium | 78.0 | 88.7 | 94.8 | 88.1 | 65.6 | 59.5 | 71.2 | | RoBERTa-Base | 79.7 | 90.1 | 95.2 | 89.2 | 67.0 | 60.9 | 75.5 | For each task, we selected the best fine-tuning hyperparameters from the lists below, and trained with the sequence length of 128: - epochs: 3, 5, 8 - batch sizes: 32, 64 - learning rates: 3e-5, 1e-4, 3e-4 ## How to use You can use this model directly with a pipeline for masked language modeling (take the case of RoBERTa-Medium): ```python >>> from transformers import pipeline >>> unmasker = pipeline('fill-mask', model='uer/chinese_roberta_L-8_H-512') >>> unmasker("中国的首都是[MASK]京。") [ {'sequence': '[CLS] 中国的首都是北京。 [SEP]', 'score': 0.8701988458633423, 'token': 1266, 'token_str': '北'}, {'sequence': '[CLS] 中国的首都是南京。 [SEP]', 'score': 0.1194809079170227, 'token': 1298, 'token_str': '南'}, {'sequence': '[CLS] 中国的首都是东京。 [SEP]', 'score': 0.0037803512532263994, 'token': 691, 'token_str': '东'}, {'sequence': '[CLS] 中国的首都是普京。 [SEP]', 'score': 0.0017127094324678183, 'token': 3249, 'token_str': '普'}, {'sequence': '[CLS] 中国的首都是望京。 [SEP]', 'score': 0.001687526935711503, 'token': 3307, 'token_str': '望'} ] ``` Here is how to use this model to get the features of a given text in PyTorch: ```python from transformers import BertTokenizer, BertModel tokenizer = BertTokenizer.from_pretrained('uer/chinese_roberta_L-8_H-512') model = BertModel.from_pretrained("uer/chinese_roberta_L-8_H-512") text = "用你喜欢的任何文本替换我。" encoded_input = tokenizer(text, return_tensors='pt') output = model(**encoded_input) ``` and in TensorFlow: ```python from transformers import BertTokenizer, TFBertModel tokenizer = BertTokenizer.from_pretrained('uer/chinese_roberta_L-8_H-512') model = TFBertModel.from_pretrained("uer/chinese_roberta_L-8_H-512") text = "用你喜欢的任何文本替换我。" encoded_input = tokenizer(text, return_tensors='tf') output = model(encoded_input) ``` ## Training data [CLUECorpusSmall](https://github.com/CLUEbenchmark/CLUECorpus2020/) is used as training data. We found that models pre-trained on CLUECorpusSmall outperform those pre-trained on CLUECorpus2020, although CLUECorpus2020 is much larger than CLUECorpusSmall. ## Training procedure Models are pre-trained by [UER-py](https://github.com/dbiir/UER-py/) on [Tencent Cloud](https://cloud.tencent.com/). We pre-train 1,000,000 steps with a sequence length of 128 and then pre-train 250,000 additional steps with a sequence length of 512. We use the same hyper-parameters on different model sizes. Taking the case of RoBERTa-Medium Stage1: ``` python3 preprocess.py --corpus_path corpora/cluecorpussmall.txt \ --vocab_path models/google_zh_vocab.txt \ --dataset_path cluecorpussmall_seq128_dataset.pt \ --processes_num 32 --seq_length 128 \ --dynamic_masking --data_processor mlm ``` ``` python3 pretrain.py --dataset_path cluecorpussmall_seq128_dataset.pt \ --vocab_path models/google_zh_vocab.txt \ --config_path models/bert/medium_config.json \ --output_model_path models/cluecorpussmall_roberta_medium_seq128_model.bin \ --world_size 8 --gpu_ranks 0 1 2 3 4 5 6 7 \ --total_steps 1000000 --save_checkpoint_steps 100000 --report_steps 50000 \ --learning_rate 1e-4 --batch_size 64 \ --data_processor mlm --target mlm ``` Stage2: ``` python3 preprocess.py --corpus_path corpora/cluecorpussmall.txt \ --vocab_path models/google_zh_vocab.txt \ --dataset_path cluecorpussmall_seq512_dataset.pt \ --processes_num 32 --seq_length 512 \ --dynamic_masking --data_processor mlm ``` ``` python3 pretrain.py --dataset_path cluecorpussmall_seq512_dataset.pt \ --vocab_path models/google_zh_vocab.txt \ --pretrained_model_path models/cluecorpussmall_roberta_medium_seq128_model.bin-1000000 \ --config_path models/bert/medium_config.json \ --output_model_path models/cluecorpussmall_roberta_medium_seq512_model.bin \ --world_size 8 --gpu_ranks 0 1 2 3 4 5 6 7 \ --total_steps 250000 --save_checkpoint_steps 50000 --report_steps 10000 \ --learning_rate 5e-5 --batch_size 16 \ --data_processor mlm --target mlm ``` Finally, we convert the pre-trained model into Huggingface's format: ``` python3 scripts/convert_bert_from_uer_to_huggingface.py --input_model_path models/cluecorpussmall_roberta_medium_seq512_model.bin-250000 \ --output_model_path pytorch_model.bin \ --layers_num 8 --type mlm ``` ### BibTeX entry and citation info ``` @article{devlin2018bert, title={Bert: Pre-training of deep bidirectional transformers for language understanding}, author={Devlin, Jacob and Chang, Ming-Wei and Lee, Kenton and Toutanova, Kristina}, journal={arXiv preprint arXiv:1810.04805}, year={2018} } @article{liu2019roberta, title={Roberta: A robustly optimized bert pretraining approach}, author={Liu, Yinhan and Ott, Myle and Goyal, Naman and Du, Jingfei and Joshi, Mandar and Chen, Danqi and Levy, Omer and Lewis, Mike and Zettlemoyer, Luke and Stoyanov, Veselin}, journal={arXiv preprint arXiv:1907.11692}, year={2019} } @article{turc2019, title={Well-Read Students Learn Better: On the Importance of Pre-training Compact Models}, author={Turc, Iulia and Chang, Ming-Wei and Lee, Kenton and Toutanova, Kristina}, journal={arXiv preprint arXiv:1908.08962v2 }, year={2019} } @article{zhao2019uer, title={UER: An Open-Source Toolkit for Pre-training Models}, author={Zhao, Zhe and Chen, Hui and Zhang, Jinbin and Zhao, Xin and Liu, Tao and Lu, Wei and Chen, Xi and Deng, Haotang and Ju, Qi and Du, Xiaoyong}, journal={EMNLP-IJCNLP 2019}, pages={241}, year={2019} } @article{zhao2023tencentpretrain, title={TencentPretrain: A Scalable and Flexible Toolkit for Pre-training Models of Different Modalities}, author={Zhao, Zhe and Li, Yudong and Hou, Cheng and Zhao, Jing and others}, journal={ACL 2023}, pages={217}, year={2023} } ``` [2_128]:https://huggingface.co/uer/chinese_roberta_L-2_H-128 [2_256]:https://huggingface.co/uer/chinese_roberta_L-2_H-256 [2_512]:https://huggingface.co/uer/chinese_roberta_L-2_H-512 [2_768]:https://huggingface.co/uer/chinese_roberta_L-2_H-768 [4_128]:https://huggingface.co/uer/chinese_roberta_L-4_H-128 [4_256]:https://huggingface.co/uer/chinese_roberta_L-4_H-256 [4_512]:https://huggingface.co/uer/chinese_roberta_L-4_H-512 [4_768]:https://huggingface.co/uer/chinese_roberta_L-4_H-768 [6_128]:https://huggingface.co/uer/chinese_roberta_L-6_H-128 [6_256]:https://huggingface.co/uer/chinese_roberta_L-6_H-256 [6_512]:https://huggingface.co/uer/chinese_roberta_L-6_H-512 [6_768]:https://huggingface.co/uer/chinese_roberta_L-6_H-768 [8_128]:https://huggingface.co/uer/chinese_roberta_L-8_H-128 [8_256]:https://huggingface.co/uer/chinese_roberta_L-8_H-256 [8_512]:https://huggingface.co/uer/chinese_roberta_L-8_H-512 [8_768]:https://huggingface.co/uer/chinese_roberta_L-8_H-768 [10_128]:https://huggingface.co/uer/chinese_roberta_L-10_H-128 [10_256]:https://huggingface.co/uer/chinese_roberta_L-10_H-256 [10_512]:https://huggingface.co/uer/chinese_roberta_L-10_H-512 [10_768]:https://huggingface.co/uer/chinese_roberta_L-10_H-768 [12_128]:https://huggingface.co/uer/chinese_roberta_L-12_H-128 [12_256]:https://huggingface.co/uer/chinese_roberta_L-12_H-256 [12_512]:https://huggingface.co/uer/chinese_roberta_L-12_H-512 [12_768]:https://huggingface.co/uer/chinese_roberta_L-12_H-768

yumjunstar/trocr-small-printedkorean-deleteunusedchar_noise

yumjunstar

2023-08-30T02:28:54Z

51

0

transformers

[ "transformers", "pytorch", "vision-encoder-decoder", "image-text-to-text", "generated_from_trainer", "endpoints_compatible", "region:us" ]

image-text-to-text

2023-08-29T17:52:22Z

--- tags: - generated_from_trainer metrics: - wer - accuracy model-index: - name: trocr-small-printedkorean-deleteunusedchar_noise results: [] ---  # trocr-small-printedkorean-deleteunusedchar_noise This model was trained from scratch on an unknown dataset. It achieves the following results on the evaluation set: - Loss: 1.3375 - Cer: 0.2783 - Wer: 0.2975 - Accuracy: 45.6667 ## Model description More information needed ## Intended uses & limitations More information needed ## Training and evaluation data More information needed ## Training procedure ### Training hyperparameters The following hyperparameters were used during training: - learning_rate: 4e-05 - train_batch_size: 128 - eval_batch_size: 192 - seed: 42 - optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08 - lr_scheduler_type: linear - num_epochs: 10 - mixed_precision_training: Native AMP ### Training results | Training Loss | Epoch | Step | Validation Loss | Cer | Wer | Accuracy | |:-------------:|:-----:|:-----:|:---------------:|:------:|:------:|:--------:| | 1.711 | 0.43 | 1000 | 1.6485 | 0.3288 | 0.3944 | 30.6667 | | 1.6849 | 0.85 | 2000 | 1.5361 | 0.3098 | 0.3809 | 32.3333 | | 1.4933 | 1.28 | 3000 | 1.4302 | 0.2935 | 0.3533 | 34.6667 | | 1.526 | 1.71 | 4000 | 1.4010 | 0.2922 | 0.3400 | 35.6667 | | 1.3422 | 2.13 | 5000 | 1.3883 | 0.2846 | 0.3331 | 36.0 | | 1.333 | 2.56 | 6000 | 1.3790 | 0.2871 | 0.3308 | 34.0 | | 1.3295 | 2.99 | 7000 | 1.3644 | 0.2876 | 0.3294 | 35.6667 | | 1.3294 | 3.42 | 8000 | 1.3588 | 0.2824 | 0.3202 | 36.6667 | | 1.3578 | 3.84 | 9000 | 1.3502 | 0.2823 | 0.3162 | 40.6667 | | 1.3029 | 4.27 | 10000 | 1.3514 | 0.2879 | 0.3228 | 37.0 | | 1.2777 | 4.7 | 11000 | 1.3507 | 0.2813 | 0.3168 | 38.3333 | | 1.1781 | 5.12 | 12000 | 1.3507 | 0.2791 | 0.3150 | 40.3333 | | 1.3025 | 5.55 | 13000 | 1.3459 | 0.2818 | 0.3099 | 41.6667 | | 1.2024 | 5.98 | 14000 | 1.3401 | 0.2801 | 0.3061 | 41.6667 | | 1.1792 | 6.4 | 15000 | 1.3412 | 0.2763 | 0.3015 | 44.6667 | | 1.1586 | 6.83 | 16000 | 1.3410 | 0.2799 | 0.3064 | 43.3333 | | 1.2098 | 7.26 | 17000 | 1.3439 | 0.2777 | 0.3030 | 43.6667 | | 1.2122 | 7.69 | 18000 | 1.3418 | 0.2816 | 0.3050 | 43.3333 | | 1.1323 | 8.11 | 19000 | 1.3409 | 0.2767 | 0.2981 | 45.3333 | | 1.2215 | 8.54 | 20000 | 1.3386 | 0.2781 | 0.3004 | 44.0 | | 1.2068 | 8.97 | 21000 | 1.3375 | 0.2762 | 0.2972 | 45.0 | | 1.0847 | 9.39 | 22000 | 1.3366 | 0.2765 | 0.2969 | 46.0 | | 1.1791 | 9.82 | 23000 | 1.3375 | 0.2783 | 0.2975 | 45.6667 | ### Framework versions - Transformers 4.28.0 - Pytorch 1.13.1+cu116 - Datasets 2.14.4 - Tokenizers 0.13.3

uer/chinese_roberta_L-8_H-768

uer

2023-08-30T02:28:34Z

110

0

transformers

[ "transformers", "pytorch", "tf", "jax", "bert", "fill-mask", "zh", "dataset:CLUECorpusSmall", "arxiv:1909.05658", "arxiv:2212.06385", "arxiv:1908.08962", "arxiv:1810.04805", "arxiv:1907.11692", "autotrain_compatible", "endpoints_compatible", "region:us" ]

fill-mask

2022-03-02T23:29:05Z

--- language: zh datasets: CLUECorpusSmall widget: - text: "北京是[MASK]国的首都。" --- # Chinese RoBERTa Miniatures ## Model description This is the set of 24 Chinese RoBERTa models pre-trained by [UER-py](https://github.com/dbiir/UER-py/), which is introduced in [this paper](https://arxiv.org/abs/1909.05658). Besides, the models could also be pre-trained by [TencentPretrain](https://github.com/Tencent/TencentPretrain) introduced in [this paper](https://arxiv.org/abs/2212.06385), which inherits UER-py to support models with parameters above one billion, and extends it to a multimodal pre-training framework. [Turc et al.](https://arxiv.org/abs/1908.08962) have shown that the standard BERT recipe is effective on a wide range of model sizes. Following their paper, we released the 24 Chinese RoBERTa models. In order to facilitate users in reproducing the results, we used a publicly available corpus and provided all training details. You can download the 24 Chinese RoBERTa miniatures either from the [UER-py Modelzoo page](https://github.com/dbiir/UER-py/wiki/Modelzoo), or via HuggingFace from the links below: | | H=128 | H=256 | H=512 | H=768 | | -------- | :-----------------------: | :-----------------------: | :-------------------------: | :-------------------------: | | **L=2** | [**2/128 (Tiny)**][2_128] | [2/256][2_256] | [2/512][2_512] | [2/768][2_768] | | **L=4** | [4/128][4_128] | [**4/256 (Mini)**][4_256] | [**4/512 (Small)**][4_512] | [4/768][4_768] | | **L=6** | [6/128][6_128] | [6/256][6_256] | [6/512][6_512] | [6/768][6_768] | | **L=8** | [8/128][8_128] | [8/256][8_256] | [**8/512 (Medium)**][8_512] | [8/768][8_768] | | **L=10** | [10/128][10_128] | [10/256][10_256] | [10/512][10_512] | [10/768][10_768] | | **L=12** | [12/128][12_128] | [12/256][12_256] | [12/512][12_512] | [**12/768 (Base)**][12_768] | Here are scores on the devlopment set of six Chinese tasks: | Model | Score | book_review | chnsenticorp | lcqmc | tnews(CLUE) | iflytek(CLUE) | ocnli(CLUE) | | -------------- | :---: | :----: | :----------: | :---: | :---------: | :-----------: | :---------: | | RoBERTa-Tiny | 72.3 | 83.4 | 91.4 | 81.8 | 62.0 | 55.0 | 60.3 | | RoBERTa-Mini | 75.9 | 85.7 | 93.7 | 86.1 | 63.9 | 58.3 | 67.4 | | RoBERTa-Small | 76.9 | 87.5 | 93.4 | 86.5 | 65.1 | 59.4 | 69.7 | | RoBERTa-Medium | 78.0 | 88.7 | 94.8 | 88.1 | 65.6 | 59.5 | 71.2 | | RoBERTa-Base | 79.7 | 90.1 | 95.2 | 89.2 | 67.0 | 60.9 | 75.5 | For each task, we selected the best fine-tuning hyperparameters from the lists below, and trained with the sequence length of 128: - epochs: 3, 5, 8 - batch sizes: 32, 64 - learning rates: 3e-5, 1e-4, 3e-4 ## How to use You can use this model directly with a pipeline for masked language modeling (take the case of RoBERTa-Medium): ```python >>> from transformers import pipeline >>> unmasker = pipeline('fill-mask', model='uer/chinese_roberta_L-8_H-512') >>> unmasker("中国的首都是[MASK]京。") [ {'sequence': '[CLS] 中国的首都是北京。 [SEP]', 'score': 0.8701988458633423, 'token': 1266, 'token_str': '北'}, {'sequence': '[CLS] 中国的首都是南京。 [SEP]', 'score': 0.1194809079170227, 'token': 1298, 'token_str': '南'}, {'sequence': '[CLS] 中国的首都是东京。 [SEP]', 'score': 0.0037803512532263994, 'token': 691, 'token_str': '东'}, {'sequence': '[CLS] 中国的首都是普京。 [SEP]', 'score': 0.0017127094324678183, 'token': 3249, 'token_str': '普'}, {'sequence': '[CLS] 中国的首都是望京。 [SEP]', 'score': 0.001687526935711503, 'token': 3307, 'token_str': '望'} ] ``` Here is how to use this model to get the features of a given text in PyTorch: ```python from transformers import BertTokenizer, BertModel tokenizer = BertTokenizer.from_pretrained('uer/chinese_roberta_L-8_H-512') model = BertModel.from_pretrained("uer/chinese_roberta_L-8_H-512") text = "用你喜欢的任何文本替换我。" encoded_input = tokenizer(text, return_tensors='pt') output = model(**encoded_input) ``` and in TensorFlow: ```python from transformers import BertTokenizer, TFBertModel tokenizer = BertTokenizer.from_pretrained('uer/chinese_roberta_L-8_H-512') model = TFBertModel.from_pretrained("uer/chinese_roberta_L-8_H-512") text = "用你喜欢的任何文本替换我。" encoded_input = tokenizer(text, return_tensors='tf') output = model(encoded_input) ``` ## Training data [CLUECorpusSmall](https://github.com/CLUEbenchmark/CLUECorpus2020/) is used as training data. We found that models pre-trained on CLUECorpusSmall outperform those pre-trained on CLUECorpus2020, although CLUECorpus2020 is much larger than CLUECorpusSmall. ## Training procedure Models are pre-trained by [UER-py](https://github.com/dbiir/UER-py/) on [Tencent Cloud](https://cloud.tencent.com/). We pre-train 1,000,000 steps with a sequence length of 128 and then pre-train 250,000 additional steps with a sequence length of 512. We use the same hyper-parameters on different model sizes. Taking the case of RoBERTa-Medium Stage1: ``` python3 preprocess.py --corpus_path corpora/cluecorpussmall.txt \ --vocab_path models/google_zh_vocab.txt \ --dataset_path cluecorpussmall_seq128_dataset.pt \ --processes_num 32 --seq_length 128 \ --dynamic_masking --data_processor mlm ``` ``` python3 pretrain.py --dataset_path cluecorpussmall_seq128_dataset.pt \ --vocab_path models/google_zh_vocab.txt \ --config_path models/bert/medium_config.json \ --output_model_path models/cluecorpussmall_roberta_medium_seq128_model.bin \ --world_size 8 --gpu_ranks 0 1 2 3 4 5 6 7 \ --total_steps 1000000 --save_checkpoint_steps 100000 --report_steps 50000 \ --learning_rate 1e-4 --batch_size 64 \ --data_processor mlm --target mlm ``` Stage2: ``` python3 preprocess.py --corpus_path corpora/cluecorpussmall.txt \ --vocab_path models/google_zh_vocab.txt \ --dataset_path cluecorpussmall_seq512_dataset.pt \ --processes_num 32 --seq_length 512 \ --dynamic_masking --data_processor mlm ``` ``` python3 pretrain.py --dataset_path cluecorpussmall_seq512_dataset.pt \ --vocab_path models/google_zh_vocab.txt \ --pretrained_model_path models/cluecorpussmall_roberta_medium_seq128_model.bin-1000000 \ --config_path models/bert/medium_config.json \ --output_model_path models/cluecorpussmall_roberta_medium_seq512_model.bin \ --world_size 8 --gpu_ranks 0 1 2 3 4 5 6 7 \ --total_steps 250000 --save_checkpoint_steps 50000 --report_steps 10000 \ --learning_rate 5e-5 --batch_size 16 \ --data_processor mlm --target mlm ``` Finally, we convert the pre-trained model into Huggingface's format: ``` python3 scripts/convert_bert_from_uer_to_huggingface.py --input_model_path models/cluecorpussmall_roberta_medium_seq512_model.bin-250000 \ --output_model_path pytorch_model.bin \ --layers_num 8 --type mlm ``` ### BibTeX entry and citation info ``` @article{devlin2018bert, title={Bert: Pre-training of deep bidirectional transformers for language understanding}, author={Devlin, Jacob and Chang, Ming-Wei and Lee, Kenton and Toutanova, Kristina}, journal={arXiv preprint arXiv:1810.04805}, year={2018} } @article{liu2019roberta, title={Roberta: A robustly optimized bert pretraining approach}, author={Liu, Yinhan and Ott, Myle and Goyal, Naman and Du, Jingfei and Joshi, Mandar and Chen, Danqi and Levy, Omer and Lewis, Mike and Zettlemoyer, Luke and Stoyanov, Veselin}, journal={arXiv preprint arXiv:1907.11692}, year={2019} } @article{turc2019, title={Well-Read Students Learn Better: On the Importance of Pre-training Compact Models}, author={Turc, Iulia and Chang, Ming-Wei and Lee, Kenton and Toutanova, Kristina}, journal={arXiv preprint arXiv:1908.08962v2 }, year={2019} } @article{zhao2019uer, title={UER: An Open-Source Toolkit for Pre-training Models}, author={Zhao, Zhe and Chen, Hui and Zhang, Jinbin and Zhao, Xin and Liu, Tao and Lu, Wei and Chen, Xi and Deng, Haotang and Ju, Qi and Du, Xiaoyong}, journal={EMNLP-IJCNLP 2019}, pages={241}, year={2019} } @article{zhao2023tencentpretrain, title={TencentPretrain: A Scalable and Flexible Toolkit for Pre-training Models of Different Modalities}, author={Zhao, Zhe and Li, Yudong and Hou, Cheng and Zhao, Jing and others}, journal={ACL 2023}, pages={217}, year={2023} } ``` [2_128]:https://huggingface.co/uer/chinese_roberta_L-2_H-128 [2_256]:https://huggingface.co/uer/chinese_roberta_L-2_H-256 [2_512]:https://huggingface.co/uer/chinese_roberta_L-2_H-512 [2_768]:https://huggingface.co/uer/chinese_roberta_L-2_H-768 [4_128]:https://huggingface.co/uer/chinese_roberta_L-4_H-128 [4_256]:https://huggingface.co/uer/chinese_roberta_L-4_H-256 [4_512]:https://huggingface.co/uer/chinese_roberta_L-4_H-512 [4_768]:https://huggingface.co/uer/chinese_roberta_L-4_H-768 [6_128]:https://huggingface.co/uer/chinese_roberta_L-6_H-128 [6_256]:https://huggingface.co/uer/chinese_roberta_L-6_H-256 [6_512]:https://huggingface.co/uer/chinese_roberta_L-6_H-512 [6_768]:https://huggingface.co/uer/chinese_roberta_L-6_H-768 [8_128]:https://huggingface.co/uer/chinese_roberta_L-8_H-128 [8_256]:https://huggingface.co/uer/chinese_roberta_L-8_H-256 [8_512]:https://huggingface.co/uer/chinese_roberta_L-8_H-512 [8_768]:https://huggingface.co/uer/chinese_roberta_L-8_H-768 [10_128]:https://huggingface.co/uer/chinese_roberta_L-10_H-128 [10_256]:https://huggingface.co/uer/chinese_roberta_L-10_H-256 [10_512]:https://huggingface.co/uer/chinese_roberta_L-10_H-512 [10_768]:https://huggingface.co/uer/chinese_roberta_L-10_H-768 [12_128]:https://huggingface.co/uer/chinese_roberta_L-12_H-128 [12_256]:https://huggingface.co/uer/chinese_roberta_L-12_H-256 [12_512]:https://huggingface.co/uer/chinese_roberta_L-12_H-512 [12_768]:https://huggingface.co/uer/chinese_roberta_L-12_H-768

uer/chinese_roberta_L-8_H-512

uer

2023-08-30T02:28:15Z

146

3

transformers

[ "transformers", "pytorch", "tf", "jax", "bert", "fill-mask", "zh", "dataset:CLUECorpusSmall", "arxiv:1909.05658", "arxiv:2212.06385", "arxiv:1908.08962", "arxiv:1810.04805", "arxiv:1907.11692", "autotrain_compatible", "endpoints_compatible", "region:us" ]

fill-mask

2022-03-02T23:29:05Z

--- language: zh datasets: CLUECorpusSmall widget: - text: "北京是[MASK]国的首都。" --- # Chinese RoBERTa Miniatures ## Model description This is the set of 24 Chinese RoBERTa models pre-trained by [UER-py](https://github.com/dbiir/UER-py/), which is introduced in [this paper](https://arxiv.org/abs/1909.05658). Besides, the models could also be pre-trained by [TencentPretrain](https://github.com/Tencent/TencentPretrain) introduced in [this paper](https://arxiv.org/abs/2212.06385), which inherits UER-py to support models with parameters above one billion, and extends it to a multimodal pre-training framework. [Turc et al.](https://arxiv.org/abs/1908.08962) have shown that the standard BERT recipe is effective on a wide range of model sizes. Following their paper, we released the 24 Chinese RoBERTa models. In order to facilitate users in reproducing the results, we used a publicly available corpus and provided all training details. You can download the 24 Chinese RoBERTa miniatures either from the [UER-py Modelzoo page](https://github.com/dbiir/UER-py/wiki/Modelzoo), or via HuggingFace from the links below: | | H=128 | H=256 | H=512 | H=768 | | -------- | :-----------------------: | :-----------------------: | :-------------------------: | :-------------------------: | | **L=2** | [**2/128 (Tiny)**][2_128] | [2/256][2_256] | [2/512][2_512] | [2/768][2_768] | | **L=4** | [4/128][4_128] | [**4/256 (Mini)**][4_256] | [**4/512 (Small)**][4_512] | [4/768][4_768] | | **L=6** | [6/128][6_128] | [6/256][6_256] | [6/512][6_512] | [6/768][6_768] | | **L=8** | [8/128][8_128] | [8/256][8_256] | [**8/512 (Medium)**][8_512] | [8/768][8_768] | | **L=10** | [10/128][10_128] | [10/256][10_256] | [10/512][10_512] | [10/768][10_768] | | **L=12** | [12/128][12_128] | [12/256][12_256] | [12/512][12_512] | [**12/768 (Base)**][12_768] | Here are scores on the devlopment set of six Chinese tasks: | Model | Score | book_review | chnsenticorp | lcqmc | tnews(CLUE) | iflytek(CLUE) | ocnli(CLUE) | | -------------- | :---: | :----: | :----------: | :---: | :---------: | :-----------: | :---------: | | RoBERTa-Tiny | 72.3 | 83.4 | 91.4 | 81.8 | 62.0 | 55.0 | 60.3 | | RoBERTa-Mini | 75.9 | 85.7 | 93.7 | 86.1 | 63.9 | 58.3 | 67.4 | | RoBERTa-Small | 76.9 | 87.5 | 93.4 | 86.5 | 65.1 | 59.4 | 69.7 | | RoBERTa-Medium | 78.0 | 88.7 | 94.8 | 88.1 | 65.6 | 59.5 | 71.2 | | RoBERTa-Base | 79.7 | 90.1 | 95.2 | 89.2 | 67.0 | 60.9 | 75.5 | For each task, we selected the best fine-tuning hyperparameters from the lists below, and trained with the sequence length of 128: - epochs: 3, 5, 8 - batch sizes: 32, 64 - learning rates: 3e-5, 1e-4, 3e-4 ## How to use You can use this model directly with a pipeline for masked language modeling (take the case of RoBERTa-Medium): ```python >>> from transformers import pipeline >>> unmasker = pipeline('fill-mask', model='uer/chinese_roberta_L-8_H-512') >>> unmasker("中国的首都是[MASK]京。") [ {'sequence': '[CLS] 中国的首都是北京。 [SEP]', 'score': 0.8701988458633423, 'token': 1266, 'token_str': '北'}, {'sequence': '[CLS] 中国的首都是南京。 [SEP]', 'score': 0.1194809079170227, 'token': 1298, 'token_str': '南'}, {'sequence': '[CLS] 中国的首都是东京。 [SEP]', 'score': 0.0037803512532263994, 'token': 691, 'token_str': '东'}, {'sequence': '[CLS] 中国的首都是普京。 [SEP]', 'score': 0.0017127094324678183, 'token': 3249, 'token_str': '普'}, {'sequence': '[CLS] 中国的首都是望京。 [SEP]', 'score': 0.001687526935711503, 'token': 3307, 'token_str': '望'} ] ``` Here is how to use this model to get the features of a given text in PyTorch: ```python from transformers import BertTokenizer, BertModel tokenizer = BertTokenizer.from_pretrained('uer/chinese_roberta_L-8_H-512') model = BertModel.from_pretrained("uer/chinese_roberta_L-8_H-512") text = "用你喜欢的任何文本替换我。" encoded_input = tokenizer(text, return_tensors='pt') output = model(**encoded_input) ``` and in TensorFlow: ```python from transformers import BertTokenizer, TFBertModel tokenizer = BertTokenizer.from_pretrained('uer/chinese_roberta_L-8_H-512') model = TFBertModel.from_pretrained("uer/chinese_roberta_L-8_H-512") text = "用你喜欢的任何文本替换我。" encoded_input = tokenizer(text, return_tensors='tf') output = model(encoded_input) ``` ## Training data [CLUECorpusSmall](https://github.com/CLUEbenchmark/CLUECorpus2020/) is used as training data. We found that models pre-trained on CLUECorpusSmall outperform those pre-trained on CLUECorpus2020, although CLUECorpus2020 is much larger than CLUECorpusSmall. ## Training procedure Models are pre-trained by [UER-py](https://github.com/dbiir/UER-py/) on [Tencent Cloud](https://cloud.tencent.com/). We pre-train 1,000,000 steps with a sequence length of 128 and then pre-train 250,000 additional steps with a sequence length of 512. We use the same hyper-parameters on different model sizes. Taking the case of RoBERTa-Medium Stage1: ``` python3 preprocess.py --corpus_path corpora/cluecorpussmall.txt \ --vocab_path models/google_zh_vocab.txt \ --dataset_path cluecorpussmall_seq128_dataset.pt \ --processes_num 32 --seq_length 128 \ --dynamic_masking --data_processor mlm ``` ``` python3 pretrain.py --dataset_path cluecorpussmall_seq128_dataset.pt \ --vocab_path models/google_zh_vocab.txt \ --config_path models/bert/medium_config.json \ --output_model_path models/cluecorpussmall_roberta_medium_seq128_model.bin \ --world_size 8 --gpu_ranks 0 1 2 3 4 5 6 7 \ --total_steps 1000000 --save_checkpoint_steps 100000 --report_steps 50000 \ --learning_rate 1e-4 --batch_size 64 \ --data_processor mlm --target mlm ``` Stage2: ``` python3 preprocess.py --corpus_path corpora/cluecorpussmall.txt \ --vocab_path models/google_zh_vocab.txt \ --dataset_path cluecorpussmall_seq512_dataset.pt \ --processes_num 32 --seq_length 512 \ --dynamic_masking --data_processor mlm ``` ``` python3 pretrain.py --dataset_path cluecorpussmall_seq512_dataset.pt \ --vocab_path models/google_zh_vocab.txt \ --pretrained_model_path models/cluecorpussmall_roberta_medium_seq128_model.bin-1000000 \ --config_path models/bert/medium_config.json \ --output_model_path models/cluecorpussmall_roberta_medium_seq512_model.bin \ --world_size 8 --gpu_ranks 0 1 2 3 4 5 6 7 \ --total_steps 250000 --save_checkpoint_steps 50000 --report_steps 10000 \ --learning_rate 5e-5 --batch_size 16 \ --data_processor mlm --target mlm ``` Finally, we convert the pre-trained model into Huggingface's format: ``` python3 scripts/convert_bert_from_uer_to_huggingface.py --input_model_path models/cluecorpussmall_roberta_medium_seq512_model.bin-250000 \ --output_model_path pytorch_model.bin \ --layers_num 8 --type mlm ``` ### BibTeX entry and citation info ``` @article{devlin2018bert, title={Bert: Pre-training of deep bidirectional transformers for language understanding}, author={Devlin, Jacob and Chang, Ming-Wei and Lee, Kenton and Toutanova, Kristina}, journal={arXiv preprint arXiv:1810.04805}, year={2018} } @article{liu2019roberta, title={Roberta: A robustly optimized bert pretraining approach}, author={Liu, Yinhan and Ott, Myle and Goyal, Naman and Du, Jingfei and Joshi, Mandar and Chen, Danqi and Levy, Omer and Lewis, Mike and Zettlemoyer, Luke and Stoyanov, Veselin}, journal={arXiv preprint arXiv:1907.11692}, year={2019} } @article{turc2019, title={Well-Read Students Learn Better: On the Importance of Pre-training Compact Models}, author={Turc, Iulia and Chang, Ming-Wei and Lee, Kenton and Toutanova, Kristina}, journal={arXiv preprint arXiv:1908.08962v2 }, year={2019} } @article{zhao2019uer, title={UER: An Open-Source Toolkit for Pre-training Models}, author={Zhao, Zhe and Chen, Hui and Zhang, Jinbin and Zhao, Xin and Liu, Tao and Lu, Wei and Chen, Xi and Deng, Haotang and Ju, Qi and Du, Xiaoyong}, journal={EMNLP-IJCNLP 2019}, pages={241}, year={2019} } @article{zhao2023tencentpretrain, title={TencentPretrain: A Scalable and Flexible Toolkit for Pre-training Models of Different Modalities}, author={Zhao, Zhe and Li, Yudong and Hou, Cheng and Zhao, Jing and others}, journal={ACL 2023}, pages={217}, year={2023} } ``` [2_128]:https://huggingface.co/uer/chinese_roberta_L-2_H-128 [2_256]:https://huggingface.co/uer/chinese_roberta_L-2_H-256 [2_512]:https://huggingface.co/uer/chinese_roberta_L-2_H-512 [2_768]:https://huggingface.co/uer/chinese_roberta_L-2_H-768 [4_128]:https://huggingface.co/uer/chinese_roberta_L-4_H-128 [4_256]:https://huggingface.co/uer/chinese_roberta_L-4_H-256 [4_512]:https://huggingface.co/uer/chinese_roberta_L-4_H-512 [4_768]:https://huggingface.co/uer/chinese_roberta_L-4_H-768 [6_128]:https://huggingface.co/uer/chinese_roberta_L-6_H-128 [6_256]:https://huggingface.co/uer/chinese_roberta_L-6_H-256 [6_512]:https://huggingface.co/uer/chinese_roberta_L-6_H-512 [6_768]:https://huggingface.co/uer/chinese_roberta_L-6_H-768 [8_128]:https://huggingface.co/uer/chinese_roberta_L-8_H-128 [8_256]:https://huggingface.co/uer/chinese_roberta_L-8_H-256 [8_512]:https://huggingface.co/uer/chinese_roberta_L-8_H-512 [8_768]:https://huggingface.co/uer/chinese_roberta_L-8_H-768 [10_128]:https://huggingface.co/uer/chinese_roberta_L-10_H-128 [10_256]:https://huggingface.co/uer/chinese_roberta_L-10_H-256 [10_512]:https://huggingface.co/uer/chinese_roberta_L-10_H-512 [10_768]:https://huggingface.co/uer/chinese_roberta_L-10_H-768 [12_128]:https://huggingface.co/uer/chinese_roberta_L-12_H-128 [12_256]:https://huggingface.co/uer/chinese_roberta_L-12_H-256 [12_512]:https://huggingface.co/uer/chinese_roberta_L-12_H-512 [12_768]:https://huggingface.co/uer/chinese_roberta_L-12_H-768

fuguizhen/q-FrozenLake-v1-4x4-noSlippery

fuguizhen

2023-08-30T02:26:51Z

0

null

[ "FrozenLake-v1-4x4-no_slippery", "q-learning", "reinforcement-learning", "custom-implementation", "model-index", "region:us" ]

reinforcement-learning

2023-08-30T02:26:49Z

--- tags: - FrozenLake-v1-4x4-no_slippery - q-learning - reinforcement-learning - custom-implementation model-index: - name: q-FrozenLake-v1-4x4-noSlippery results: - task: type: reinforcement-learning name: reinforcement-learning dataset: name: FrozenLake-v1-4x4-no_slippery type: FrozenLake-v1-4x4-no_slippery metrics: - type: mean_reward value: 1.00 +/- 0.00 name: mean_reward verified: false --- # **Q-Learning** Agent playing1 **FrozenLake-v1** This is a trained model of a **Q-Learning** agent playing **FrozenLake-v1** . ## Usage model = load_from_hub(repo_id="fuguizhen/q-FrozenLake-v1-4x4-noSlippery", filename="q-learning.pkl") # Don't forget to check if you need to add additional attributes (is_slippery=False etc) env = gym.make(model["env_id"])

uer/chinese_roberta_L-6_H-256

uer

2023-08-30T02:26:30Z

108

1

transformers

[ "transformers", "pytorch", "tf", "jax", "bert", "fill-mask", "zh", "dataset:CLUECorpusSmall", "arxiv:1909.05658", "arxiv:2212.06385", "arxiv:1908.08962", "arxiv:1810.04805", "arxiv:1907.11692", "autotrain_compatible", "endpoints_compatible", "region:us" ]

fill-mask

2022-03-02T23:29:05Z

--- language: zh datasets: CLUECorpusSmall widget: - text: "北京是[MASK]国的首都。" --- # Chinese RoBERTa Miniatures ## Model description This is the set of 24 Chinese RoBERTa models pre-trained by [UER-py](https://github.com/dbiir/UER-py/), which is introduced in [this paper](https://arxiv.org/abs/1909.05658). Besides, the models could also be pre-trained by [TencentPretrain](https://github.com/Tencent/TencentPretrain) introduced in [this paper](https://arxiv.org/abs/2212.06385), which inherits UER-py to support models with parameters above one billion, and extends it to a multimodal pre-training framework. [Turc et al.](https://arxiv.org/abs/1908.08962) have shown that the standard BERT recipe is effective on a wide range of model sizes. Following their paper, we released the 24 Chinese RoBERTa models. In order to facilitate users in reproducing the results, we used a publicly available corpus and provided all training details. You can download the 24 Chinese RoBERTa miniatures either from the [UER-py Modelzoo page](https://github.com/dbiir/UER-py/wiki/Modelzoo), or via HuggingFace from the links below: | | H=128 | H=256 | H=512 | H=768 | | -------- | :-----------------------: | :-----------------------: | :-------------------------: | :-------------------------: | | **L=2** | [**2/128 (Tiny)**][2_128] | [2/256][2_256] | [2/512][2_512] | [2/768][2_768] | | **L=4** | [4/128][4_128] | [**4/256 (Mini)**][4_256] | [**4/512 (Small)**][4_512] | [4/768][4_768] | | **L=6** | [6/128][6_128] | [6/256][6_256] | [6/512][6_512] | [6/768][6_768] | | **L=8** | [8/128][8_128] | [8/256][8_256] | [**8/512 (Medium)**][8_512] | [8/768][8_768] | | **L=10** | [10/128][10_128] | [10/256][10_256] | [10/512][10_512] | [10/768][10_768] | | **L=12** | [12/128][12_128] | [12/256][12_256] | [12/512][12_512] | [**12/768 (Base)**][12_768] | Here are scores on the devlopment set of six Chinese tasks: | Model | Score | book_review | chnsenticorp | lcqmc | tnews(CLUE) | iflytek(CLUE) | ocnli(CLUE) | | -------------- | :---: | :----: | :----------: | :---: | :---------: | :-----------: | :---------: | | RoBERTa-Tiny | 72.3 | 83.4 | 91.4 | 81.8 | 62.0 | 55.0 | 60.3 | | RoBERTa-Mini | 75.9 | 85.7 | 93.7 | 86.1 | 63.9 | 58.3 | 67.4 | | RoBERTa-Small | 76.9 | 87.5 | 93.4 | 86.5 | 65.1 | 59.4 | 69.7 | | RoBERTa-Medium | 78.0 | 88.7 | 94.8 | 88.1 | 65.6 | 59.5 | 71.2 | | RoBERTa-Base | 79.7 | 90.1 | 95.2 | 89.2 | 67.0 | 60.9 | 75.5 | For each task, we selected the best fine-tuning hyperparameters from the lists below, and trained with the sequence length of 128: - epochs: 3, 5, 8 - batch sizes: 32, 64 - learning rates: 3e-5, 1e-4, 3e-4 ## How to use You can use this model directly with a pipeline for masked language modeling (take the case of RoBERTa-Medium): ```python >>> from transformers import pipeline >>> unmasker = pipeline('fill-mask', model='uer/chinese_roberta_L-8_H-512') >>> unmasker("中国的首都是[MASK]京。") [ {'sequence': '[CLS] 中国的首都是北京。 [SEP]', 'score': 0.8701988458633423, 'token': 1266, 'token_str': '北'}, {'sequence': '[CLS] 中国的首都是南京。 [SEP]', 'score': 0.1194809079170227, 'token': 1298, 'token_str': '南'}, {'sequence': '[CLS] 中国的首都是东京。 [SEP]', 'score': 0.0037803512532263994, 'token': 691, 'token_str': '东'}, {'sequence': '[CLS] 中国的首都是普京。 [SEP]', 'score': 0.0017127094324678183, 'token': 3249, 'token_str': '普'}, {'sequence': '[CLS] 中国的首都是望京。 [SEP]', 'score': 0.001687526935711503, 'token': 3307, 'token_str': '望'} ] ``` Here is how to use this model to get the features of a given text in PyTorch: ```python from transformers import BertTokenizer, BertModel tokenizer = BertTokenizer.from_pretrained('uer/chinese_roberta_L-8_H-512') model = BertModel.from_pretrained("uer/chinese_roberta_L-8_H-512") text = "用你喜欢的任何文本替换我。" encoded_input = tokenizer(text, return_tensors='pt') output = model(**encoded_input) ``` and in TensorFlow: ```python from transformers import BertTokenizer, TFBertModel tokenizer = BertTokenizer.from_pretrained('uer/chinese_roberta_L-8_H-512') model = TFBertModel.from_pretrained("uer/chinese_roberta_L-8_H-512") text = "用你喜欢的任何文本替换我。" encoded_input = tokenizer(text, return_tensors='tf') output = model(encoded_input) ``` ## Training data [CLUECorpusSmall](https://github.com/CLUEbenchmark/CLUECorpus2020/) is used as training data. We found that models pre-trained on CLUECorpusSmall outperform those pre-trained on CLUECorpus2020, although CLUECorpus2020 is much larger than CLUECorpusSmall. ## Training procedure Models are pre-trained by [UER-py](https://github.com/dbiir/UER-py/) on [Tencent Cloud](https://cloud.tencent.com/). We pre-train 1,000,000 steps with a sequence length of 128 and then pre-train 250,000 additional steps with a sequence length of 512. We use the same hyper-parameters on different model sizes. Taking the case of RoBERTa-Medium Stage1: ``` python3 preprocess.py --corpus_path corpora/cluecorpussmall.txt \ --vocab_path models/google_zh_vocab.txt \ --dataset_path cluecorpussmall_seq128_dataset.pt \ --processes_num 32 --seq_length 128 \ --dynamic_masking --data_processor mlm ``` ``` python3 pretrain.py --dataset_path cluecorpussmall_seq128_dataset.pt \ --vocab_path models/google_zh_vocab.txt \ --config_path models/bert/medium_config.json \ --output_model_path models/cluecorpussmall_roberta_medium_seq128_model.bin \ --world_size 8 --gpu_ranks 0 1 2 3 4 5 6 7 \ --total_steps 1000000 --save_checkpoint_steps 100000 --report_steps 50000 \ --learning_rate 1e-4 --batch_size 64 \ --data_processor mlm --target mlm ``` Stage2: ``` python3 preprocess.py --corpus_path corpora/cluecorpussmall.txt \ --vocab_path models/google_zh_vocab.txt \ --dataset_path cluecorpussmall_seq512_dataset.pt \ --processes_num 32 --seq_length 512 \ --dynamic_masking --data_processor mlm ``` ``` python3 pretrain.py --dataset_path cluecorpussmall_seq512_dataset.pt \ --vocab_path models/google_zh_vocab.txt \ --pretrained_model_path models/cluecorpussmall_roberta_medium_seq128_model.bin-1000000 \ --config_path models/bert/medium_config.json \ --output_model_path models/cluecorpussmall_roberta_medium_seq512_model.bin \ --world_size 8 --gpu_ranks 0 1 2 3 4 5 6 7 \ --total_steps 250000 --save_checkpoint_steps 50000 --report_steps 10000 \ --learning_rate 5e-5 --batch_size 16 \ --data_processor mlm --target mlm ``` Finally, we convert the pre-trained model into Huggingface's format: ``` python3 scripts/convert_bert_from_uer_to_huggingface.py --input_model_path models/cluecorpussmall_roberta_medium_seq512_model.bin-250000 \ --output_model_path pytorch_model.bin \ --layers_num 8 --type mlm ``` ### BibTeX entry and citation info ``` @article{devlin2018bert, title={Bert: Pre-training of deep bidirectional transformers for language understanding}, author={Devlin, Jacob and Chang, Ming-Wei and Lee, Kenton and Toutanova, Kristina}, journal={arXiv preprint arXiv:1810.04805}, year={2018} } @article{liu2019roberta, title={Roberta: A robustly optimized bert pretraining approach}, author={Liu, Yinhan and Ott, Myle and Goyal, Naman and Du, Jingfei and Joshi, Mandar and Chen, Danqi and Levy, Omer and Lewis, Mike and Zettlemoyer, Luke and Stoyanov, Veselin}, journal={arXiv preprint arXiv:1907.11692}, year={2019} } @article{turc2019, title={Well-Read Students Learn Better: On the Importance of Pre-training Compact Models}, author={Turc, Iulia and Chang, Ming-Wei and Lee, Kenton and Toutanova, Kristina}, journal={arXiv preprint arXiv:1908.08962v2 }, year={2019} } @article{zhao2019uer, title={UER: An Open-Source Toolkit for Pre-training Models}, author={Zhao, Zhe and Chen, Hui and Zhang, Jinbin and Zhao, Xin and Liu, Tao and Lu, Wei and Chen, Xi and Deng, Haotang and Ju, Qi and Du, Xiaoyong}, journal={EMNLP-IJCNLP 2019}, pages={241}, year={2019} } @article{zhao2023tencentpretrain, title={TencentPretrain: A Scalable and Flexible Toolkit for Pre-training Models of Different Modalities}, author={Zhao, Zhe and Li, Yudong and Hou, Cheng and Zhao, Jing and others}, journal={ACL 2023}, pages={217}, year={2023} } ``` [2_128]:https://huggingface.co/uer/chinese_roberta_L-2_H-128 [2_256]:https://huggingface.co/uer/chinese_roberta_L-2_H-256 [2_512]:https://huggingface.co/uer/chinese_roberta_L-2_H-512 [2_768]:https://huggingface.co/uer/chinese_roberta_L-2_H-768 [4_128]:https://huggingface.co/uer/chinese_roberta_L-4_H-128 [4_256]:https://huggingface.co/uer/chinese_roberta_L-4_H-256 [4_512]:https://huggingface.co/uer/chinese_roberta_L-4_H-512 [4_768]:https://huggingface.co/uer/chinese_roberta_L-4_H-768 [6_128]:https://huggingface.co/uer/chinese_roberta_L-6_H-128 [6_256]:https://huggingface.co/uer/chinese_roberta_L-6_H-256 [6_512]:https://huggingface.co/uer/chinese_roberta_L-6_H-512 [6_768]:https://huggingface.co/uer/chinese_roberta_L-6_H-768 [8_128]:https://huggingface.co/uer/chinese_roberta_L-8_H-128 [8_256]:https://huggingface.co/uer/chinese_roberta_L-8_H-256 [8_512]:https://huggingface.co/uer/chinese_roberta_L-8_H-512 [8_768]:https://huggingface.co/uer/chinese_roberta_L-8_H-768 [10_128]:https://huggingface.co/uer/chinese_roberta_L-10_H-128 [10_256]:https://huggingface.co/uer/chinese_roberta_L-10_H-256 [10_512]:https://huggingface.co/uer/chinese_roberta_L-10_H-512 [10_768]:https://huggingface.co/uer/chinese_roberta_L-10_H-768 [12_128]:https://huggingface.co/uer/chinese_roberta_L-12_H-128 [12_256]:https://huggingface.co/uer/chinese_roberta_L-12_H-256 [12_512]:https://huggingface.co/uer/chinese_roberta_L-12_H-512 [12_768]:https://huggingface.co/uer/chinese_roberta_L-12_H-768

uer/chinese_roberta_L-6_H-128

uer

2023-08-30T02:26:10Z

119

0

transformers

[ "transformers", "pytorch", "tf", "jax", "bert", "fill-mask", "zh", "dataset:CLUECorpusSmall", "arxiv:1909.05658", "arxiv:2212.06385", "arxiv:1908.08962", "arxiv:1810.04805", "arxiv:1907.11692", "autotrain_compatible", "endpoints_compatible", "region:us" ]

fill-mask

2022-03-02T23:29:05Z

--- language: zh datasets: CLUECorpusSmall widget: - text: "北京是[MASK]国的首都。" --- # Chinese RoBERTa Miniatures ## Model description This is the set of 24 Chinese RoBERTa models pre-trained by [UER-py](https://github.com/dbiir/UER-py/), which is introduced in [this paper](https://arxiv.org/abs/1909.05658). Besides, the models could also be pre-trained by [TencentPretrain](https://github.com/Tencent/TencentPretrain) introduced in [this paper](https://arxiv.org/abs/2212.06385), which inherits UER-py to support models with parameters above one billion, and extends it to a multimodal pre-training framework. [Turc et al.](https://arxiv.org/abs/1908.08962) have shown that the standard BERT recipe is effective on a wide range of model sizes. Following their paper, we released the 24 Chinese RoBERTa models. In order to facilitate users in reproducing the results, we used a publicly available corpus and provided all training details. You can download the 24 Chinese RoBERTa miniatures either from the [UER-py Modelzoo page](https://github.com/dbiir/UER-py/wiki/Modelzoo), or via HuggingFace from the links below: | | H=128 | H=256 | H=512 | H=768 | | -------- | :-----------------------: | :-----------------------: | :-------------------------: | :-------------------------: | | **L=2** | [**2/128 (Tiny)**][2_128] | [2/256][2_256] | [2/512][2_512] | [2/768][2_768] | | **L=4** | [4/128][4_128] | [**4/256 (Mini)**][4_256] | [**4/512 (Small)**][4_512] | [4/768][4_768] | | **L=6** | [6/128][6_128] | [6/256][6_256] | [6/512][6_512] | [6/768][6_768] | | **L=8** | [8/128][8_128] | [8/256][8_256] | [**8/512 (Medium)**][8_512] | [8/768][8_768] | | **L=10** | [10/128][10_128] | [10/256][10_256] | [10/512][10_512] | [10/768][10_768] | | **L=12** | [12/128][12_128] | [12/256][12_256] | [12/512][12_512] | [**12/768 (Base)**][12_768] | Here are scores on the devlopment set of six Chinese tasks: | Model | Score | book_review | chnsenticorp | lcqmc | tnews(CLUE) | iflytek(CLUE) | ocnli(CLUE) | | -------------- | :---: | :----: | :----------: | :---: | :---------: | :-----------: | :---------: | | RoBERTa-Tiny | 72.3 | 83.4 | 91.4 | 81.8 | 62.0 | 55.0 | 60.3 | | RoBERTa-Mini | 75.9 | 85.7 | 93.7 | 86.1 | 63.9 | 58.3 | 67.4 | | RoBERTa-Small | 76.9 | 87.5 | 93.4 | 86.5 | 65.1 | 59.4 | 69.7 | | RoBERTa-Medium | 78.0 | 88.7 | 94.8 | 88.1 | 65.6 | 59.5 | 71.2 | | RoBERTa-Base | 79.7 | 90.1 | 95.2 | 89.2 | 67.0 | 60.9 | 75.5 | For each task, we selected the best fine-tuning hyperparameters from the lists below, and trained with the sequence length of 128: - epochs: 3, 5, 8 - batch sizes: 32, 64 - learning rates: 3e-5, 1e-4, 3e-4 ## How to use You can use this model directly with a pipeline for masked language modeling (take the case of RoBERTa-Medium): ```python >>> from transformers import pipeline >>> unmasker = pipeline('fill-mask', model='uer/chinese_roberta_L-8_H-512') >>> unmasker("中国的首都是[MASK]京。") [ {'sequence': '[CLS] 中国的首都是北京。 [SEP]', 'score': 0.8701988458633423, 'token': 1266, 'token_str': '北'}, {'sequence': '[CLS] 中国的首都是南京。 [SEP]', 'score': 0.1194809079170227, 'token': 1298, 'token_str': '南'}, {'sequence': '[CLS] 中国的首都是东京。 [SEP]', 'score': 0.0037803512532263994, 'token': 691, 'token_str': '东'}, {'sequence': '[CLS] 中国的首都是普京。 [SEP]', 'score': 0.0017127094324678183, 'token': 3249, 'token_str': '普'}, {'sequence': '[CLS] 中国的首都是望京。 [SEP]', 'score': 0.001687526935711503, 'token': 3307, 'token_str': '望'} ] ``` Here is how to use this model to get the features of a given text in PyTorch: ```python from transformers import BertTokenizer, BertModel tokenizer = BertTokenizer.from_pretrained('uer/chinese_roberta_L-8_H-512') model = BertModel.from_pretrained("uer/chinese_roberta_L-8_H-512") text = "用你喜欢的任何文本替换我。" encoded_input = tokenizer(text, return_tensors='pt') output = model(**encoded_input) ``` and in TensorFlow: ```python from transformers import BertTokenizer, TFBertModel tokenizer = BertTokenizer.from_pretrained('uer/chinese_roberta_L-8_H-512') model = TFBertModel.from_pretrained("uer/chinese_roberta_L-8_H-512") text = "用你喜欢的任何文本替换我。" encoded_input = tokenizer(text, return_tensors='tf') output = model(encoded_input) ``` ## Training data [CLUECorpusSmall](https://github.com/CLUEbenchmark/CLUECorpus2020/) is used as training data. We found that models pre-trained on CLUECorpusSmall outperform those pre-trained on CLUECorpus2020, although CLUECorpus2020 is much larger than CLUECorpusSmall. ## Training procedure Models are pre-trained by [UER-py](https://github.com/dbiir/UER-py/) on [Tencent Cloud](https://cloud.tencent.com/). We pre-train 1,000,000 steps with a sequence length of 128 and then pre-train 250,000 additional steps with a sequence length of 512. We use the same hyper-parameters on different model sizes. Taking the case of RoBERTa-Medium Stage1: ``` python3 preprocess.py --corpus_path corpora/cluecorpussmall.txt \ --vocab_path models/google_zh_vocab.txt \ --dataset_path cluecorpussmall_seq128_dataset.pt \ --processes_num 32 --seq_length 128 \ --dynamic_masking --data_processor mlm ``` ``` python3 pretrain.py --dataset_path cluecorpussmall_seq128_dataset.pt \ --vocab_path models/google_zh_vocab.txt \ --config_path models/bert/medium_config.json \ --output_model_path models/cluecorpussmall_roberta_medium_seq128_model.bin \ --world_size 8 --gpu_ranks 0 1 2 3 4 5 6 7 \ --total_steps 1000000 --save_checkpoint_steps 100000 --report_steps 50000 \ --learning_rate 1e-4 --batch_size 64 \ --data_processor mlm --target mlm ``` Stage2: ``` python3 preprocess.py --corpus_path corpora/cluecorpussmall.txt \ --vocab_path models/google_zh_vocab.txt \ --dataset_path cluecorpussmall_seq512_dataset.pt \ --processes_num 32 --seq_length 512 \ --dynamic_masking --data_processor mlm ``` ``` python3 pretrain.py --dataset_path cluecorpussmall_seq512_dataset.pt \ --vocab_path models/google_zh_vocab.txt \ --pretrained_model_path models/cluecorpussmall_roberta_medium_seq128_model.bin-1000000 \ --config_path models/bert/medium_config.json \ --output_model_path models/cluecorpussmall_roberta_medium_seq512_model.bin \ --world_size 8 --gpu_ranks 0 1 2 3 4 5 6 7 \ --total_steps 250000 --save_checkpoint_steps 50000 --report_steps 10000 \ --learning_rate 5e-5 --batch_size 16 \ --data_processor mlm --target mlm ``` Finally, we convert the pre-trained model into Huggingface's format: ``` python3 scripts/convert_bert_from_uer_to_huggingface.py --input_model_path models/cluecorpussmall_roberta_medium_seq512_model.bin-250000 \ --output_model_path pytorch_model.bin \ --layers_num 8 --type mlm ``` ### BibTeX entry and citation info ``` @article{devlin2018bert, title={Bert: Pre-training of deep bidirectional transformers for language understanding}, author={Devlin, Jacob and Chang, Ming-Wei and Lee, Kenton and Toutanova, Kristina}, journal={arXiv preprint arXiv:1810.04805}, year={2018} } @article{liu2019roberta, title={Roberta: A robustly optimized bert pretraining approach}, author={Liu, Yinhan and Ott, Myle and Goyal, Naman and Du, Jingfei and Joshi, Mandar and Chen, Danqi and Levy, Omer and Lewis, Mike and Zettlemoyer, Luke and Stoyanov, Veselin}, journal={arXiv preprint arXiv:1907.11692}, year={2019} } @article{turc2019, title={Well-Read Students Learn Better: On the Importance of Pre-training Compact Models}, author={Turc, Iulia and Chang, Ming-Wei and Lee, Kenton and Toutanova, Kristina}, journal={arXiv preprint arXiv:1908.08962v2 }, year={2019} } @article{zhao2019uer, title={UER: An Open-Source Toolkit for Pre-training Models}, author={Zhao, Zhe and Chen, Hui and Zhang, Jinbin and Zhao, Xin and Liu, Tao and Lu, Wei and Chen, Xi and Deng, Haotang and Ju, Qi and Du, Xiaoyong}, journal={EMNLP-IJCNLP 2019}, pages={241}, year={2019} } @article{zhao2023tencentpretrain, title={TencentPretrain: A Scalable and Flexible Toolkit for Pre-training Models of Different Modalities}, author={Zhao, Zhe and Li, Yudong and Hou, Cheng and Zhao, Jing and others}, journal={ACL 2023}, pages={217}, year={2023} } ``` [2_128]:https://huggingface.co/uer/chinese_roberta_L-2_H-128 [2_256]:https://huggingface.co/uer/chinese_roberta_L-2_H-256 [2_512]:https://huggingface.co/uer/chinese_roberta_L-2_H-512 [2_768]:https://huggingface.co/uer/chinese_roberta_L-2_H-768 [4_128]:https://huggingface.co/uer/chinese_roberta_L-4_H-128 [4_256]:https://huggingface.co/uer/chinese_roberta_L-4_H-256 [4_512]:https://huggingface.co/uer/chinese_roberta_L-4_H-512 [4_768]:https://huggingface.co/uer/chinese_roberta_L-4_H-768 [6_128]:https://huggingface.co/uer/chinese_roberta_L-6_H-128 [6_256]:https://huggingface.co/uer/chinese_roberta_L-6_H-256 [6_512]:https://huggingface.co/uer/chinese_roberta_L-6_H-512 [6_768]:https://huggingface.co/uer/chinese_roberta_L-6_H-768 [8_128]:https://huggingface.co/uer/chinese_roberta_L-8_H-128 [8_256]:https://huggingface.co/uer/chinese_roberta_L-8_H-256 [8_512]:https://huggingface.co/uer/chinese_roberta_L-8_H-512 [8_768]:https://huggingface.co/uer/chinese_roberta_L-8_H-768 [10_128]:https://huggingface.co/uer/chinese_roberta_L-10_H-128 [10_256]:https://huggingface.co/uer/chinese_roberta_L-10_H-256 [10_512]:https://huggingface.co/uer/chinese_roberta_L-10_H-512 [10_768]:https://huggingface.co/uer/chinese_roberta_L-10_H-768 [12_128]:https://huggingface.co/uer/chinese_roberta_L-12_H-128 [12_256]:https://huggingface.co/uer/chinese_roberta_L-12_H-256 [12_512]:https://huggingface.co/uer/chinese_roberta_L-12_H-512 [12_768]:https://huggingface.co/uer/chinese_roberta_L-12_H-768

uer/chinese_roberta_L-4_H-768

uer

2023-08-30T02:25:50Z

109

0

transformers

[ "transformers", "pytorch", "tf", "jax", "bert", "fill-mask", "zh", "dataset:CLUECorpusSmall", "arxiv:1909.05658", "arxiv:2212.06385", "arxiv:1908.08962", "arxiv:1810.04805", "arxiv:1907.11692", "autotrain_compatible", "endpoints_compatible", "region:us" ]

fill-mask

2022-03-02T23:29:05Z

--- language: zh datasets: CLUECorpusSmall widget: - text: "北京是[MASK]国的首都。" --- # Chinese RoBERTa Miniatures ## Model description This is the set of 24 Chinese RoBERTa models pre-trained by [UER-py](https://github.com/dbiir/UER-py/), which is introduced in [this paper](https://arxiv.org/abs/1909.05658). Besides, the models could also be pre-trained by [TencentPretrain](https://github.com/Tencent/TencentPretrain) introduced in [this paper](https://arxiv.org/abs/2212.06385), which inherits UER-py to support models with parameters above one billion, and extends it to a multimodal pre-training framework. [Turc et al.](https://arxiv.org/abs/1908.08962) have shown that the standard BERT recipe is effective on a wide range of model sizes. Following their paper, we released the 24 Chinese RoBERTa models. In order to facilitate users in reproducing the results, we used a publicly available corpus and provided all training details. You can download the 24 Chinese RoBERTa miniatures either from the [UER-py Modelzoo page](https://github.com/dbiir/UER-py/wiki/Modelzoo), or via HuggingFace from the links below: | | H=128 | H=256 | H=512 | H=768 | | -------- | :-----------------------: | :-----------------------: | :-------------------------: | :-------------------------: | | **L=2** | [**2/128 (Tiny)**][2_128] | [2/256][2_256] | [2/512][2_512] | [2/768][2_768] | | **L=4** | [4/128][4_128] | [**4/256 (Mini)**][4_256] | [**4/512 (Small)**][4_512] | [4/768][4_768] | | **L=6** | [6/128][6_128] | [6/256][6_256] | [6/512][6_512] | [6/768][6_768] | | **L=8** | [8/128][8_128] | [8/256][8_256] | [**8/512 (Medium)**][8_512] | [8/768][8_768] | | **L=10** | [10/128][10_128] | [10/256][10_256] | [10/512][10_512] | [10/768][10_768] | | **L=12** | [12/128][12_128] | [12/256][12_256] | [12/512][12_512] | [**12/768 (Base)**][12_768] | Here are scores on the devlopment set of six Chinese tasks: | Model | Score | book_review | chnsenticorp | lcqmc | tnews(CLUE) | iflytek(CLUE) | ocnli(CLUE) | | -------------- | :---: | :----: | :----------: | :---: | :---------: | :-----------: | :---------: | | RoBERTa-Tiny | 72.3 | 83.4 | 91.4 | 81.8 | 62.0 | 55.0 | 60.3 | | RoBERTa-Mini | 75.9 | 85.7 | 93.7 | 86.1 | 63.9 | 58.3 | 67.4 | | RoBERTa-Small | 76.9 | 87.5 | 93.4 | 86.5 | 65.1 | 59.4 | 69.7 | | RoBERTa-Medium | 78.0 | 88.7 | 94.8 | 88.1 | 65.6 | 59.5 | 71.2 | | RoBERTa-Base | 79.7 | 90.1 | 95.2 | 89.2 | 67.0 | 60.9 | 75.5 | For each task, we selected the best fine-tuning hyperparameters from the lists below, and trained with the sequence length of 128: - epochs: 3, 5, 8 - batch sizes: 32, 64 - learning rates: 3e-5, 1e-4, 3e-4 ## How to use You can use this model directly with a pipeline for masked language modeling (take the case of RoBERTa-Medium): ```python >>> from transformers import pipeline >>> unmasker = pipeline('fill-mask', model='uer/chinese_roberta_L-8_H-512') >>> unmasker("中国的首都是[MASK]京。") [ {'sequence': '[CLS] 中国的首都是北京。 [SEP]', 'score': 0.8701988458633423, 'token': 1266, 'token_str': '北'}, {'sequence': '[CLS] 中国的首都是南京。 [SEP]', 'score': 0.1194809079170227, 'token': 1298, 'token_str': '南'}, {'sequence': '[CLS] 中国的首都是东京。 [SEP]', 'score': 0.0037803512532263994, 'token': 691, 'token_str': '东'}, {'sequence': '[CLS] 中国的首都是普京。 [SEP]', 'score': 0.0017127094324678183, 'token': 3249, 'token_str': '普'}, {'sequence': '[CLS] 中国的首都是望京。 [SEP]', 'score': 0.001687526935711503, 'token': 3307, 'token_str': '望'} ] ``` Here is how to use this model to get the features of a given text in PyTorch: ```python from transformers import BertTokenizer, BertModel tokenizer = BertTokenizer.from_pretrained('uer/chinese_roberta_L-8_H-512') model = BertModel.from_pretrained("uer/chinese_roberta_L-8_H-512") text = "用你喜欢的任何文本替换我。" encoded_input = tokenizer(text, return_tensors='pt') output = model(**encoded_input) ``` and in TensorFlow: ```python from transformers import BertTokenizer, TFBertModel tokenizer = BertTokenizer.from_pretrained('uer/chinese_roberta_L-8_H-512') model = TFBertModel.from_pretrained("uer/chinese_roberta_L-8_H-512") text = "用你喜欢的任何文本替换我。" encoded_input = tokenizer(text, return_tensors='tf') output = model(encoded_input) ``` ## Training data [CLUECorpusSmall](https://github.com/CLUEbenchmark/CLUECorpus2020/) is used as training data. We found that models pre-trained on CLUECorpusSmall outperform those pre-trained on CLUECorpus2020, although CLUECorpus2020 is much larger than CLUECorpusSmall. ## Training procedure Models are pre-trained by [UER-py](https://github.com/dbiir/UER-py/) on [Tencent Cloud](https://cloud.tencent.com/). We pre-train 1,000,000 steps with a sequence length of 128 and then pre-train 250,000 additional steps with a sequence length of 512. We use the same hyper-parameters on different model sizes. Taking the case of RoBERTa-Medium Stage1: ``` python3 preprocess.py --corpus_path corpora/cluecorpussmall.txt \ --vocab_path models/google_zh_vocab.txt \ --dataset_path cluecorpussmall_seq128_dataset.pt \ --processes_num 32 --seq_length 128 \ --dynamic_masking --data_processor mlm ``` ``` python3 pretrain.py --dataset_path cluecorpussmall_seq128_dataset.pt \ --vocab_path models/google_zh_vocab.txt \ --config_path models/bert/medium_config.json \ --output_model_path models/cluecorpussmall_roberta_medium_seq128_model.bin \ --world_size 8 --gpu_ranks 0 1 2 3 4 5 6 7 \ --total_steps 1000000 --save_checkpoint_steps 100000 --report_steps 50000 \ --learning_rate 1e-4 --batch_size 64 \ --data_processor mlm --target mlm ``` Stage2: ``` python3 preprocess.py --corpus_path corpora/cluecorpussmall.txt \ --vocab_path models/google_zh_vocab.txt \ --dataset_path cluecorpussmall_seq512_dataset.pt \ --processes_num 32 --seq_length 512 \ --dynamic_masking --data_processor mlm ``` ``` python3 pretrain.py --dataset_path cluecorpussmall_seq512_dataset.pt \ --vocab_path models/google_zh_vocab.txt \ --pretrained_model_path models/cluecorpussmall_roberta_medium_seq128_model.bin-1000000 \ --config_path models/bert/medium_config.json \ --output_model_path models/cluecorpussmall_roberta_medium_seq512_model.bin \ --world_size 8 --gpu_ranks 0 1 2 3 4 5 6 7 \ --total_steps 250000 --save_checkpoint_steps 50000 --report_steps 10000 \ --learning_rate 5e-5 --batch_size 16 \ --data_processor mlm --target mlm ``` Finally, we convert the pre-trained model into Huggingface's format: ``` python3 scripts/convert_bert_from_uer_to_huggingface.py --input_model_path models/cluecorpussmall_roberta_medium_seq512_model.bin-250000 \ --output_model_path pytorch_model.bin \ --layers_num 8 --type mlm ``` ### BibTeX entry and citation info ``` @article{devlin2018bert, title={Bert: Pre-training of deep bidirectional transformers for language understanding}, author={Devlin, Jacob and Chang, Ming-Wei and Lee, Kenton and Toutanova, Kristina}, journal={arXiv preprint arXiv:1810.04805}, year={2018} } @article{liu2019roberta, title={Roberta: A robustly optimized bert pretraining approach}, author={Liu, Yinhan and Ott, Myle and Goyal, Naman and Du, Jingfei and Joshi, Mandar and Chen, Danqi and Levy, Omer and Lewis, Mike and Zettlemoyer, Luke and Stoyanov, Veselin}, journal={arXiv preprint arXiv:1907.11692}, year={2019} } @article{turc2019, title={Well-Read Students Learn Better: On the Importance of Pre-training Compact Models}, author={Turc, Iulia and Chang, Ming-Wei and Lee, Kenton and Toutanova, Kristina}, journal={arXiv preprint arXiv:1908.08962v2 }, year={2019} } @article{zhao2019uer, title={UER: An Open-Source Toolkit for Pre-training Models}, author={Zhao, Zhe and Chen, Hui and Zhang, Jinbin and Zhao, Xin and Liu, Tao and Lu, Wei and Chen, Xi and Deng, Haotang and Ju, Qi and Du, Xiaoyong}, journal={EMNLP-IJCNLP 2019}, pages={241}, year={2019} } @article{zhao2023tencentpretrain, title={TencentPretrain: A Scalable and Flexible Toolkit for Pre-training Models of Different Modalities}, author={Zhao, Zhe and Li, Yudong and Hou, Cheng and Zhao, Jing and others}, journal={ACL 2023}, pages={217}, year={2023} } ``` [2_128]:https://huggingface.co/uer/chinese_roberta_L-2_H-128 [2_256]:https://huggingface.co/uer/chinese_roberta_L-2_H-256 [2_512]:https://huggingface.co/uer/chinese_roberta_L-2_H-512 [2_768]:https://huggingface.co/uer/chinese_roberta_L-2_H-768 [4_128]:https://huggingface.co/uer/chinese_roberta_L-4_H-128 [4_256]:https://huggingface.co/uer/chinese_roberta_L-4_H-256 [4_512]:https://huggingface.co/uer/chinese_roberta_L-4_H-512 [4_768]:https://huggingface.co/uer/chinese_roberta_L-4_H-768 [6_128]:https://huggingface.co/uer/chinese_roberta_L-6_H-128 [6_256]:https://huggingface.co/uer/chinese_roberta_L-6_H-256 [6_512]:https://huggingface.co/uer/chinese_roberta_L-6_H-512 [6_768]:https://huggingface.co/uer/chinese_roberta_L-6_H-768 [8_128]:https://huggingface.co/uer/chinese_roberta_L-8_H-128 [8_256]:https://huggingface.co/uer/chinese_roberta_L-8_H-256 [8_512]:https://huggingface.co/uer/chinese_roberta_L-8_H-512 [8_768]:https://huggingface.co/uer/chinese_roberta_L-8_H-768 [10_128]:https://huggingface.co/uer/chinese_roberta_L-10_H-128 [10_256]:https://huggingface.co/uer/chinese_roberta_L-10_H-256 [10_512]:https://huggingface.co/uer/chinese_roberta_L-10_H-512 [10_768]:https://huggingface.co/uer/chinese_roberta_L-10_H-768 [12_128]:https://huggingface.co/uer/chinese_roberta_L-12_H-128 [12_256]:https://huggingface.co/uer/chinese_roberta_L-12_H-256 [12_512]:https://huggingface.co/uer/chinese_roberta_L-12_H-512 [12_768]:https://huggingface.co/uer/chinese_roberta_L-12_H-768

uer/chinese_roberta_L-4_H-256

uer

2023-08-30T02:25:07Z

117

3

transformers

[ "transformers", "pytorch", "tf", "jax", "bert", "fill-mask", "zh", "dataset:CLUECorpusSmall", "arxiv:1909.05658", "arxiv:2212.06385", "arxiv:1908.08962", "arxiv:1810.04805", "arxiv:1907.11692", "autotrain_compatible", "endpoints_compatible", "region:us" ]

fill-mask

2022-03-02T23:29:05Z

--- language: zh datasets: CLUECorpusSmall widget: - text: "北京是[MASK]国的首都。" --- # Chinese RoBERTa Miniatures ## Model description This is the set of 24 Chinese RoBERTa models pre-trained by [UER-py](https://github.com/dbiir/UER-py/), which is introduced in [this paper](https://arxiv.org/abs/1909.05658). Besides, the models could also be pre-trained by [TencentPretrain](https://github.com/Tencent/TencentPretrain) introduced in [this paper](https://arxiv.org/abs/2212.06385), which inherits UER-py to support models with parameters above one billion, and extends it to a multimodal pre-training framework. [Turc et al.](https://arxiv.org/abs/1908.08962) have shown that the standard BERT recipe is effective on a wide range of model sizes. Following their paper, we released the 24 Chinese RoBERTa models. In order to facilitate users in reproducing the results, we used a publicly available corpus and provided all training details. You can download the 24 Chinese RoBERTa miniatures either from the [UER-py Modelzoo page](https://github.com/dbiir/UER-py/wiki/Modelzoo), or via HuggingFace from the links below: | | H=128 | H=256 | H=512 | H=768 | | -------- | :-----------------------: | :-----------------------: | :-------------------------: | :-------------------------: | | **L=2** | [**2/128 (Tiny)**][2_128] | [2/256][2_256] | [2/512][2_512] | [2/768][2_768] | | **L=4** | [4/128][4_128] | [**4/256 (Mini)**][4_256] | [**4/512 (Small)**][4_512] | [4/768][4_768] | | **L=6** | [6/128][6_128] | [6/256][6_256] | [6/512][6_512] | [6/768][6_768] | | **L=8** | [8/128][8_128] | [8/256][8_256] | [**8/512 (Medium)**][8_512] | [8/768][8_768] | | **L=10** | [10/128][10_128] | [10/256][10_256] | [10/512][10_512] | [10/768][10_768] | | **L=12** | [12/128][12_128] | [12/256][12_256] | [12/512][12_512] | [**12/768 (Base)**][12_768] | Here are scores on the devlopment set of six Chinese tasks: | Model | Score | book_review | chnsenticorp | lcqmc | tnews(CLUE) | iflytek(CLUE) | ocnli(CLUE) | | -------------- | :---: | :----: | :----------: | :---: | :---------: | :-----------: | :---------: | | RoBERTa-Tiny | 72.3 | 83.4 | 91.4 | 81.8 | 62.0 | 55.0 | 60.3 | | RoBERTa-Mini | 75.9 | 85.7 | 93.7 | 86.1 | 63.9 | 58.3 | 67.4 | | RoBERTa-Small | 76.9 | 87.5 | 93.4 | 86.5 | 65.1 | 59.4 | 69.7 | | RoBERTa-Medium | 78.0 | 88.7 | 94.8 | 88.1 | 65.6 | 59.5 | 71.2 | | RoBERTa-Base | 79.7 | 90.1 | 95.2 | 89.2 | 67.0 | 60.9 | 75.5 | For each task, we selected the best fine-tuning hyperparameters from the lists below, and trained with the sequence length of 128: - epochs: 3, 5, 8 - batch sizes: 32, 64 - learning rates: 3e-5, 1e-4, 3e-4 ## How to use You can use this model directly with a pipeline for masked language modeling (take the case of RoBERTa-Medium): ```python >>> from transformers import pipeline >>> unmasker = pipeline('fill-mask', model='uer/chinese_roberta_L-8_H-512') >>> unmasker("中国的首都是[MASK]京。") [ {'sequence': '[CLS] 中国的首都是北京。 [SEP]', 'score': 0.8701988458633423, 'token': 1266, 'token_str': '北'}, {'sequence': '[CLS] 中国的首都是南京。 [SEP]', 'score': 0.1194809079170227, 'token': 1298, 'token_str': '南'}, {'sequence': '[CLS] 中国的首都是东京。 [SEP]', 'score': 0.0037803512532263994, 'token': 691, 'token_str': '东'}, {'sequence': '[CLS] 中国的首都是普京。 [SEP]', 'score': 0.0017127094324678183, 'token': 3249, 'token_str': '普'}, {'sequence': '[CLS] 中国的首都是望京。 [SEP]', 'score': 0.001687526935711503, 'token': 3307, 'token_str': '望'} ] ``` Here is how to use this model to get the features of a given text in PyTorch: ```python from transformers import BertTokenizer, BertModel tokenizer = BertTokenizer.from_pretrained('uer/chinese_roberta_L-8_H-512') model = BertModel.from_pretrained("uer/chinese_roberta_L-8_H-512") text = "用你喜欢的任何文本替换我。" encoded_input = tokenizer(text, return_tensors='pt') output = model(**encoded_input) ``` and in TensorFlow: ```python from transformers import BertTokenizer, TFBertModel tokenizer = BertTokenizer.from_pretrained('uer/chinese_roberta_L-8_H-512') model = TFBertModel.from_pretrained("uer/chinese_roberta_L-8_H-512") text = "用你喜欢的任何文本替换我。" encoded_input = tokenizer(text, return_tensors='tf') output = model(encoded_input) ``` ## Training data [CLUECorpusSmall](https://github.com/CLUEbenchmark/CLUECorpus2020/) is used as training data. We found that models pre-trained on CLUECorpusSmall outperform those pre-trained on CLUECorpus2020, although CLUECorpus2020 is much larger than CLUECorpusSmall. ## Training procedure Models are pre-trained by [UER-py](https://github.com/dbiir/UER-py/) on [Tencent Cloud](https://cloud.tencent.com/). We pre-train 1,000,000 steps with a sequence length of 128 and then pre-train 250,000 additional steps with a sequence length of 512. We use the same hyper-parameters on different model sizes. Taking the case of RoBERTa-Medium Stage1: ``` python3 preprocess.py --corpus_path corpora/cluecorpussmall.txt \ --vocab_path models/google_zh_vocab.txt \ --dataset_path cluecorpussmall_seq128_dataset.pt \ --processes_num 32 --seq_length 128 \ --dynamic_masking --data_processor mlm ``` ``` python3 pretrain.py --dataset_path cluecorpussmall_seq128_dataset.pt \ --vocab_path models/google_zh_vocab.txt \ --config_path models/bert/medium_config.json \ --output_model_path models/cluecorpussmall_roberta_medium_seq128_model.bin \ --world_size 8 --gpu_ranks 0 1 2 3 4 5 6 7 \ --total_steps 1000000 --save_checkpoint_steps 100000 --report_steps 50000 \ --learning_rate 1e-4 --batch_size 64 \ --data_processor mlm --target mlm ``` Stage2: ``` python3 preprocess.py --corpus_path corpora/cluecorpussmall.txt \ --vocab_path models/google_zh_vocab.txt \ --dataset_path cluecorpussmall_seq512_dataset.pt \ --processes_num 32 --seq_length 512 \ --dynamic_masking --data_processor mlm ``` ``` python3 pretrain.py --dataset_path cluecorpussmall_seq512_dataset.pt \ --vocab_path models/google_zh_vocab.txt \ --pretrained_model_path models/cluecorpussmall_roberta_medium_seq128_model.bin-1000000 \ --config_path models/bert/medium_config.json \ --output_model_path models/cluecorpussmall_roberta_medium_seq512_model.bin \ --world_size 8 --gpu_ranks 0 1 2 3 4 5 6 7 \ --total_steps 250000 --save_checkpoint_steps 50000 --report_steps 10000 \ --learning_rate 5e-5 --batch_size 16 \ --data_processor mlm --target mlm ``` Finally, we convert the pre-trained model into Huggingface's format: ``` python3 scripts/convert_bert_from_uer_to_huggingface.py --input_model_path models/cluecorpussmall_roberta_medium_seq512_model.bin-250000 \ --output_model_path pytorch_model.bin \ --layers_num 8 --type mlm ``` ### BibTeX entry and citation info ``` @article{devlin2018bert, title={Bert: Pre-training of deep bidirectional transformers for language understanding}, author={Devlin, Jacob and Chang, Ming-Wei and Lee, Kenton and Toutanova, Kristina}, journal={arXiv preprint arXiv:1810.04805}, year={2018} } @article{liu2019roberta, title={Roberta: A robustly optimized bert pretraining approach}, author={Liu, Yinhan and Ott, Myle and Goyal, Naman and Du, Jingfei and Joshi, Mandar and Chen, Danqi and Levy, Omer and Lewis, Mike and Zettlemoyer, Luke and Stoyanov, Veselin}, journal={arXiv preprint arXiv:1907.11692}, year={2019} } @article{turc2019, title={Well-Read Students Learn Better: On the Importance of Pre-training Compact Models}, author={Turc, Iulia and Chang, Ming-Wei and Lee, Kenton and Toutanova, Kristina}, journal={arXiv preprint arXiv:1908.08962v2 }, year={2019} } @article{zhao2019uer, title={UER: An Open-Source Toolkit for Pre-training Models}, author={Zhao, Zhe and Chen, Hui and Zhang, Jinbin and Zhao, Xin and Liu, Tao and Lu, Wei and Chen, Xi and Deng, Haotang and Ju, Qi and Du, Xiaoyong}, journal={EMNLP-IJCNLP 2019}, pages={241}, year={2019} } @article{zhao2023tencentpretrain, title={TencentPretrain: A Scalable and Flexible Toolkit for Pre-training Models of Different Modalities}, author={Zhao, Zhe and Li, Yudong and Hou, Cheng and Zhao, Jing and others}, journal={ACL 2023}, pages={217}, year={2023} } ``` [2_128]:https://huggingface.co/uer/chinese_roberta_L-2_H-128 [2_256]:https://huggingface.co/uer/chinese_roberta_L-2_H-256 [2_512]:https://huggingface.co/uer/chinese_roberta_L-2_H-512 [2_768]:https://huggingface.co/uer/chinese_roberta_L-2_H-768 [4_128]:https://huggingface.co/uer/chinese_roberta_L-4_H-128 [4_256]:https://huggingface.co/uer/chinese_roberta_L-4_H-256 [4_512]:https://huggingface.co/uer/chinese_roberta_L-4_H-512 [4_768]:https://huggingface.co/uer/chinese_roberta_L-4_H-768 [6_128]:https://huggingface.co/uer/chinese_roberta_L-6_H-128 [6_256]:https://huggingface.co/uer/chinese_roberta_L-6_H-256 [6_512]:https://huggingface.co/uer/chinese_roberta_L-6_H-512 [6_768]:https://huggingface.co/uer/chinese_roberta_L-6_H-768 [8_128]:https://huggingface.co/uer/chinese_roberta_L-8_H-128 [8_256]:https://huggingface.co/uer/chinese_roberta_L-8_H-256 [8_512]:https://huggingface.co/uer/chinese_roberta_L-8_H-512 [8_768]:https://huggingface.co/uer/chinese_roberta_L-8_H-768 [10_128]:https://huggingface.co/uer/chinese_roberta_L-10_H-128 [10_256]:https://huggingface.co/uer/chinese_roberta_L-10_H-256 [10_512]:https://huggingface.co/uer/chinese_roberta_L-10_H-512 [10_768]:https://huggingface.co/uer/chinese_roberta_L-10_H-768 [12_128]:https://huggingface.co/uer/chinese_roberta_L-12_H-128 [12_256]:https://huggingface.co/uer/chinese_roberta_L-12_H-256 [12_512]:https://huggingface.co/uer/chinese_roberta_L-12_H-512 [12_768]:https://huggingface.co/uer/chinese_roberta_L-12_H-768

yoo11/taskC_distilbert_model

yoo11

2023-08-30T02:21:22Z

61

0

transformers

[ "transformers", "tf", "distilbert", "text-classification", "generated_from_keras_callback", "base_model:distilbert/distilbert-base-uncased", "base_model:finetune:distilbert/distilbert-base-uncased", "license:apache-2.0", "autotrain_compatible", "endpoints_compatible", "region:us" ]

text-classification

2023-08-30T02:17:13Z

--- license: apache-2.0 base_model: distilbert-base-uncased tags: - generated_from_keras_callback model-index: - name: yoo11/taskC_distilbert_model results: [] ---  # yoo11/taskC_distilbert_model This model is a fine-tuned version of [distilbert-base-uncased](https://huggingface.co/distilbert-base-uncased) on an unknown dataset. It achieves the following results on the evaluation set: - Train Loss: 1.2046 - Validation Loss: 1.3859 - Train Accuracy: 0.5247 - Epoch: 2 ## Model description More information needed ## Intended uses & limitations More information needed ## Training and evaluation data More information needed ## Training procedure ### Training hyperparameters The following hyperparameters were used during training: - optimizer: {'name': 'Adam', 'weight_decay': None, 'clipnorm': None, 'global_clipnorm': None, 'clipvalue': None, 'use_ema': False, 'ema_momentum': 0.99, 'ema_overwrite_frequency': None, 'jit_compile': True, 'is_legacy_optimizer': False, 'learning_rate': {'class_name': 'PolynomialDecay', 'config': {'initial_learning_rate': 2e-05, 'decay_steps': 1060, 'end_learning_rate': 0.0, 'power': 1.0, 'cycle': False, 'name': None}}, 'beta_1': 0.9, 'beta_2': 0.999, 'epsilon': 1e-08, 'amsgrad': False} - training_precision: float32 ### Training results | Train Loss | Validation Loss | Train Accuracy | Epoch | |:----------:|:---------------:|:--------------:|:-----:| | 1.9115 | 1.6364 | 0.4444 | 0 | | 1.4956 | 1.4539 | 0.4835 | 1 | | 1.2046 | 1.3859 | 0.5247 | 2 | ### Framework versions - Transformers 4.32.1 - TensorFlow 2.12.0 - Datasets 2.14.4 - Tokenizers 0.13.3

Xieyiyiyi/ceshi

Xieyiyiyi

2023-08-30T02:12:23Z

0

null

[ "license:bsl-1.0", "region:us" ]

null

2023-02-21T08:33:29Z

--- license: bsl-1.0 --- [[toc]] - [Refrence111](#Refrence111) ## Refrence 1111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111 1111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111 1111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111 ## Refrence111

Chanblock/Llama-2-7b-chat-hf-250_data_final

Chanblock

2023-08-30T02:08:21Z

0

null

[ "tensorboard", "generated_from_trainer", "base_model:NousResearch/Llama-2-7b-hf", "base_model:finetune:NousResearch/Llama-2-7b-hf", "region:us" ]

null

2023-08-30T02:03:56Z

--- base_model: NousResearch/Llama-2-7b-hf tags: - generated_from_trainer model-index: - name: Llama-2-7b-chat-hf-250_data_final results: [] ---  # Llama-2-7b-chat-hf-250_data_final This model is a fine-tuned version of [NousResearch/Llama-2-7b-hf](https://huggingface.co/NousResearch/Llama-2-7b-hf) on an unknown dataset. ## Model description More information needed ## Intended uses & limitations More information needed ## Training and evaluation data More information needed ## Training procedure ### Training hyperparameters The following hyperparameters were used during training: - learning_rate: 0.0002 - train_batch_size: 3 - eval_batch_size: 3 - seed: 42 - optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08 - lr_scheduler_type: linear - num_epochs: 1 ### Training results ### Framework versions - Transformers 4.31.0 - Pytorch 2.0.1+cu118 - Datasets 2.14.4 - Tokenizers 0.13.3

jason-lee08/my_model

jason-lee08

2023-08-30T01:21:53Z

186

0

transformers

[ "transformers", "safetensors", "gpt_neo", "text-generation", "generated_from_trainer", "base_model:roneneldan/TinyStories-33M", "base_model:finetune:roneneldan/TinyStories-33M", "autotrain_compatible", "endpoints_compatible", "region:us" ]

text-generation

2023-08-30T01:20:23Z

--- base_model: roneneldan/TinyStories-33M tags: - generated_from_trainer model-index: - name: my_model results: [] ---  # my_model This model is a fine-tuned version of [roneneldan/TinyStories-33M](https://huggingface.co/roneneldan/TinyStories-33M) on the None dataset. It achieves the following results on the evaluation set: - Loss: 1.0283 ## Model description More information needed ## Intended uses & limitations More information needed ## Training and evaluation data More information needed ## Training procedure ### Training hyperparameters The following hyperparameters were used during training: - learning_rate: 3e-05 - train_batch_size: 32 - eval_batch_size: 8 - seed: 42 - gradient_accumulation_steps: 16 - total_train_batch_size: 512 - optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08 - lr_scheduler_type: linear - training_steps: 8 ### Training results | Training Loss | Epoch | Step | Validation Loss | |:-------------:|:-----:|:----:|:---------------:| | No log | 0.11 | 2 | 1.1872 | | No log | 0.22 | 4 | 1.0699 | | No log | 0.33 | 6 | 1.0373 | | No log | 0.44 | 8 | 1.0283 | ### Framework versions - Transformers 4.32.1 - Pytorch 2.0.1 - Datasets 2.12.0 - Tokenizers 0.13.2

TomyAI/jojimizugi

TomyAI

2023-08-30T01:18:03Z

0

7

null

[ "license:creativeml-openrail-m", "region:us" ]

null

2023-08-30T00:03:53Z

--- license: creativeml-openrail-m --- 可愛い水着のLoRAです。プロンプト：形状： swimsuit:ワンピース型 tankini:タンクトップ型～チューブトップ型 bikini:三角ビキニ装飾： frilled:主に肩～胸のフリフリ skirt:スカート型模様： printed:模様入 gingham check:ギンガムチェック polca dot:水玉模様 striped:横縞 floral printed:花柄 scale printed:鱗柄（人魚風）これ以外でも(模様の名前) printed swimsuitとすれば大体反応します。ウェイトを1,1,1,1,1,1,1,1,1,0,0,0,1,1,1,0,0 にすると顔への影響が減りますが、柄の精度が落ちます。 ![jojimizugi](jojimizugi.png) ![jojimizugi_grid](jojimizugi_grid.jpg)

Adrian101-hnd/ppo-LunarLander-v2

Adrian101-hnd

2023-08-30T01:05:09Z

1

0

stable-baselines3

[ "stable-baselines3", "LunarLander-v2", "deep-reinforcement-learning", "reinforcement-learning", "model-index", "region:us" ]

reinforcement-learning

2023-08-30T01:04:46Z

--- library_name: stable-baselines3 tags: - LunarLander-v2 - deep-reinforcement-learning - reinforcement-learning - stable-baselines3 model-index: - name: PPO results: - task: type: reinforcement-learning name: reinforcement-learning dataset: name: LunarLander-v2 type: LunarLander-v2 metrics: - type: mean_reward value: 252.32 +/- 21.07 name: mean_reward verified: false --- # **PPO** Agent playing **LunarLander-v2** This is a trained model of a **PPO** agent playing **LunarLander-v2** using the [stable-baselines3 library](https://github.com/DLR-RM/stable-baselines3). ## Usage (with Stable-baselines3) TODO: Add your code ```python from stable_baselines3 import ... from huggingface_sb3 import load_from_hub ... ```

AzulkerAI/Crickey

AzulkerAI

2023-08-30T00:50:12Z

0

null

[ "region:us" ]

null

2023-08-30T00:38:11Z

blahblahblahblah look its funny thingy

touchtech/fashion-images-perspectives-vit-huge-patch14-224-in21k

touchtech

2023-08-30T00:48:28Z

6

0

transformers

[ "transformers", "pytorch", "vit", "image-classification", "vision", "generated_from_trainer", "dataset:imagefolder", "base_model:google/vit-huge-patch14-224-in21k", "base_model:finetune:google/vit-huge-patch14-224-in21k", "license:apache-2.0", "model-index", "autotrain_compatible", "endpoints_compatible", "region:us" ]

image-classification

2023-08-29T19:30:32Z

--- license: apache-2.0 base_model: google/vit-huge-patch14-224-in21k tags: - image-classification - vision - generated_from_trainer datasets: - imagefolder metrics: - accuracy model-index: - name: fashion-images-perspectives-vit-huge-patch14-224-in21k results: - task: name: Image Classification type: image-classification dataset: name: touchtech/fashion-images-perspectives type: imagefolder config: default split: train args: default metrics: - name: Accuracy type: accuracy value: 0.9149976711690732 ---  # fashion-images-perspectives-vit-huge-patch14-224-in21k This model is a fine-tuned version of [google/vit-huge-patch14-224-in21k](https://huggingface.co/google/vit-huge-patch14-224-in21k) on the touchtech/fashion-images-perspectives dataset. It achieves the following results on the evaluation set: - Loss: 0.2604 - Accuracy: 0.9150 ## Model description More information needed ## Intended uses & limitations More information needed ## Training and evaluation data More information needed ## Training procedure ### Training hyperparameters The following hyperparameters were used during training: - learning_rate: 2e-05 - train_batch_size: 8 - eval_batch_size: 8 - seed: 1337 - optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08 - lr_scheduler_type: linear - num_epochs: 5.0 ### Training results | Training Loss | Epoch | Step | Validation Loss | Accuracy | |:-------------:|:-----:|:-----:|:---------------:|:--------:| | 0.5932 | 1.0 | 3042 | 0.3863 | 0.8889 | | 0.4286 | 2.0 | 6084 | 0.4201 | 0.8642 | | 0.3628 | 3.0 | 9126 | 0.2820 | 0.9101 | | 0.3183 | 4.0 | 12168 | 0.2604 | 0.9150 | | 0.2648 | 5.0 | 15210 | 0.2710 | 0.9150 | ### Framework versions - Transformers 4.33.0.dev0 - Pytorch 2.0.1+cu118 - Datasets 2.14.4 - Tokenizers 0.13.3

yoo11/taskA_distilbert_model

yoo11

2023-08-30T00:23:08Z

61

0

transformers

[ "transformers", "tf", "distilbert", "text-classification", "generated_from_keras_callback", "base_model:distilbert/distilbert-base-uncased", "base_model:finetune:distilbert/distilbert-base-uncased", "license:apache-2.0", "autotrain_compatible", "endpoints_compatible", "region:us" ]

text-classification

2023-08-30T00:15:17Z

--- license: apache-2.0 base_model: distilbert-base-uncased tags: - generated_from_keras_callback model-index: - name: yoo11/taskA_distilbert_model results: [] ---  # yoo11/taskA_distilbert_model This model is a fine-tuned version of [distilbert-base-uncased](https://huggingface.co/distilbert-base-uncased) on an unknown dataset. It achieves the following results on the evaluation set: - Train Loss: 0.1501 - Validation Loss: 0.4055 - Train Accuracy: 0.863 - Epoch: 2 ## Model description More information needed ## Intended uses & limitations More information needed ## Training and evaluation data More information needed ## Training procedure ### Training hyperparameters The following hyperparameters were used during training: - optimizer: {'name': 'Adam', 'weight_decay': None, 'clipnorm': None, 'global_clipnorm': None, 'clipvalue': None, 'use_ema': False, 'ema_momentum': 0.99, 'ema_overwrite_frequency': None, 'jit_compile': True, 'is_legacy_optimizer': False, 'learning_rate': {'class_name': 'PolynomialDecay', 'config': {'initial_learning_rate': 2e-05, 'decay_steps': 4375, 'end_learning_rate': 0.0, 'power': 1.0, 'cycle': False, 'name': None}}, 'beta_1': 0.9, 'beta_2': 0.999, 'epsilon': 1e-08, 'amsgrad': False} - training_precision: float32 ### Training results | Train Loss | Validation Loss | Train Accuracy | Epoch | |:----------:|:---------------:|:--------------:|:-----:| | 0.3800 | 0.3247 | 0.861 | 0 | | 0.2582 | 0.3318 | 0.861 | 1 | | 0.1501 | 0.4055 | 0.863 | 2 | ### Framework versions - Transformers 4.32.1 - TensorFlow 2.12.0 - Datasets 2.14.4 - Tokenizers 0.13.3