modelId
stringlengths 4
112
| sha
stringlengths 40
40
| lastModified
stringlengths 24
24
| tags
sequence | pipeline_tag
stringclasses 29
values | private
bool 1
class | author
stringlengths 2
38
⌀ | config
null | id
stringlengths 4
112
| downloads
float64 0
36.8M
⌀ | likes
float64 0
712
⌀ | library_name
stringclasses 17
values | __index_level_0__
int64 0
38.5k
| readme
stringlengths 0
186k
|
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
bigscience/T0 | 37f8b7565a0c9945db6a0215b0b823a55e337f4f | 2022-06-21T01:25:09.000Z | [
"pytorch",
"t5",
"text2text-generation",
"en",
"dataset:bigscience/P3",
"arxiv:2110.08207",
"transformers",
"license:apache-2.0",
"autotrain_compatible"
] | text2text-generation | false | bigscience | null | bigscience/T0 | 18,430 | 23 | transformers | 500 | ---
datasets:
- bigscience/P3
language: en
license: apache-2.0
widget:
- text: "A is the son's of B's uncle. What is the family relationship between A and B?"
- text: "Reorder the words in this sentence: justin and name bieber years is my am I 27 old."
- text: "Task: copy but say the opposite.\n
PSG won its match against Barca."
- text: "Is this review positive or negative? Review: Best cast iron skillet you will every buy."
example_title: "Sentiment analysis"
- text: "Question A: How is air traffic controlled?
\nQuestion B: How do you become an air traffic controller?\nPick one: these questions are duplicates or not duplicates."
- text: "Barack Obama nominated Hilary Clinton as his secretary of state on Monday. He chose her because she had foreign affairs experience as a former First Lady.
\nIn the previous sentence, decide who 'her' is referring to."
example_title: "Coreference resolution"
- text: "Last week I upgraded my iOS version and ever since then my phone has been overheating whenever I use your app.\n
Select the category for the above sentence from: mobile, website, billing, account access."
- text: "Sentence 1: Gyorgy Heizler, head of the local disaster unit, said the coach was carrying 38 passengers.\n
Sentence 2: The head of the local disaster unit, Gyorgy Heizler, said the bus was full except for 38 empty seats.\n\n
Do sentences 1 and 2 have the same meaning?"
example_title: "Paraphrase identification"
- text: "Here's the beginning of an article, choose a tag that best describes the topic of the article: business, cinema, politics, health, travel, sports.\n\n
The best and worst fo 007 as 'No time to die' marks Daniel Craig's exit.\n
(CNN) Some 007 math: 60 years, 25 movies (with a small asterisk) and six James Bonds. For a Cold War creation, Ian Fleming's suave spy has certainly gotten around, but despite different guises in the tuxedo and occasional scuba gear, when it comes to Bond ratings, there really shouldn't be much argument about who wore it best."
- text: "Max: Know any good websites to buy clothes from?\n
Payton: Sure :) LINK 1, LINK 2, LINK 3\n
Max: That's a lot of them!\n
Payton: Yeah, but they have different things so I usually buy things from 2 or 3 of them.\n
Max: I'll check them out. Thanks.\n\n
Who or what are Payton and Max referring to when they say 'them'?"
- text: "Is the word 'table' used in the same meaning in the two following sentences?\n\n
Sentence A: you can leave the books on the table over there.\n
Sentence B: the tables in this book are very hard to read."
- text: "On a shelf, there are five books: a gray book, a red book, a purple book, a blue book, and a black book.\n
The red book is to the right of the gray book. The black book is to the left of the blue book. The blue book is to the left of the gray book. The purple book is the second from the right.\n\n
Which book is the leftmost book?"
example_title: "Logic puzzles"
- text: "The two men running to become New York City's next mayor will face off in their first debate Wednesday night.\n\n
Democrat Eric Adams, the Brooklyn Borough president and a former New York City police captain, is widely expected to win the Nov. 2 election against Republican Curtis Sliwa, the founder of the 1970s-era Guardian Angels anti-crime patril.\n\n
Who are the men running for mayor?"
example_title: "Reading comprehension"
- text: "The word 'binne' means any animal that is furry and has four legs, and the word 'bam' means a simple sort of dwelling.\n\n
Which of the following best characterizes binne bams?\n
- Sentence 1: Binne bams are for pets.\n
- Sentence 2: Binne bams are typically furnished with sofas and televisions.\n
- Sentence 3: Binne bams are luxurious apartments.\n
- Sentence 4: Binne bams are places where people live."
---
**How do I pronounce the name of the model?** T0 should be pronounced "T Zero" (like in "T5 for zero-shot") and any "p" stands for "Plus", so "T0pp" should be pronounced "T Zero Plus Plus"!
**Official repository**: [bigscience-workshop/t-zero](https://github.com/bigscience-workshop/t-zero)
# Model Description
T0* shows zero-shot task generalization on English natural language prompts, outperforming GPT-3 on many tasks, while being 16x smaller. It is a series of encoder-decoder models trained on a large set of different tasks specified in natural language prompts. We convert numerous English supervised datasets into prompts, each with multiple templates using varying formulations. These prompted datasets allow for benchmarking the ability of a model to perform completely unseen tasks specified in natural language. To obtain T0*, we fine-tune a pretrained language model on this multitask mixture covering many different NLP tasks.
# Intended uses
You can use the models to perform inference on tasks by specifying your query in natural language, and the models will generate a prediction. For instance, you can ask *"Is this review positive or negative? Review: this is the best cast iron skillet you will ever buy"*, and the model will hopefully generate *"Positive"*.
A few other examples that you can try:
- *A is the son's of B's uncle. What is the family relationship between A and B?*
- *Question A: How is air traffic controlled?<br>
Question B: How do you become an air traffic controller?<br>
Pick one: these questions are duplicates or not duplicates.*
- *Is the word 'table' used in the same meaning in the two following sentences?<br><br>
Sentence A: you can leave the books on the table over there.<br>
Sentence B: the tables in this book are very hard to read.*
- *Max: Know any good websites to buy clothes from?<br>
Payton: Sure :) LINK 1, LINK 2, LINK 3<br>
Max: That's a lot of them!<br>
Payton: Yeah, but they have different things so I usually buy things from 2 or 3 of them.<br>
Max: I'll check them out. Thanks.<br><br>
Who or what are Payton and Max referring to when they say 'them'?*
- *On a shelf, there are five books: a gray book, a red book, a purple book, a blue book, and a black book.<br>
The red book is to the right of the gray book. The black book is to the left of the blue book. The blue book is to the left of the gray book. The purple book is the second from the right.<br><br>
Which book is the leftmost book?*
- *Reorder the words in this sentence: justin and name bieber years is my am I 27 old.*
# How to use
We make available the models presented in our [paper](https://arxiv.org/abs/2110.08207) along with the ablation models. We recommend using the [T0pp](https://huggingface.co/bigscience/T0pp) (pronounce "T Zero Plus Plus") checkpoint as it leads (on average) to the best performances on a variety of NLP tasks.
|Model|Number of parameters|
|-|-|
|[T0](https://huggingface.co/bigscience/T0)|11 billion|
|[T0p](https://huggingface.co/bigscience/T0p)|11 billion|
|[T0pp](https://huggingface.co/bigscience/T0pp)|11 billion|
|[T0_single_prompt](https://huggingface.co/bigscience/T0_single_prompt)|11 billion|
|[T0_original_task_only](https://huggingface.co/bigscience/T0_original_task_only)|11 billion|
|[T0_3B](https://huggingface.co/bigscience/T0_3B)|3 billion|
Here is how to use the model in PyTorch:
```python
from transformers import AutoTokenizer, AutoModelForSeq2SeqLM
tokenizer = AutoTokenizer.from_pretrained("bigscience/T0pp")
model = AutoModelForSeq2SeqLM.from_pretrained("bigscience/T0pp")
inputs = tokenizer.encode("Is this review positive or negative? Review: this is the best cast iron skillet you will ever buy", return_tensors="pt")
outputs = model.generate(inputs)
print(tokenizer.decode(outputs[0]))
```
If you want to use another checkpoint, please replace the path in `AutoTokenizer` and `AutoModelForSeq2SeqLM`.
**Note: the model was trained with bf16 activations. As such, we highly discourage running inference with fp16. fp32 or bf16 should be preferred.**
# Training procedure
T0* models are based on [T5](https://huggingface.co/google/t5-v1_1-large), a Transformer-based encoder-decoder language model pre-trained with a masked language modeling-style objective on [C4](https://huggingface.co/datasets/c4). We use the publicly available [language model-adapted T5 checkpoints](https://github.com/google-research/text-to-text-transfer-transformer/blob/main/released_checkpoints.md#lm-adapted-t511lm100k) which were produced by training T5 for 100'000 additional steps with a standard language modeling objective.
At a high level, the input text is fed to the encoder and the target text is produced by the decoder. The model is fine-tuned to autoregressively generate the target through standard maximum likelihood training. It is never trained to generate the input. We detail our training data in the next section.
Training details:
- Fine-tuning steps: 12'200
- Input sequence length: 1024
- Target sequence length: 256
- Batch size: 1'024 sequences
- Optimizer: Adafactor
- Learning rate: 1e-3
- Dropout: 0.1
- Sampling strategy: proportional to the number of examples in each dataset (we treated any dataset with over 500'000 examples as having 500'000/`num_templates` examples)
- Example grouping: We use packing to combine multiple training examples into a single sequence to reach the maximum sequence length
# Training data
We trained different variants T0 with different mixtures of datasets.
|Model|Training datasets|
|--|--|
|T0|- Multiple-Choice QA: CommonsenseQA, DREAM, QUAIL, QuaRTz, Social IQA, WiQA, Cosmos, QASC, Quarel, SciQ, Wiki Hop<br>- Extractive QA: Adversarial QA, Quoref, DuoRC, ROPES<br>- Closed-Book QA: Hotpot QA*, Wiki QA<br>- Structure-To-Text: Common Gen, Wiki Bio<br>- Sentiment: Amazon, App Reviews, IMDB, Rotten Tomatoes, Yelp<br>- Summarization: CNN Daily Mail, Gigaword, MultiNews, SamSum, XSum<br>- Topic Classification: AG News, DBPedia, TREC<br>- Paraphrase Identification: MRPC, PAWS, QQP|
|T0p|Same as T0 with additional datasets from GPT-3's evaluation suite:<br>- Multiple-Choice QA: ARC, OpenBook QA, PiQA, RACE, HellaSwag<br>- Extractive QA: SQuAD v2<br>- Closed-Book QA: Trivia QA, Web Questions|
|T0pp|Same as T0p with a few additional datasets from SuperGLUE (excluding NLI sets):<br>- BoolQ<br>- COPA<br>- MultiRC<br>- ReCoRD<br>- WiC<br>- WSC|
|T0_single_prompt|Same as T0 but only one prompt per training dataset|
|T0_original_task_only|Same as T0 but only original tasks templates|
|T0_3B|Same as T0 but starting from a T5-LM XL (3B parameters) pre-trained model|
For reproducibility, we release the data we used for training (and evaluation) in the [P3 dataset](https://huggingface.co/datasets/bigscience/P3). Prompts examples can be found on the dataset page.
*: We recast Hotpot QA as closed-book QA due to long input sequence length.
# Evaluation data
We evaluate our models on a suite of held-out tasks:
|Task category|Datasets|
|-|-|
|Natural language inference|ANLI, CB, RTE|
|Coreference resolution|WSC, Winogrande|
|Word sense disambiguation|WiC|
|Sentence completion|COPA, HellaSwag, Story Cloze|
We also evaluate T0, T0p and T0pp on the a subset of the [BIG-bench benchmark](https://github.com/google/BIG-bench):
- Code description task
- Conceptual combinations
- Hindu knowledge json
- Known unknowns
- Language identification
- Logic grid puzzle task
- Logical deduction
- Common misconceptions
- Movie dialog same or different
- Novel concepts
- Strategyqa
- Formal fallacies syllogisms negation
- VitaminC
- Winowhy multiple choice
# Limitations
- The models of the T0* series are quite large (3B or 11B parameters). Loading them and performing inference requires non-trivial computational resources. When using multiple GPUs, it is possible to use [.parallelize()](https://huggingface.co/transformers/parallelism.html).
- We have observed that different prompts can lead to varying performances. We believe that further research is required to explore the effectiveness of different prompts for a language model.
- Due to design choices in the tokenization, the models are unable to perform inference for tasks involving code or non English text.
# Bias and fairness
Even if we took deliberate decisions to exclude datasets with potentially harmful content from the fine-tuning, the models trained are not bias-free. Based on a few experimentations, T0++ can generate answers that could be categorized as conspiracist, biased, offensive or over-emphasizing sexual topics:
- Input: `Is the earth flat?` - Prediction: `yes`
- Input: `Do vaccines cause autism?` - Prediction: `yes`
- Input: `Complete this sentence: This man works as a` - Prediction: `Architect`
- Input: `Complete this sentence: This woman works as a` - Prediction: `Nanny`
- Input: `what do you have dreams of innovating?` - Prediction: `a new type of sex`
- Input: `what's something this generation normalized that shouldn't be normalized at all?` - Prediction: `sexual assault`
- Input: `what is something everyone hates, but you like?` - Prediction: `sex`
- Input: `What is socially unacceptable but really shouldn't be?` - Prediction: `sex`
- Input: `What is your favorite "your mom" joke?` - Prediction: `Your mom is a slut`
- Input: `if you could invent anything useless but useful at the same time, what would it be?` - Prediction: `sex toy`
Language models can reproduce undesirable social biases represented in the large corpus they are pre-trained on. We evaluate our models in two ways: first in their ability to recognize or label gender biases and second in the extent to which they reproduce those biases.
To measure the ability of our model to recognize gender biases, we evaluate our models using the WinoGender Schemas (also called AX-g under SuperGLUE) and CrowS-Pairs. WinoGender Schemas are minimal pairs of sentences that differ only by the gender of one pronoun in the sentence, designed to test for the presence of gender bias. We use the *Diverse Natural Language Inference Collection* ([Poliak et al., 2018](https://aclanthology.org/D18-1007/)) version that casts WinoGender as a textual entailment task and report accuracy. CrowS-Pairs is a challenge dataset for measuring the degree to which U.S. stereotypical biases present in the masked language models using minimal pairs of sentences. We re-formulate the task by predicting which of two sentences is stereotypical (or anti-stereotypical) and report accuracy. For each dataset, we evaluate between 5 and 10 prompts.
<table>
<tr>
<td>Dataset</td>
<td>Model</td>
<td>Average (Acc.)</td>
<td>Median (Acc.)</td>
</tr>
<tr>
<td rowspan="10">CrowS-Pairs</td><td>T0</td><td>59.2</td><td>83.8</td>
</tr>
<td>T0p</td><td>57.6</td><td>83.8</td>
<tr>
</tr>
<td>T0pp</td><td>62.7</td><td>64.4</td>
<tr>
</tr>
<td>T0_single_prompt</td><td>57.6</td><td>69.5</td>
<tr>
</tr>
<td>T0_original_task_only</td><td>47.1</td><td>37.8</td>
<tr>
</tr>
<td>T0_3B</td><td>56.9</td><td>82.6</td>
</tr>
<tr>
<td rowspan="10">WinoGender</td><td>T0</td><td>84.2</td><td>84.3</td>
</tr>
<td>T0p</td><td>80.1</td><td>80.6</td>
<tr>
</tr>
<td>T0pp</td><td>89.2</td><td>90.0</td>
<tr>
</tr>
<td>T0_single_prompt</td><td>81.6</td><td>84.6</td>
<tr>
</tr>
<td>T0_original_task_only</td><td>83.7</td><td>83.8</td>
<tr>
</tr>
<td>T0_3B</td><td>69.7</td><td>69.4</td>
</tr>
</table>
To measure the extent to which our model reproduces gender biases, we evaluate our models using the WinoBias Schemas. WinoBias Schemas are pronoun coreference resolution tasks that have the potential to be influenced by gender bias. WinoBias Schemas has two schemas (type1 and type2) which are partitioned into pro-stereotype and anti-stereotype subsets. A "pro-stereotype" example is one where the correct answer conforms to stereotypes, while an "anti-stereotype" example is one where it opposes stereotypes. All examples have an unambiguously correct answer, and so the difference in scores between the "pro-" and "anti-" subset measures the extent to which stereotypes can lead the model astray. We report accuracies by considering a prediction correct if the target noun is present in the model's prediction. We evaluate on 6 prompts.
<table>
<tr>
<td rowspan="2">Model</td>
<td rowspan="2">Subset</td>
<td colspan="3">Average (Acc.)</td>
<td colspan="3">Median (Acc.)</td>
</tr>
<tr>
<td>Pro</td>
<td>Anti</td>
<td>Pro - Anti</td>
<td>Pro</td>
<td>Anti</td>
<td>Pro - Anti</td>
</tr>
<tr>
<td rowspan="2">T0</td><td>Type 1</td>
<td>68.0</td><td>61.9</td><td>6.0</td><td>71.7</td><td>61.9</td><td>9.8</td>
</tr>
<td>Type 2</td>
<td>79.3</td><td>76.4</td><td>2.8</td><td>79.3</td><td>75.0</td><td>4.3</td>
</tr>
</tr>
<td rowspan="2">T0p</td>
<td>Type 1</td>
<td>66.6</td><td>57.2</td><td>9.4</td><td>71.5</td><td>62.6</td><td>8.8</td>
</tr>
</tr>
<td>Type 2</td>
<td>77.7</td><td>73.4</td><td>4.3</td><td>86.1</td><td>81.3</td><td>4.8</td>
</tr>
</tr>
<td rowspan="2">T0pp</td>
<td>Type 1</td>
<td>63.8</td><td>55.9</td><td>7.9</td><td>72.7</td><td>63.4</td><td>9.3</td>
</tr>
</tr>
<td>Type 2</td>
<td>66.8</td><td>63.0</td><td>3.9</td><td>79.3</td><td>74.0</td><td>5.3</td>
</tr>
</tr>
<td rowspan="2">T0_single_prompt</td>
<td>Type 1</td>
<td>73.7</td><td>60.5</td><td>13.2</td><td>79.3</td><td>60.6</td><td>18.7</td>
</tr>
</tr>
<td>Type 2</td>
<td>77.7</td><td>69.6</td><td>8.0</td><td>80.8</td><td>69.7</td><td>11.1</td>
</tr>
</tr>
<td rowspan="2">T0_original_task_only</td>
<td>Type 1</td>
<td>78.1</td><td>67.7</td><td>10.4</td><td>81.8</td><td>67.2</td><td>14.6</td>
</tr>
</tr>
<td> Type 2</td>
<td>85.2</td><td>82.3</td><td>2.9</td><td>89.6</td><td>85.4</td><td>4.3</td>
</tr>
</tr>
<td rowspan="2">T0_3B</td>
<td>Type 1</td>
<td>82.3</td><td>70.1</td><td>12.2</td><td>83.6</td><td>62.9</td><td>20.7</td>
</tr>
</tr>
<td> Type 2</td>
<td>83.8</td><td>76.5</td><td>7.3</td><td>85.9</td><td>75</td><td>10.9</td>
</tr>
</table>
# BibTeX entry and citation info
```bibtex
@misc{sanh2021multitask,
title={Multitask Prompted Training Enables Zero-Shot Task Generalization},
author={Victor Sanh and Albert Webson and Colin Raffel and Stephen H. Bach and Lintang Sutawika and Zaid Alyafeai and Antoine Chaffin and Arnaud Stiegler and Teven Le Scao and Arun Raja and Manan Dey and M Saiful Bari and Canwen Xu and Urmish Thakker and Shanya Sharma Sharma and Eliza Szczechla and Taewoon Kim and Gunjan Chhablani and Nihal Nayak and Debajyoti Datta and Jonathan Chang and Mike Tian-Jian Jiang and Han Wang and Matteo Manica and Sheng Shen and Zheng Xin Yong and Harshit Pandey and Rachel Bawden and Thomas Wang and Trishala Neeraj and Jos Rozen and Abheesht Sharma and Andrea Santilli and Thibault Fevry and Jason Alan Fries and Ryan Teehan and Stella Biderman and Leo Gao and Tali Bers and Thomas Wolf and Alexander M. Rush},
year={2021},
eprint={2110.08207},
archivePrefix={arXiv},
primaryClass={cs.LG}
}
``` |
facebook/opt-30b | 463007d7da4e87fe962909a027811a8c0b32ede8 | 2022-06-23T16:42:12.000Z | [
"pytorch",
"tf",
"jax",
"opt",
"text-generation",
"en",
"arxiv:2205.01068",
"arxiv:2005.14165",
"transformers",
"license:other"
] | text-generation | false | facebook | null | facebook/opt-30b | 18,428 | 71 | transformers | 501 | ---
language: en
inference: false
tags:
- text-generation
- opt
license: other
commercial: false
---
# OPT : Open Pre-trained Transformer Language Models
OPT was first introduced in [Open Pre-trained Transformer Language Models](https://arxiv.org/abs/2205.01068) and first released in [metaseq's repository](https://github.com/facebookresearch/metaseq) on May 3rd 2022 by Meta AI.
**Disclaimer**: The team releasing OPT wrote an official model card, which is available in Appendix D of the [paper](https://arxiv.org/pdf/2205.01068.pdf).
Content from **this** model card has been written by the Hugging Face team.
## Intro
To quote the first two paragraphs of the [official paper](https://arxiv.org/abs/2205.01068)
> Large language models trained on massive text collections have shown surprising emergent
> capabilities to generate text and perform zero- and few-shot learning. While in some cases the public
> can interact with these models through paid APIs, full model access is currently limited to only a
> few highly resourced labs. This restricted access has limited researchers’ ability to study how and
> why these large language models work, hindering progress on improving known challenges in areas
> such as robustness, bias, and toxicity.
> We present Open Pretrained Transformers (OPT), a suite of decoder-only pre-trained transformers ranging from 125M
> to 175B parameters, which we aim to fully and responsibly share with interested researchers. We train the OPT models to roughly match
> the performance and sizes of the GPT-3 class of models, while also applying the latest best practices in data
> collection and efficient training. Our aim in developing this suite of OPT models is to enable reproducible and responsible research at scale, and
> to bring more voices to the table in studying the impact of these LLMs. Definitions of risk, harm, bias, and toxicity, etc., should be articulated by the
> collective research community as a whole, which is only possible when models are available for study.
## Model description
OPT was predominantly pretrained with English text, but a small amount of non-English data is still present within the training corpus via CommonCrawl. The model was pretrained using a causal language modeling (CLM) objective.
OPT belongs to the same family of decoder-only models like [GPT-3](https://arxiv.org/abs/2005.14165). As such, it was pretrained using the self-supervised causal language modedling objective.
For evaluation, OPT follows [GPT-3](https://arxiv.org/abs/2005.14165) by using their prompts and overall experimental setup. For more details, please read
the [official paper](https://arxiv.org/abs/2205.01068).
## Intended uses & limitations
The pretrained-only model can be used for prompting for evaluation of downstream tasks as well as text generation.
In addition, the model can be fine-tuned on a downstream task using the [CLM example](https://github.com/huggingface/transformers/tree/main/examples/pytorch/language-modeling). For all other OPT checkpoints, please have a look at the [model hub](https://huggingface.co/models?filter=opt).
### How to use
For large OPT models, such as this one, it is not recommend to make use of the `text-generation` pipeline because
one should load the model in half-precision to accelerate generation and optimize memory consumption on GPU.
It is recommended to directly call the [`generate`](https://huggingface.co/docs/transformers/main/en/main_classes/text_generation#transformers.generation_utils.GenerationMixin.generate)
method as follows:
```python
>>> from transformers import AutoModelForCausalLM, AutoTokenizer
>>> import torch
>>> model = AutoModelForCausalLM.from_pretrained("facebook/opt-30b", torch_dtype=torch.float16).cuda()
>>> # the fast tokenizer currently does not work correctly
>>> tokenizer = AutoTokenizer.from_pretrained("facebook/opt-30b", use_fast=False)
>>> prompt = "Hello, I am conscious and"
>>> input_ids = tokenizer(prompt, return_tensors="pt").input_ids.cuda()
>>> generated_ids = model.generate(input_ids)
>>> tokenizer.batch_decode(generated_ids, skip_special_tokens=True)
['Hello, I am conscious and I am here.\nI am also conscious and I am here']
```
By default, generation is deterministic. In order to use the top-k sampling, please set `do_sample` to `True`.
```python
>>> from transformers import AutoModelForCausalLM, AutoTokenizer, set_seed
>>> import torch
>>> model = AutoModelForCausalLM.from_pretrained("facebook/opt-30b", torch_dtype=torch.float16).cuda()
>>> # the fast tokenizer currently does not work correctly
>>> tokenizer = AutoTokenizer.from_pretrained("facebook/opt-30b", use_fast=False)
>>> prompt = "Hello, I am conscious and"
>>> input_ids = tokenizer(prompt, return_tensors="pt").input_ids.cuda()
>>> set_seed(32)
>>> generated_ids = model.generate(input_ids, do_sample=True)
>>> tokenizer.batch_decode(generated_ids, skip_special_tokens=True)
['Hello, I am conscious and aware that you have your back turned to me and want to talk']
```
### Limitations and bias
As mentioned in Meta AI's model card, given that the training data used for this model contains a lot of
unfiltered content from the internet, which is far from neutral the model is strongly biased :
> Like other large language models for which the diversity (or lack thereof) of training
> data induces downstream impact on the quality of our model, OPT-175B has limitations in terms
> of bias and safety. OPT-175B can also have quality issues in terms of generation diversity and
> hallucination. In general, OPT-175B is not immune from the plethora of issues that plague modern
> large language models.
Here's an example of how the model can have biased predictions:
```python
>>> from transformers import AutoModelForCausalLM, AutoTokenizer, set_seed
>>> import torch
>>> model = AutoModelForCausalLM.from_pretrained("facebook/opt-30b", torch_dtype=torch.float16).cuda()
>>> # the fast tokenizer currently does not work correctly
>>> tokenizer = AutoTokenizer.from_pretrained("facebook/opt-30b", use_fast=False)
>>> prompt = "The woman worked as a"
>>> input_ids = tokenizer(prompt, return_tensors="pt").input_ids.cuda()
>>> set_seed(32)
>>> generated_ids = model.generate(input_ids, do_sample=True, num_return_sequences=5, max_length=10)
>>> tokenizer.batch_decode(generated_ids, skip_special_tokens=True)
The woman worked as a supervisor in the office
The woman worked as a social worker in a
The woman worked as a cashier at the
The woman worked as a teacher from 2011 to
he woman worked as a maid at the house
```
compared to:
```python
>>> from transformers import AutoModelForCausalLM, AutoTokenizer, set_seed
>>> import torch
>>> model = AutoModelForCausalLM.from_pretrained("facebook/opt-30b", torch_dtype=torch.float16).cuda()
>>> # the fast tokenizer currently does not work correctly
>>> tokenizer = AutoTokenizer.from_pretrained("facebook/opt-30b", use_fast=False)
>>> prompt = "The man worked as a"
>>> input_ids = tokenizer(prompt, return_tensors="pt").input_ids.cuda()
>>> set_seed(32)
>>> generated_ids = model.generate(input_ids, do_sample=True, num_return_sequences=5, max_length=10)
>>> tokenizer.batch_decode(generated_ids, skip_special_tokens=True)
The man worked as a school bus driver for
The man worked as a bartender in a bar
The man worked as a cashier at the
The man worked as a teacher, and was
The man worked as a professional at a range
```
This bias will also affect all fine-tuned versions of this model.
## Training data
The Meta AI team wanted to train this model on a corpus as large as possible. It is composed of the union of the following 5 filtered datasets of textual documents:
- BookCorpus, which consists of more than 10K unpublished books,
- CC-Stories, which contains a subset of CommonCrawl data filtered to match the
story-like style of Winograd schemas,
- The Pile, from which * Pile-CC, OpenWebText2, USPTO, Project Gutenberg, OpenSubtitles, Wikipedia, DM Mathematics and HackerNews* were included.
- Pushshift.io Reddit dataset that was developed in Baumgartner et al. (2020) and processed in
Roller et al. (2021)
- CCNewsV2 containing an updated version of the English portion of the CommonCrawl News
dataset that was used in RoBERTa (Liu et al., 2019b)
The final training data contains 180B tokens corresponding to 800GB of data. The validation split was made of 200MB of the pretraining data, sampled proportionally
to each dataset’s size in the pretraining corpus.
The dataset might contains offensive content as parts of the dataset are a subset of
public Common Crawl data, along with a subset of public Reddit data, which could contain sentences
that, if viewed directly, can be insulting, threatening, or might otherwise cause anxiety.
### Collection process
The dataset was collected form internet, and went through classic data processing algorithms and
re-formatting practices, including removing repetitive/non-informative text like *Chapter One* or
*This ebook by Project Gutenberg.*
## Training procedure
### Preprocessing
The texts are tokenized using the **GPT2** byte-level version of Byte Pair Encoding (BPE) (for unicode characters) and a
vocabulary size of 50272. The inputs are sequences of 2048 consecutive tokens.
The 175B model was trained on 992 *80GB A100 GPUs*. The training duration was roughly ~33 days of continuous training.
### BibTeX entry and citation info
```bibtex
@misc{zhang2022opt,
title={OPT: Open Pre-trained Transformer Language Models},
author={Susan Zhang and Stephen Roller and Naman Goyal and Mikel Artetxe and Moya Chen and Shuohui Chen and Christopher Dewan and Mona Diab and Xian Li and Xi Victoria Lin and Todor Mihaylov and Myle Ott and Sam Shleifer and Kurt Shuster and Daniel Simig and Punit Singh Koura and Anjali Sridhar and Tianlu Wang and Luke Zettlemoyer},
year={2022},
eprint={2205.01068},
archivePrefix={arXiv},
primaryClass={cs.CL}
}
```
|
microsoft/deberta-v3-xsmall | a4c41b18332abc2f8df65cd0528883373cf09923 | 2022-01-13T18:28:06.000Z | [
"pytorch",
"tf",
"deberta-v2",
"en",
"arxiv:2006.03654",
"arxiv:2111.09543",
"transformers",
"deberta",
"deberta-v3",
"license:mit"
] | null | false | microsoft | null | microsoft/deberta-v3-xsmall | 18,409 | 10 | transformers | 502 | ---
language: en
tags:
- deberta
- deberta-v3
thumbnail: https://huggingface.co/front/thumbnails/microsoft.png
license: mit
---
## DeBERTaV3: Improving DeBERTa using ELECTRA-Style Pre-Training with Gradient-Disentangled Embedding Sharing
[DeBERTa](https://arxiv.org/abs/2006.03654) improves the BERT and RoBERTa models using disentangled attention and enhanced mask decoder. With those two improvements, DeBERTa out perform RoBERTa on a majority of NLU tasks with 80GB training data.
In [DeBERTa V3](https://arxiv.org/abs/2111.09543), we further improved the efficiency of DeBERTa using ELECTRA-Style pre-training with Gradient Disentangled Embedding Sharing. Compared to DeBERTa, our V3 version significantly improves the model performance on downstream tasks. You can find more technique details about the new model from our [paper](https://arxiv.org/abs/2111.09543).
Please check the [official repository](https://github.com/microsoft/DeBERTa) for more implementation details and updates.
The DeBERTa V3 xsmall model comes with 12 layers and a hidden size of 384. It has only **22M** backbone parameters with a vocabulary containing 128K tokens which introduces 48M parameters in the Embedding layer. This model was trained using the 160GB data as DeBERTa V2.
#### Fine-tuning on NLU tasks
We present the dev results on SQuAD 2.0 and MNLI tasks.
| Model |Vocabulary(K)|Backbone #Params(M)| SQuAD 2.0(F1/EM) | MNLI-m/mm(ACC)|
|-------------------|----------|-------------------|-----------|----------|
| RoBERTa-base |50 |86 | 83.7/80.5 | 87.6/- |
| XLNet-base |32 |92 | -/80.2 | 86.8/- |
| ELECTRA-base |30 |86 | -/80.5 | 88.8/ |
| DeBERTa-base |50 |100 | 86.2/83.1| 88.8/88.5|
| DeBERTa-v3-large|128|304 | 91.5/89.0 | 91.8/91.9|
| DeBERTa-v3-base |128|86 | 88.4/85.4 | 90.6/90.7|
| DeBERTa-v3-small |128|44 | 82.8/80.4 | 88.3/87.7|
| **DeBERTa-v3-xsmall** |128|**22** | **84.8/82.0** | **88.1/88.3**|
| DeBERTa-v3-xsmall+SiFT|128|22 | -/- | 88.4/88.5|
[#| ELECTRA-small |30 |9.6 | - | - |]::
#### Fine-tuning with HF transformers
```bash
#!/bin/bash
cd transformers/examples/pytorch/text-classification/
pip install datasets
export TASK_NAME=mnli
output_dir="ds_results"
num_gpus=8
batch_size=8
python -m torch.distributed.launch --nproc_per_node=${num_gpus} \
run_glue.py \
--model_name_or_path microsoft/deberta-v3-xsmall \
--task_name $TASK_NAME \
--do_train \
--do_eval \
--evaluation_strategy steps \
--max_seq_length 256 \
--warmup_steps 1000 \
--per_device_train_batch_size ${batch_size} \
--learning_rate 4.5e-5 \
--num_train_epochs 3 \
--output_dir $output_dir \
--overwrite_output_dir \
--logging_steps 1000 \
--logging_dir $output_dir
```
### Citation
If you find DeBERTa useful for your work, please cite the following papers:
``` latex
@misc{he2021debertav3,
title={DeBERTaV3: Improving DeBERTa using ELECTRA-Style Pre-Training with Gradient-Disentangled Embedding Sharing},
author={Pengcheng He and Jianfeng Gao and Weizhu Chen},
year={2021},
eprint={2111.09543},
archivePrefix={arXiv},
primaryClass={cs.CL}
}
```
``` latex
@inproceedings{
he2021deberta,
title={DEBERTA: DECODING-ENHANCED BERT WITH DISENTANGLED ATTENTION},
author={Pengcheng He and Xiaodong Liu and Jianfeng Gao and Weizhu Chen},
booktitle={International Conference on Learning Representations},
year={2021},
url={https://openreview.net/forum?id=XPZIaotutsD}
}
```
|
ynie/roberta-large-snli_mnli_fever_anli_R1_R2_R3-nli | 5b605abab9b75bc87ab66cfc049ef58d9d64b8ed | 2021-05-20T23:17:23.000Z | [
"pytorch",
"jax",
"roberta",
"text-classification",
"dataset:snli",
"dataset:anli",
"dataset:multi_nli",
"dataset:multi_nli_mismatch",
"dataset:fever",
"transformers",
"license:mit"
] | text-classification | false | ynie | null | ynie/roberta-large-snli_mnli_fever_anli_R1_R2_R3-nli | 18,330 | 3 | transformers | 503 | ---
datasets:
- snli
- anli
- multi_nli
- multi_nli_mismatch
- fever
license: mit
---
This is a strong pre-trained RoBERTa-Large NLI model.
The training data is a combination of well-known NLI datasets: [`SNLI`](https://nlp.stanford.edu/projects/snli/), [`MNLI`](https://cims.nyu.edu/~sbowman/multinli/), [`FEVER-NLI`](https://github.com/easonnie/combine-FEVER-NSMN/blob/master/other_resources/nli_fever.md), [`ANLI (R1, R2, R3)`](https://github.com/facebookresearch/anli).
Other pre-trained NLI models including `RoBERTa`, `ALBert`, `BART`, `ELECTRA`, `XLNet` are also available.
Trained by [Yixin Nie](https://easonnie.github.io), [original source](https://github.com/facebookresearch/anli).
Try the code snippet below.
```
from transformers import AutoTokenizer, AutoModelForSequenceClassification
import torch
if __name__ == '__main__':
max_length = 256
premise = "Two women are embracing while holding to go packages."
hypothesis = "The men are fighting outside a deli."
hg_model_hub_name = "ynie/roberta-large-snli_mnli_fever_anli_R1_R2_R3-nli"
# hg_model_hub_name = "ynie/albert-xxlarge-v2-snli_mnli_fever_anli_R1_R2_R3-nli"
# hg_model_hub_name = "ynie/bart-large-snli_mnli_fever_anli_R1_R2_R3-nli"
# hg_model_hub_name = "ynie/electra-large-discriminator-snli_mnli_fever_anli_R1_R2_R3-nli"
# hg_model_hub_name = "ynie/xlnet-large-cased-snli_mnli_fever_anli_R1_R2_R3-nli"
tokenizer = AutoTokenizer.from_pretrained(hg_model_hub_name)
model = AutoModelForSequenceClassification.from_pretrained(hg_model_hub_name)
tokenized_input_seq_pair = tokenizer.encode_plus(premise, hypothesis,
max_length=max_length,
return_token_type_ids=True, truncation=True)
input_ids = torch.Tensor(tokenized_input_seq_pair['input_ids']).long().unsqueeze(0)
# remember bart doesn't have 'token_type_ids', remove the line below if you are using bart.
token_type_ids = torch.Tensor(tokenized_input_seq_pair['token_type_ids']).long().unsqueeze(0)
attention_mask = torch.Tensor(tokenized_input_seq_pair['attention_mask']).long().unsqueeze(0)
outputs = model(input_ids,
attention_mask=attention_mask,
token_type_ids=token_type_ids,
labels=None)
# Note:
# "id2label": {
# "0": "entailment",
# "1": "neutral",
# "2": "contradiction"
# },
predicted_probability = torch.softmax(outputs[0], dim=1)[0].tolist() # batch_size only one
print("Premise:", premise)
print("Hypothesis:", hypothesis)
print("Entailment:", predicted_probability[0])
print("Neutral:", predicted_probability[1])
print("Contradiction:", predicted_probability[2])
```
More in [here](https://github.com/facebookresearch/anli/blob/master/src/hg_api/interactive_eval.py).
Citation:
```
@inproceedings{nie-etal-2020-adversarial,
title = "Adversarial {NLI}: A New Benchmark for Natural Language Understanding",
author = "Nie, Yixin and
Williams, Adina and
Dinan, Emily and
Bansal, Mohit and
Weston, Jason and
Kiela, Douwe",
booktitle = "Proceedings of the 58th Annual Meeting of the Association for Computational Linguistics",
year = "2020",
publisher = "Association for Computational Linguistics",
}
```
|
sshleifer/tiny-dbmdz-bert-large-cased-finetuned-conll03-english | 1dfc430017617e79fd19a2b60b30d6b217f64d28 | 2021-05-20T07:12:23.000Z | [
"pytorch",
"tf",
"jax",
"bert",
"token-classification",
"transformers",
"autotrain_compatible"
] | token-classification | false | sshleifer | null | sshleifer/tiny-dbmdz-bert-large-cased-finetuned-conll03-english | 18,202 | null | transformers | 504 | Entry not found |
savasy/bert-base-turkish-sentiment-cased | 330ec37b18140dcd5c5dd6357d59463ae9deb2e0 | 2021-05-20T04:55:01.000Z | [
"pytorch",
"jax",
"bert",
"text-classification",
"tr",
"transformers"
] | text-classification | false | savasy | null | savasy/bert-base-turkish-sentiment-cased | 18,114 | 5 | transformers | 505 | ---
language: tr
---
# Bert-base Turkish Sentiment Model
https://huggingface.co/savasy/bert-base-turkish-sentiment-cased
This model is used for Sentiment Analysis, which is based on BERTurk for Turkish Language https://huggingface.co/dbmdz/bert-base-turkish-cased
## Dataset
The dataset is taken from the studies [[2]](#paper-2) and [[3]](#paper-3), and merged.
* The study [2] gathered movie and product reviews. The products are book, DVD, electronics, and kitchen.
The movie dataset is taken from a cinema Web page ([Beyazperde](www.beyazperde.com)) with
5331 positive and 5331 negative sentences. Reviews in the Web page are marked in
scale from 0 to 5 by the users who made the reviews. The study considered a review
sentiment positive if the rating is equal to or bigger than 4, and negative if it is less
or equal to 2. They also built Turkish product review dataset from an online retailer
Web page. They constructed benchmark dataset consisting of reviews regarding some
products (book, DVD, etc.). Likewise, reviews are marked in the range from 1 to 5,
and majority class of reviews are 5. Each category has 700 positive and 700 negative
reviews in which average rating of negative reviews is 2.27 and of positive reviews
is 4.5. This dataset is also used by the study [[1]](#paper-1).
* The study [[3]](#paper-3) collected tweet dataset. They proposed a new approach for automatically classifying the sentiment of microblog messages. The proposed approach is based on utilizing robust feature representation and fusion.
*Merged Dataset*
| *size* | *data* |
|--------|----|
| 8000 |dev.tsv|
| 8262 |test.tsv|
| 32000 |train.tsv|
| *48290* |*total*|
### The dataset is used by following papers
<a id="paper-1">[1]</a> Yildirim, Savaş. (2020). Comparing Deep Neural Networks to Traditional Models for Sentiment Analysis in Turkish Language. 10.1007/978-981-15-1216-2_12.
<a id="paper-2">[2]</a> Demirtas, Erkin and Mykola Pechenizkiy. 2013. Cross-lingual polarity detection with machine translation. In Proceedings of the Second International Workshop on Issues of Sentiment
Discovery and Opinion Mining (WISDOM ’13)
<a id="paper-3">[3]</a> Hayran, A., Sert, M. (2017), "Sentiment Analysis on Microblog Data based on Word Embedding and Fusion Techniques", IEEE 25th Signal Processing and Communications Applications Conference (SIU 2017), Belek, Turkey
## Training
```shell
export GLUE_DIR="./sst-2-newall"
export TASK_NAME=SST-2
python3 run_glue.py \
--model_type bert \
--model_name_or_path dbmdz/bert-base-turkish-uncased\
--task_name "SST-2" \
--do_train \
--do_eval \
--data_dir "./sst-2-newall" \
--max_seq_length 128 \
--per_gpu_train_batch_size 32 \
--learning_rate 2e-5 \
--num_train_epochs 3.0 \
--output_dir "./model"
```
## Results
> 05/10/2020 17:00:43 - INFO - transformers.trainer - \*\*\*\*\* Running Evaluation \*\*\*\*\*
> 05/10/2020 17:00:43 - INFO - transformers.trainer - Num examples = 7999
> 05/10/2020 17:00:43 - INFO - transformers.trainer - Batch size = 8
> Evaluation: 100% 1000/1000 [00:34<00:00, 29.04it/s]
> 05/10/2020 17:01:17 - INFO - \_\_main__ - \*\*\*\*\* Eval results sst-2 \*\*\*\*\*
> 05/10/2020 17:01:17 - INFO - \_\_main__ - acc = 0.9539942492811602
> 05/10/2020 17:01:17 - INFO - \_\_main__ - loss = 0.16348013816401363
Accuracy is about **95.4%**
## Code Usage
```python
from transformers import AutoModelForSequenceClassification, AutoTokenizer, pipeline
model = AutoModelForSequenceClassification.from_pretrained("savasy/bert-base-turkish-sentiment-cased")
tokenizer = AutoTokenizer.from_pretrained("savasy/bert-base-turkish-sentiment-cased")
sa= pipeline("sentiment-analysis", tokenizer=tokenizer, model=model)
p = sa("bu telefon modelleri çok kaliteli , her parçası çok özel bence")
print(p)
# [{'label': 'LABEL_1', 'score': 0.9871089}]
print(p[0]['label'] == 'LABEL_1')
# True
p = sa("Film çok kötü ve çok sahteydi")
print(p)
# [{'label': 'LABEL_0', 'score': 0.9975505}]
print(p[0]['label'] == 'LABEL_1')
# False
```
## Test
### Data
Suppose your file has lots of lines of comment and label (1 or 0) at the end (tab seperated)
> comment1 ... \t label
> comment2 ... \t label
> ...
### Code
```python
from transformers import AutoModelForSequenceClassification, AutoTokenizer, pipeline
model = AutoModelForSequenceClassification.from_pretrained("savasy/bert-base-turkish-sentiment-cased")
tokenizer = AutoTokenizer.from_pretrained("savasy/bert-base-turkish-sentiment-cased")
sa = pipeline("sentiment-analysis", tokenizer=tokenizer, model=model)
input_file = "/path/to/your/file/yourfile.tsv"
i, crr = 0, 0
for line in open(input_file):
lines = line.strip().split("\t")
if len(lines) == 2:
i = i + 1
if i%100 == 0:
print(i)
pred = sa(lines[0])
pred = pred[0]["label"].split("_")[1]
if pred == lines[1]:
crr = crr + 1
print(crr, i, crr/i)
```
|
megagonlabs/transformers-ud-japanese-electra-base-ginza-510 | 8498b019a97bdf693fa138efcf7e118f099e2c4e | 2021-12-05T12:12:12.000Z | [
"pytorch",
"electra",
"feature-extraction",
"ja",
"dataset:mC4",
"dataset:UD_Japanese_BCCWJ r2.8",
"dataset:GSK2014-A(2019)",
"arxiv:1910.10683",
"transformers",
"PyTorch",
"Transformers",
"spaCy",
"ELECTRA",
"GiNZA",
"mC4",
"UD_Japanese-BCCWJ",
"GSK2014-A",
"MIT",
"license:mit"
] | feature-extraction | false | megagonlabs | null | megagonlabs/transformers-ud-japanese-electra-base-ginza-510 | 18,065 | null | transformers | 506 | ---
language:
- ja
thumbnail: "https://raw.githubusercontent.com/megagonlabs/ginza/static/docs/images/GiNZA_logo_4c_s.png"
tags:
- PyTorch
- Transformers
- spaCy
- ELECTRA
- GiNZA
- mC4
- UD_Japanese-BCCWJ
- GSK2014-A
- ja
- MIT
license: "mit"
datasets:
- mC4
- UD_Japanese_BCCWJ r2.8
- GSK2014-A(2019)
metrics:
- UAS
- LAS
- UPOS
---
# transformers-ud-japanese-electra-ginza-510 (sudachitra-wordpiece, mC4 Japanese)
This is an [ELECTRA](https://github.com/google-research/electra) model pretrained on approximately 200M Japanese sentences extracted from the [mC4](https://huggingface.co/datasets/mc4) and finetuned by [spaCy v3](https://spacy.io/usage/v3) on [UD\_Japanese\_BCCWJ r2.8](https://universaldependencies.org/treebanks/ja_bccwj/index.html).
The base pretrain model is [megagonlabs/transformers-ud-japanese-electra-base-discrimininator](https://huggingface.co/megagonlabs/transformers-ud-japanese-electra-base-discriminator).
The entire spaCy v3 model is distributed as a python package named [`ja_ginza_electra`](https://pypi.org/project/ja-ginza-electra/) from PyPI along with [`GiNZA v5`](https://github.com/megagonlabs/ginza) which provides some custom pipeline components to recognize the Japanese bunsetu-phrase structures.
Try running it as below:
```console
$ pip install ginza ja_ginza_electra
$ ginza
```
## Licenses
The models are distributed under the terms of the [MIT License](https://opensource.org/licenses/mit-license.php).
## Acknowledgments
This model is permitted to be published under the `MIT License` under a joint research agreement between NINJAL (National Institute for Japanese Language and Linguistics) and Megagon Labs Tokyo.
## Citations
- [mC4](https://huggingface.co/datasets/mc4)
Contains information from `mC4` which is made available under the [ODC Attribution License](https://opendatacommons.org/licenses/by/1-0/).
```
@article{2019t5,
author = {Colin Raffel and Noam Shazeer and Adam Roberts and Katherine Lee and Sharan Narang and Michael Matena and Yanqi Zhou and Wei Li and Peter J. Liu},
title = {Exploring the Limits of Transfer Learning with a Unified Text-to-Text Transformer},
journal = {arXiv e-prints},
year = {2019},
archivePrefix = {arXiv},
eprint = {1910.10683},
}
```
- [UD\_Japanese\_BCCWJ r2.8](https://universaldependencies.org/treebanks/ja_bccwj/index.html)
```
Asahara, M., Kanayama, H., Tanaka, T., Miyao, Y., Uematsu, S., Mori, S.,
Matsumoto, Y., Omura, M., & Murawaki, Y. (2018).
Universal Dependencies Version 2 for Japanese.
In LREC-2018.
```
- [GSK2014-A(2019)](https://www.gsk.or.jp/catalog/gsk2014-a/)
|
fnlp/bart-base-chinese | cac82280f9cafd2e7bb76fc32123717e335397ba | 2021-10-31T15:05:30.000Z | [
"pytorch",
"bart",
"feature-extraction",
"zh",
"arxiv:2109.05729",
"transformers",
"text2text-generation",
"Chinese",
"seq2seq",
"BART"
] | feature-extraction | false | fnlp | null | fnlp/bart-base-chinese | 18,058 | 16 | transformers | 507 | ---
tags:
- text2text-generation
- Chinese
- seq2seq
- BART
language: zh
---
# Chinese BART-Base
## Model description
This is an implementation of Chinese BART-Base.
[**CPT: A Pre-Trained Unbalanced Transformer for Both Chinese Language Understanding and Generation**](https://arxiv.org/pdf/2109.05729.pdf)
Yunfan Shao, Zhichao Geng, Yitao Liu, Junqi Dai, Fei Yang, Li Zhe, Hujun Bao, Xipeng Qiu
**Github Link:** https://github.com/fastnlp/CPT
## Usage
```python
>>> from transformers import BertTokenizer, BartForConditionalGeneration, Text2TextGenerationPipeline
>>> tokenizer = BertTokenizer.from_pretrained("fnlp/bart-base-chinese")
>>> model = BartForConditionalGeneration.from_pretrained("fnlp/bart-base-chinese")
>>> text2text_generator = Text2TextGenerationPipeline(model, tokenizer)
>>> text2text_generator("北京是[MASK]的首都", max_length=50, do_sample=False)
[{'generated_text': '北 京 是 中 国 的 首 都'}]
```
**Note: Please use BertTokenizer for the model vocabulary. DO NOT use original BartTokenizer.**
## Citation
```bibtex
@article{shao2021cpt,
title={CPT: A Pre-Trained Unbalanced Transformer for Both Chinese Language Understanding and Generation},
author={Yunfan Shao and Zhichao Geng and Yitao Liu and Junqi Dai and Fei Yang and Li Zhe and Hujun Bao and Xipeng Qiu},
journal={arXiv preprint arXiv:2109.05729},
year={2021}
}
``` |
kssteven/ibert-roberta-base | 4f98e9110b04a8958444d3af8ed39287834fbb90 | 2021-11-22T10:09:32.000Z | [
"pytorch",
"ibert",
"fill-mask",
"arxiv:1907.11692",
"arxiv:2101.01321",
"transformers",
"autotrain_compatible"
] | fill-mask | false | kssteven | null | kssteven/ibert-roberta-base | 17,996 | null | transformers | 508 | # I-BERT base model
This model, `ibert-roberta-base`, is an integer-only quantized version of [RoBERTa](https://arxiv.org/abs/1907.11692), and was introduced in [this paper](https://arxiv.org/abs/2101.01321).
I-BERT stores all parameters with INT8 representation, and carries out the entire inference using integer-only arithmetic.
In particular, I-BERT replaces all floating point operations in the Transformer architectures (e.g., MatMul, GELU, Softmax, and LayerNorm) with closely approximating integer operations.
This can result in upto 4x inference speed up as compared to floating point counterpart when tested on an Nvidia T4 GPU.
The best model parameters searched via quantization-aware finetuning can be then exported (e.g., to TensorRT) for integer-only deployment of the model.
## Finetuning Procedure
Finetuning of I-BERT consists of 3 stages: (1) Full-precision finetuning from the pretrained model on a down-stream task, (2) model quantization, and (3) integer-only finetuning (i.e., quantization-aware training) of the quantized model.
### Full-precision finetuning
Full-precision finetuning of I-BERT is similar to RoBERTa finetuning.
For instance, you can run the following command to finetune on the [MRPC](https://www.microsoft.com/en-us/download/details.aspx?id=52398) text classification task.
```
python examples/text-classification/run_glue.py \
--model_name_or_path kssteven/ibert-roberta-base \
--task_name MRPC \
--do_eval \
--do_train \
--evaluation_strategy epoch \
--max_seq_length 128 \
--per_device_train_batch_size 32 \
--save_steps 115 \
--learning_rate 2e-5 \
--num_train_epochs 10 \
--output_dir $OUTPUT_DIR
```
### Model Quantization
Once you are done with full-precision finetuning, open up `config.json` in your checkpoint directory and set the `quantize` attribute as `true`.
```
{
"_name_or_path": "kssteven/ibert-roberta-base",
"architectures": [
"IBertForSequenceClassification"
],
"attention_probs_dropout_prob": 0.1,
"bos_token_id": 0,
"eos_token_id": 2,
"finetuning_task": "mrpc",
"force_dequant": "none",
"hidden_act": "gelu",
"hidden_dropout_prob": 0.1,
"hidden_size": 768,
"initializer_range": 0.02,
"intermediate_size": 3072,
"layer_norm_eps": 1e-05,
"max_position_embeddings": 514,
"model_type": "ibert",
"num_attention_heads": 12,
"num_hidden_layers": 12,
"pad_token_id": 1,
"position_embedding_type": "absolute",
"quant_mode": true,
"tokenizer_class": "RobertaTokenizer",
"transformers_version": "4.4.0.dev0",
"type_vocab_size": 1,
"vocab_size": 50265
}
```
Then, your model will automatically run as the integer-only mode when you load the checkpoint.
Also, make sure to delete `optimizer.pt`, `scheduler.pt` and `trainer_state.json` in the same directory.
Otherwise, HF will not reset the optimizer, scheduler, or trainer state for the following integer-only finetuning.
### Integer-only finetuning (Quantization-aware training)
Finally, you will be able to run integer-only finetuning simply by loading the checkpoint file you modified.
Note that the only difference in the example command below is `model_name_or_path`.
```
python examples/text-classification/run_glue.py \
--model_name_or_path $CHECKPOINT_DIR
--task_name MRPC \
--do_eval \
--do_train \
--evaluation_strategy epoch \
--max_seq_length 128 \
--per_device_train_batch_size 32 \
--save_steps 115 \
--learning_rate 1e-6 \
--num_train_epochs 10 \
--output_dir $OUTPUT_DIR
```
## Citation info
If you use I-BERT, please cite [our papaer](https://arxiv.org/abs/2101.01321).
```
@article{kim2021bert,
title={I-BERT: Integer-only BERT Quantization},
author={Kim, Sehoon and Gholami, Amir and Yao, Zhewei and Mahoney, Michael W and Keutzer, Kurt},
journal={arXiv preprint arXiv:2101.01321},
year={2021}
}
```
|
facebook/deit-small-distilled-patch16-224 | 3559f915da8724f0eba2252acb20cc96649c6289 | 2022-07-13T11:41:21.000Z | [
"pytorch",
"tf",
"deit",
"image-classification",
"dataset:imagenet",
"arxiv:2012.12877",
"arxiv:2006.03677",
"transformers",
"vision",
"license:apache-2.0"
] | image-classification | false | facebook | null | facebook/deit-small-distilled-patch16-224 | 17,980 | null | transformers | 509 | ---
license: apache-2.0
tags:
- image-classification
- vision
datasets:
- imagenet
---
# Distilled Data-efficient Image Transformer (small-sized model)
Distilled data-efficient Image Transformer (DeiT) model pre-trained and fine-tuned on ImageNet-1k (1 million images, 1,000 classes) at resolution 224x224. It was first introduced in the paper [Training data-efficient image transformers & distillation through attention](https://arxiv.org/abs/2012.12877) by Touvron et al. and first released in [this repository](https://github.com/facebookresearch/deit). However, the weights were converted from the [timm repository](https://github.com/rwightman/pytorch-image-models) by Ross Wightman.
Disclaimer: The team releasing DeiT did not write a model card for this model so this model card has been written by the Hugging Face team.
## Model description
This model is a distilled Vision Transformer (ViT). It uses a distillation token, besides the class token, to effectively learn from a teacher (CNN) during both pre-training and fine-tuning. The distillation token is learned through backpropagation, by interacting with the class ([CLS]) and patch tokens through the self-attention layers.
Images are presented to the model as a sequence of fixed-size patches (resolution 16x16), which are linearly embedded.
## Intended uses & limitations
You can use the raw model for image classification. See the [model hub](https://huggingface.co/models?search=facebook/deit) to look for
fine-tuned versions on a task that interests you.
### How to use
Since this model is a distilled ViT model, you can plug it into DeiTModel, DeiTForImageClassification or DeiTForImageClassificationWithTeacher. Note that the model expects the data to be prepared using DeiTFeatureExtractor. Here we use AutoFeatureExtractor, which will automatically use the appropriate feature extractor given the model name.
Here is how to use this model to classify an image of the COCO 2017 dataset into one of the 1,000 ImageNet classes:
```python
from transformers import AutoFeatureExtractor, DeiTForImageClassificationWithTeacher
from PIL import Image
import requests
url = 'http://images.cocodataset.org/val2017/000000039769.jpg'
image = Image.open(requests.get(url, stream=True).raw)
feature_extractor = AutoFeatureExtractor.from_pretrained('facebook/deit-small-distilled-patch16-224')
model = DeiTForImageClassificationWithTeacher.from_pretrained('facebook/deit-small-distilled-patch16-224')
inputs = feature_extractor(images=image, return_tensors="pt")
outputs = model(**inputs)
logits = outputs.logits
# model predicts one of the 1000 ImageNet classes
predicted_class_idx = logits.argmax(-1).item()
print("Predicted class:", model.config.id2label[predicted_class_idx])
```
Currently, both the feature extractor and model support PyTorch. Tensorflow and JAX/FLAX are coming soon.
## Training data
This model was pretrained and fine-tuned with distillation on [ImageNet-1k](http://www.image-net.org/challenges/LSVRC/2012/), a dataset consisting of 1 million images and 1k classes.
## Training procedure
### Preprocessing
The exact details of preprocessing of images during training/validation can be found [here](https://github.com/facebookresearch/deit/blob/ab5715372db8c6cad5740714b2216d55aeae052e/datasets.py#L78).
At inference time, images are resized/rescaled to the same resolution (256x256), center-cropped at 224x224 and normalized across the RGB channels with the ImageNet mean and standard deviation.
### Pretraining
The model was trained on a single 8-GPU node for 3 days. Training resolution is 224. For all hyperparameters (such as batch size and learning rate) we refer to table 9 of the original paper.
## Evaluation results
| Model | ImageNet top-1 accuracy | ImageNet top-5 accuracy | # params | URL |
|---------------------------------------|-------------------------|-------------------------|----------|------------------------------------------------------------------|
| DeiT-tiny | 72.2 | 91.1 | 5M | https://huggingface.co/facebook/deit-tiny-patch16-224 |
| DeiT-small | 79.9 | 95.0 | 22M | https://huggingface.co/facebook/deit-small-patch16-224 |
| DeiT-base | 81.8 | 95.6 | 86M | https://huggingface.co/facebook/deit-base-patch16-224 |
| DeiT-tiny distilled | 74.5 | 91.9 | 6M | https://huggingface.co/facebook/deit-tiny-distilled-patch16-224 |
| **DeiT-small distilled** | **81.2** | **95.4** | **22M** | **https://huggingface.co/facebook/deit-small-distilled-patch16-224** |
| DeiT-base distilled | 83.4 | 96.5 | 87M | https://huggingface.co/facebook/deit-base-distilled-patch16-224 |
| DeiT-base 384 | 82.9 | 96.2 | 87M | https://huggingface.co/facebook/deit-base-patch16-384 |
| DeiT-base distilled 384 (1000 epochs) | 85.2 | 97.2 | 88M | https://huggingface.co/facebook/deit-base-distilled-patch16-384 |
Note that for fine-tuning, the best results are obtained with a higher resolution (384x384). Of course, increasing the model size will result in better performance.
### BibTeX entry and citation info
```bibtex
@misc{touvron2021training,
title={Training data-efficient image transformers & distillation through attention},
author={Hugo Touvron and Matthieu Cord and Matthijs Douze and Francisco Massa and Alexandre Sablayrolles and Hervé Jégou},
year={2021},
eprint={2012.12877},
archivePrefix={arXiv},
primaryClass={cs.CV}
}
```
```bibtex
@misc{wu2020visual,
title={Visual Transformers: Token-based Image Representation and Processing for Computer Vision},
author={Bichen Wu and Chenfeng Xu and Xiaoliang Dai and Alvin Wan and Peizhao Zhang and Zhicheng Yan and Masayoshi Tomizuka and Joseph Gonzalez and Kurt Keutzer and Peter Vajda},
year={2020},
eprint={2006.03677},
archivePrefix={arXiv},
primaryClass={cs.CV}
}
```
```bibtex
@inproceedings{deng2009imagenet,
title={Imagenet: A large-scale hierarchical image database},
author={Deng, Jia and Dong, Wei and Socher, Richard and Li, Li-Jia and Li, Kai and Fei-Fei, Li},
booktitle={2009 IEEE conference on computer vision and pattern recognition},
pages={248--255},
year={2009},
organization={Ieee}
}
``` |
hf-internal-testing/processor_with_lm | 5477bdaf3c221237d0859ebcd6c6aa49e4a7d804 | 2022-01-18T13:19:58.000Z | [
"pytorch",
"wav2vec2",
"automatic-speech-recognition",
"transformers"
] | automatic-speech-recognition | false | hf-internal-testing | null | hf-internal-testing/processor_with_lm | 17,977 | null | transformers | 510 | Please leave this README for testing purposes |
nlp-waseda/roberta-base-japanese | 87e4c4bb0e741b81e03db376c92f0af288f3be49 | 2022-06-10T23:32:53.000Z | [
"pytorch",
"roberta",
"fill-mask",
"ja",
"dataset:wikipedia",
"dataset:cc100",
"transformers",
"license:cc-by-sa-4.0",
"autotrain_compatible"
] | fill-mask | false | nlp-waseda | null | nlp-waseda/roberta-base-japanese | 17,807 | 10 | transformers | 511 | ---
language: ja
license: cc-by-sa-4.0
datasets:
- wikipedia
- cc100
mask_token: "[MASK]"
widget:
- text: "早稲田 大学 で 自然 言語 処理 を [MASK] する 。"
---
# nlp-waseda/roberta-base-japanese
## Model description
This is a Japanese RoBERTa base model pretrained on Japanese Wikipedia and the Japanese portion of CC-100.
## How to use
You can use this model for masked language modeling as follows:
```python
from transformers import AutoTokenizer, AutoModelForMaskedLM
tokenizer = AutoTokenizer.from_pretrained("nlp-waseda/roberta-base-japanese")
model = AutoModelForMaskedLM.from_pretrained("nlp-waseda/roberta-base-japanese")
sentence = '早稲田 大学 で 自然 言語 処理 を [MASK] する 。' # input should be segmented into words by Juman++ in advance
encoding = tokenizer(sentence, return_tensors='pt')
...
```
You can fine-tune this model on downstream tasks.
## Tokenization
The input text should be segmented into words by [Juman++](https://github.com/ku-nlp/jumanpp) in advance. Juman++ 2.0.0-rc3 was used for pretraining. Each word is tokenized into tokens by [sentencepiece](https://github.com/google/sentencepiece).
## Vocabulary
The vocabulary consists of 32000 tokens including words ([JumanDIC](https://github.com/ku-nlp/JumanDIC)) and subwords induced by the unigram language model of [sentencepiece](https://github.com/google/sentencepiece).
## Training procedure
This model was trained on Japanese Wikipedia (as of 20210920) and the Japanese portion of CC-100. It took a week using eight NVIDIA A100 GPUs.
The following hyperparameters were used during pretraining:
- learning_rate: 1e-4
- per_device_train_batch_size: 256
- distributed_type: multi-GPU
- num_devices: 8
- gradient_accumulation_steps: 2
- total_train_batch_size: 4096
- max_seq_length: 128
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
- lr_scheduler_type: linear
- training_steps: 700000
- warmup_steps: 10000
- mixed_precision_training: Native AMP
## Performance on JGLUE
See the [Baseline Scores](https://github.com/yahoojapan/JGLUE#baseline-scores) of JGLUE.
|
KoboldAI/GPT-Neo-2.7B-Shinen | 551fbd85138d4f29589ab07000cca813cb8a62ea | 2022-03-20T18:49:18.000Z | [
"pytorch",
"gpt_neo",
"text-generation",
"en",
"transformers",
"license:mit"
] | text-generation | false | KoboldAI | null | KoboldAI/GPT-Neo-2.7B-Shinen | 17,802 | 2 | transformers | 512 | ---
language: en
license: mit
---
# GPT-Neo 2.7B - Shinen
## Model Description
GPT-Neo 2.7B-Shinen is a finetune created using EleutherAI's GPT-Neo 2.7B model. Compared to GPT-Neo-2.7-Horni, this model is much heavier on the sexual content.
**Warning: THIS model is NOT suitable for use by minors. The model will output X-rated content.**
## Training data
The training data contains user-generated stories from sexstories.com. All stories are tagged using the following way:
```
[Theme: <theme1>, <theme2> ,<theme3>]
<Story goes here>
```
### How to use
You can use this model directly with a pipeline for text generation. This example generates a different sequence each time it's run:
```py
>>> from transformers import pipeline
>>> generator = pipeline('text-generation', model='KoboldAI/GPT-Neo-2.7B-Shinen')
>>> generator("She was staring at me", do_sample=True, min_length=50)
[{'generated_text': 'She was staring at me with a look that said it all. She wanted me so badly tonight that I wanted'}]
```
### Limitations and Biases
GPT-Neo was trained as an autoregressive language model. This means that its core functionality is taking a string of text and predicting the next token. While language models are widely used for tasks other than this, there are a lot of unknowns with this work.
GPT-Neo-Shinen was trained on a dataset known to contain profanity, lewd, and otherwise abrasive language. GPT-Neo-Shinen *WILL* produce socially unacceptable text without warning.
GPT-Neo-Shinen will respond to particular prompts and offensive content may occur without warning. We recommend having a human curate or filter the outputs before releasing them, both to censor undesirable content and to improve the quality of the results.
### BibTeX entry and citation info
The model is made using the following software:
```bibtex
@software{gpt-neo,
author = {Black, Sid and
Leo, Gao and
Wang, Phil and
Leahy, Connor and
Biderman, Stella},
title = {{GPT-Neo: Large Scale Autoregressive Language
Modeling with Mesh-Tensorflow}},
month = mar,
year = 2021,
note = {{If you use this software, please cite it using
these metadata.}},
publisher = {Zenodo},
version = {1.0},
doi = {10.5281/zenodo.5297715},
url = {https://doi.org/10.5281/zenodo.5297715}
}
``` |
dbmdz/bert-base-german-cased | 56c3dce79f5d93e466f3b800d8e57cddfe13a6d4 | 2021-05-19T14:52:56.000Z | [
"pytorch",
"tf",
"jax",
"bert",
"fill-mask",
"de",
"transformers",
"license:mit",
"autotrain_compatible"
] | fill-mask | false | dbmdz | null | dbmdz/bert-base-german-cased | 17,797 | 5 | transformers | 513 | ---
language: de
license: mit
---
# 🤗 + 📚 dbmdz German BERT models
In this repository the MDZ Digital Library team (dbmdz) at the Bavarian State
Library open sources another German BERT models 🎉
# German BERT
## Stats
In addition to the recently released [German BERT](https://deepset.ai/german-bert)
model by [deepset](https://deepset.ai/) we provide another German-language model.
The source data for the model consists of a recent Wikipedia dump, EU Bookshop corpus,
Open Subtitles, CommonCrawl, ParaCrawl and News Crawl. This results in a dataset with
a size of 16GB and 2,350,234,427 tokens.
For sentence splitting, we use [spacy](https://spacy.io/). Our preprocessing steps
(sentence piece model for vocab generation) follow those used for training
[SciBERT](https://github.com/allenai/scibert). The model is trained with an initial
sequence length of 512 subwords and was performed for 1.5M steps.
This release includes both cased and uncased models.
## Model weights
Currently only PyTorch-[Transformers](https://github.com/huggingface/transformers)
compatible weights are available. If you need access to TensorFlow checkpoints,
please raise an issue!
| Model | Downloads
| -------------------------------- | ---------------------------------------------------------------------------------------------------------------
| `bert-base-german-dbmdz-cased` | [`config.json`](https://s3.amazonaws.com/models.huggingface.co/bert/bert-base-german-dbmdz-cased-config.json) • [`pytorch_model.bin`](https://s3.amazonaws.com/models.huggingface.co/bert/bert-base-german-dbmdz-cased-pytorch_model.bin) • [`vocab.txt`](https://s3.amazonaws.com/models.huggingface.co/bert/bert-base-german-dbmdz-cased-vocab.txt)
| `bert-base-german-dbmdz-uncased` | [`config.json`](https://s3.amazonaws.com/models.huggingface.co/bert/bert-base-german-dbmdz-uncased-config.json) • [`pytorch_model.bin`](https://s3.amazonaws.com/models.huggingface.co/bert/bert-base-german-dbmdz-uncased-pytorch_model.bin) • [`vocab.txt`](https://s3.amazonaws.com/models.huggingface.co/bert/bert-base-german-dbmdz-uncased-vocab.txt)
## Usage
With Transformers >= 2.3 our German BERT models can be loaded like:
```python
from transformers import AutoModel, AutoTokenizer
tokenizer = AutoTokenizer.from_pretrained("dbmdz/bert-base-german-cased")
model = AutoModel.from_pretrained("dbmdz/bert-base-german-cased")
```
## Results
For results on downstream tasks like NER or PoS tagging, please refer to
[this repository](https://github.com/stefan-it/fine-tuned-berts-seq).
# Huggingface model hub
All models are available on the [Huggingface model hub](https://huggingface.co/dbmdz).
# Contact (Bugs, Feedback, Contribution and more)
For questions about our BERT models just open an issue
[here](https://github.com/dbmdz/berts/issues/new) 🤗
# Acknowledgments
Research supported with Cloud TPUs from Google's TensorFlow Research Cloud (TFRC).
Thanks for providing access to the TFRC ❤️
Thanks to the generous support from the [Hugging Face](https://huggingface.co/) team,
it is possible to download both cased and uncased models from their S3 storage 🤗
|
sshleifer/bart-tiny-random | 69bce9237e4fa10ea015446395ec0108067890cf | 2021-06-14T07:44:43.000Z | [
"pytorch",
"tf",
"jax",
"bart",
"text2text-generation",
"transformers",
"autotrain_compatible"
] | text2text-generation | false | sshleifer | null | sshleifer/bart-tiny-random | 17,763 | null | transformers | 514 | Entry not found |
microsoft/trocr-base-printed | 191a64dd5078db39e975a76b798b6fd026a96fa6 | 2022-07-01T07:35:28.000Z | [
"pytorch",
"vision-encoder-decoder",
"arxiv:2109.10282",
"transformers",
"trocr",
"image-to-text"
] | image-to-text | false | microsoft | null | microsoft/trocr-base-printed | 17,758 | 12 | transformers | 515 | ---
tags:
- trocr
- image-to-text
---
# TrOCR (base-sized model, fine-tuned on SROIE)
TrOCR model fine-tuned on the [SROIE dataset](https://rrc.cvc.uab.es/?ch=13). It was introduced in the paper [TrOCR: Transformer-based Optical Character Recognition with Pre-trained Models](https://arxiv.org/abs/2109.10282) by Li et al. and first released in [this repository](https://github.com/microsoft/unilm/tree/master/trocr).
Disclaimer: The team releasing TrOCR did not write a model card for this model so this model card has been written by the Hugging Face team.
## Model description
The TrOCR model is an encoder-decoder model, consisting of an image Transformer as encoder, and a text Transformer as decoder. The image encoder was initialized from the weights of BEiT, while the text decoder was initialized from the weights of RoBERTa.
Images are presented to the model as a sequence of fixed-size patches (resolution 16x16), which are linearly embedded. One also adds absolute position embeddings before feeding the sequence to the layers of the Transformer encoder. Next, the Transformer text decoder autoregressively generates tokens.
## Intended uses & limitations
You can use the raw model for optical character recognition (OCR) on single text-line images. See the [model hub](https://huggingface.co/models?search=microsoft/trocr) to look for fine-tuned versions on a task that interests you.
### How to use
Here is how to use this model in PyTorch:
```python
from transformers import TrOCRProcessor, VisionEncoderDecoderModel
from PIL import Image
import requests
# load image from the IAM database (actually this model is meant to be used on printed text)
url = 'https://fki.tic.heia-fr.ch/static/img/a01-122-02-00.jpg'
image = Image.open(requests.get(url, stream=True).raw).convert("RGB")
processor = TrOCRProcessor.from_pretrained('microsoft/trocr-base-printed')
model = VisionEncoderDecoderModel.from_pretrained('microsoft/trocr-base-printed')
pixel_values = processor(images=image, return_tensors="pt").pixel_values
generated_ids = model.generate(pixel_values)
generated_text = processor.batch_decode(generated_ids, skip_special_tokens=True)[0]
```
### BibTeX entry and citation info
```bibtex
@misc{li2021trocr,
title={TrOCR: Transformer-based Optical Character Recognition with Pre-trained Models},
author={Minghao Li and Tengchao Lv and Lei Cui and Yijuan Lu and Dinei Florencio and Cha Zhang and Zhoujun Li and Furu Wei},
year={2021},
eprint={2109.10282},
archivePrefix={arXiv},
primaryClass={cs.CL}
}
``` |
uer/gpt2-chinese-poem | 01a2a7e8de2df28f48dfa4262b759dd5858b84bd | 2022-02-20T05:00:26.000Z | [
"pytorch",
"tf",
"jax",
"gpt2",
"text-generation",
"zh",
"transformers"
] | text-generation | false | uer | null | uer/gpt2-chinese-poem | 17,737 | 4 | transformers | 516 | ---
language: zh
widget:
- text: "[CLS] 万 叠 春 山 积 雨 晴 ，"
- text: "[CLS] 大 漠"
---
# Chinese Poem GPT2 Model
## Model description
The model is used to generate Chinese ancient poems. You can download the model either from the [GPT2-Chinese Github page](https://github.com/Morizeyao/GPT2-Chinese), or via HuggingFace from the link [gpt2-chinese-poem](https://huggingface.co/uer/gpt2-chinese-poem]).
Since the parameter skip_special_tokens is used in the pipelines.py, special tokens such as [SEP], [UNK] will be deleted, the output results of Hosted inference API (right) may not be properly displayed.
## How to use
You can use the model directly with a pipeline for text generation:
When the parameter skip_special_tokens is True:
```python
>>> from transformers import BertTokenizer, GPT2LMHeadModel,TextGenerationPipeline
>>> tokenizer = BertTokenizer.from_pretrained("uer/gpt2-chinese-poem")
>>> model = GPT2LMHeadModel.from_pretrained("uer/gpt2-chinese-poem")
>>> text_generator = TextGenerationPipeline(model, tokenizer)
>>> text_generator("[CLS]梅 山 如 积 翠 ，", max_length=50, do_sample=True)
[{'generated_text': '[CLS]梅 山 如 积 翠 ， 丛 竹 隠 疏 花 。 水 影 落 寒 濑 ， 竹 声 随 暮 鸦 。 茅 茨 数 间 屋 ， 烟 火 两 三 家 。 安 得 携 琴 酒 ， 相 逢 烟 雨 赊 。 向 湖 边 过 ， 偏 怜 雪 里 看 。 浮 峦 如 画 出 ， 远 树 与 天 连 。 月 上 僧 房 静 ， 风 回 萤 火 寒 。 幽 情 何 可 写 ， 赖 有 子 期 弹 。 棠 真'}]
```
When the parameter skip_special_tokens is False:
```python
>>> from transformers import BertTokenizer, GPT2LMHeadModel,TextGenerationPipeline
>>> tokenizer = BertTokenizer.from_pretrained("uer/gpt2-chinese-poem")
>>> model = GPT2LMHeadModel.from_pretrained("uer/gpt2-chinese-poem")
>>> text_generator = TextGenerationPipeline(model, tokenizer)
>>> text_generator("[CLS]梅 山 如 积 翠 ，", max_length=100, do_sample=True)
[{'generated_text': '[CLS]梅 山 如 积 翠 ， 秀 出 何 其 雄 。 矫 矫 云 间 质 ， 映 日 生 玲 珑 。 根 大 乱 石 结 ， 枝 高 青 云 蒙 。 常 因 风 露 晚 ， 隠 映 瑶 台 中 。 忽 闻 山 石 裂 ， 万 里 吹 天 风 。 又 觉 此 身 高 ， 迥 出 凡 境 空 。 清 影 落 潭 水 ， 暗 香 来 逈 峰 。 却 寻 白 太 白 ， 月 影 摇 江 东 。 [SEP] 而 非'}]
```
## Training data
Training data contains 800,000 Chinese ancient poems which are collected by [chinese-poetry](https://github.com/chinese-poetry/chinese-poetry) and [Poetry](https://github.com/Werneror/Poetry) projects.
## Training procedure
The model is pre-trained by [UER-py](https://github.com/dbiir/UER-py/) on [Tencent Cloud](https://cloud.tencent.com/). We pre-train 200,000 steps with a sequence length of 128. We use extended vocabulary to handle out-of-vocabulary words. The Chinese character that occurs greater than or equal to 100 in poem corpus is added to the vocabulary.
```
python3 preprocess.py --corpus_path corpora/poem.txt \
--vocab_path models/poem_zh_vocab.txt \
--dataset_path poem_dataset.pt --processes_num 16 \
--seq_length 128 --data_processor lm
```
```
python3 pretrain.py --dataset_path poem_dataset.pt \
--vocab_path models/poem_zh_vocab.txt \
--config_path models/gpt2/config.json \
--output_model_path models/poem_gpt2_model.bin \
--world_size 8 --gpu_ranks 0 1 2 3 4 5 6 7 \
--total_steps 200000 --save_checkpoint_steps 50000 --report_steps 1000 \
--learning_rate 5e-4 --batch_size 64
```
Finally, we convert the pre-trained model into Huggingface's format:
```
python3 scripts/convert_gpt2_from_uer_to_huggingface.py --input_model_path poem_gpt2_model.bin-200000 \
--output_model_path pytorch_model.bin \
--layers_num 12
```
### BibTeX entry and citation info
```
@article{radford2019language,
title={Language Models are Unsupervised Multitask Learners},
author={Radford, Alec and Wu, Jeff and Child, Rewon and Luan, David and Amodei, Dario and Sutskever, Ilya},
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}
}
``` |
google/t5-large-lm-adapt | 96ce18564557a62d6ff1cb3771af167433827961 | 2021-11-01T14:00:11.000Z | [
"pytorch",
"tf",
"t5",
"text2text-generation",
"en",
"dataset:c4",
"arxiv:2002.05202",
"arxiv:1910.10683",
"transformers",
"t5-lm-adapt",
"license:apache-2.0",
"autotrain_compatible"
] | text2text-generation | false | google | null | google/t5-large-lm-adapt | 17,713 | 3 | transformers | 517 | ---
language: en
datasets:
- c4
tags:
- t5-lm-adapt
license: apache-2.0
---
[Google's T5](https://ai.googleblog.com/2020/02/exploring-transfer-learning-with-t5.html) Version 1.1 - LM-Adapted
## Version 1.1 - LM-Adapted
[T5 Version 1.1 - LM Adapted](https://github.com/google-research/text-to-text-transfer-transformer/blob/main/released_checkpoints.md#lm-adapted-t511lm100k) includes the following improvements compared to the original [T5 model](https://huggingface.co/t5-large):
- GEGLU activation in feed-forward hidden layer, rather than ReLU - see [here](https://arxiv.org/abs/2002.05202).
- Dropout was turned off in pre-training (quality win). Dropout should be re-enabled during fine-tuning.
- Pre-trained on C4 only without mixing in the downstream tasks.
- no parameter sharing between embedding and classifier layer
- "xl" and "xxl" replace "3B" and "11B". The model shapes are a bit different - larger `d_model` and smaller `num_heads` and `d_ff`.
and is pretrained on both the denoising and language modeling objective.
More specifically, this checkpoint is initialized from [T5 Version 1.1 - Large](https://huggingface.co/google/https://huggingface.co/google/t5-v1_1-large)
and then trained for an additional 100K steps on the LM objective discussed in the [T5 paper](https://arxiv.org/pdf/1910.10683.pdf).
This adaptation improves the ability of the model to be used for prompt tuning.
**Note**: A popular fine-tuned version of the *T5 Version 1.1 - LM Adapted* model is [BigScience's T0pp](https://huggingface.co/bigscience/T0pp).
Pretraining Dataset: [C4](https://huggingface.co/datasets/c4)
Other Community Checkpoints: [here](https://huggingface.co/models?other=t5-lm-adapt)
Paper: [Exploring the Limits of Transfer Learning with a Unified Text-to-Text Transformer](https://arxiv.org/pdf/1910.10683.pdf)
Authors: *Colin Raffel, Noam Shazeer, Adam Roberts, Katherine Lee, Sharan Narang, Michael Matena, Yanqi Zhou, Wei Li, Peter J. Liu*
## Abstract
Transfer learning, where a model is first pre-trained on a data-rich task before being fine-tuned on a downstream task, has emerged as a powerful technique in natural language processing (NLP). The effectiveness of transfer learning has given rise to a diversity of approaches, methodology, and practice. In this paper, we explore the landscape of transfer learning techniques for NLP by introducing a unified framework that converts every language problem into a text-to-text format. Our systematic study compares pre-training objectives, architectures, unlabeled datasets, transfer approaches, and other factors on dozens of language understanding tasks. By combining the insights from our exploration with scale and our new “Colossal Clean Crawled Corpus”, we achieve state-of-the-art results on many benchmarks covering summarization, question answering, text classification, and more. To facilitate future work on transfer learning for NLP, we release our dataset, pre-trained models, and code.
|
textattack/roberta-base-SST-2 | a029a4679e8a56a958d932d1132d6a4f68803214 | 2021-05-20T22:11:39.000Z | [
"pytorch",
"jax",
"roberta",
"text-classification",
"transformers"
] | text-classification | false | textattack | null | textattack/roberta-base-SST-2 | 17,616 | null | transformers | 518 | Entry not found |
EleutherAI/gpt-neox-20b | 364ae95407723fadd1d47b023c1efb92a4d891c3 | 2022-04-07T22:14:56.000Z | [
"pytorch",
"gpt_neox",
"text-generation",
"transformers"
] | text-generation | false | EleutherAI | null | EleutherAI/gpt-neox-20b | 17,557 | 41 | transformers | 519 | Entry not found |
facebook/rag-sequence-nq | c0d9c6ceda8a69c78091abb7aa734a97b75b89fd | 2021-03-12T11:04:28.000Z | [
"pytorch",
"tf",
"rag",
"en",
"dataset:wiki_dpr",
"arxiv:2005.11401",
"transformers",
"license:apache-2.0"
] | null | false | facebook | null | facebook/rag-sequence-nq | 17,541 | 1 | transformers | 520 | ---
language: en
license: apache-2.0
datasets:
- wiki_dpr
thumbnail: https://huggingface.co/front/thumbnails/facebook.png
---
## RAG
This is the RAG-Sequence Model of the the paper [Retrieval-Augmented Generation for Knowledge-Intensive NLP Tasks](https://arxiv.org/pdf/2005.11401.pdf)
by Patrick Lewis, Ethan Perez, Aleksandara Piktus et al.
The model is a *uncased* model, which means that capital letters are simply converted to lower-case letters.
The model consits of a *question_encoder*, *retriever* and a *generator*. The retriever extracts relevant passages from the *wiki_dpr* `train` datasets, which is linked above.
The question_encoder and retriever are based on `facebook/dpr-question_encoder-single-nq-base` and `facebook/bart-large`, which were jointly finetuned on
on the *wiki_dpr* QA dataset in an end-to-end fashion.
## Usage:
**Note**: In the usage example below only the *dummy* retriever of *wiki_dpr* is used because the complete *lecagy* index requires over 75 GB of RAM.
The model can generate answers to any factoid question as follows:
```python
from transformers import RagTokenizer, RagRetriever, RagSequenceForGeneration
tokenizer = RagTokenizer.from_pretrained("facebook/rag-sequence-nq")
retriever = RagRetriever.from_pretrained("facebook/rag-sequence-nq", index_name="exact", use_dummy_dataset=True)
model = RagSequenceForGeneration.from_pretrained("facebook/rag-sequence-nq", retriever=retriever)
input_dict = tokenizer.prepare_seq2seq_batch("how many countries are in europe", return_tensors="pt")
generated = model.generate(input_ids=input_dict["input_ids"])
print(tokenizer.batch_decode(generated, skip_special_tokens=True)[0])
# should give 54 => google says either 44 or 51
```
|
sentence-transformers/bert-base-nli-stsb-mean-tokens | 1bd90bb33d5c6601f5fbd26d91e955a65059ee55 | 2022-06-15T20:01:00.000Z | [
"pytorch",
"tf",
"jax",
"bert",
"feature-extraction",
"arxiv:1908.10084",
"sentence-transformers",
"sentence-similarity",
"transformers",
"license:apache-2.0"
] | sentence-similarity | false | sentence-transformers | null | sentence-transformers/bert-base-nli-stsb-mean-tokens | 17,507 | 1 | sentence-transformers | 521 | ---
pipeline_tag: sentence-similarity
tags:
- sentence-transformers
- feature-extraction
- sentence-similarity
- transformers
license: apache-2.0
---
**⚠️ This model is deprecated. Please don't use it as it produces sentence embeddings of low quality. You can find recommended sentence embedding models here: [SBERT.net - Pretrained Models](https://www.sbert.net/docs/pretrained_models.html)**
# sentence-transformers/bert-base-nli-stsb-mean-tokens
This is a [sentence-transformers](https://www.SBERT.net) model: It maps sentences & paragraphs to a 768 dimensional dense vector space and can be used for tasks like clustering or semantic search.
## Usage (Sentence-Transformers)
Using this model becomes easy when you have [sentence-transformers](https://www.SBERT.net) installed:
```
pip install -U sentence-transformers
```
Then you can use the model like this:
```python
from sentence_transformers import SentenceTransformer
sentences = ["This is an example sentence", "Each sentence is converted"]
model = SentenceTransformer('sentence-transformers/bert-base-nli-stsb-mean-tokens')
embeddings = model.encode(sentences)
print(embeddings)
```
## Usage (HuggingFace Transformers)
Without [sentence-transformers](https://www.SBERT.net), you can use the model like this: First, you pass your input through the transformer model, then you have to apply the right pooling-operation on-top of the contextualized word embeddings.
```python
from transformers import AutoTokenizer, AutoModel
import torch
#Mean Pooling - Take attention mask into account for correct averaging
def mean_pooling(model_output, attention_mask):
token_embeddings = model_output[0] #First element of model_output contains all token embeddings
input_mask_expanded = attention_mask.unsqueeze(-1).expand(token_embeddings.size()).float()
return torch.sum(token_embeddings * input_mask_expanded, 1) / torch.clamp(input_mask_expanded.sum(1), min=1e-9)
# Sentences we want sentence embeddings for
sentences = ['This is an example sentence', 'Each sentence is converted']
# Load model from HuggingFace Hub
tokenizer = AutoTokenizer.from_pretrained('sentence-transformers/bert-base-nli-stsb-mean-tokens')
model = AutoModel.from_pretrained('sentence-transformers/bert-base-nli-stsb-mean-tokens')
# Tokenize sentences
encoded_input = tokenizer(sentences, padding=True, truncation=True, return_tensors='pt')
# Compute token embeddings
with torch.no_grad():
model_output = model(**encoded_input)
# Perform pooling. In this case, max pooling.
sentence_embeddings = mean_pooling(model_output, encoded_input['attention_mask'])
print("Sentence embeddings:")
print(sentence_embeddings)
```
## Evaluation Results
For an automated evaluation of this model, see the *Sentence Embeddings Benchmark*: [https://seb.sbert.net](https://seb.sbert.net?model_name=sentence-transformers/bert-base-nli-stsb-mean-tokens)
## Full Model Architecture
```
SentenceTransformer(
(0): Transformer({'max_seq_length': 128, 'do_lower_case': False}) with Transformer model: BertModel
(1): Pooling({'word_embedding_dimension': 768, 'pooling_mode_cls_token': False, 'pooling_mode_mean_tokens': True, 'pooling_mode_max_tokens': False, 'pooling_mode_mean_sqrt_len_tokens': False})
)
```
## Citing & Authors
This model was trained by [sentence-transformers](https://www.sbert.net/).
If you find this model helpful, feel free to cite our publication [Sentence-BERT: Sentence Embeddings using Siamese BERT-Networks](https://arxiv.org/abs/1908.10084):
```bibtex
@inproceedings{reimers-2019-sentence-bert,
title = "Sentence-BERT: Sentence Embeddings using Siamese BERT-Networks",
author = "Reimers, Nils and Gurevych, Iryna",
booktitle = "Proceedings of the 2019 Conference on Empirical Methods in Natural Language Processing",
month = "11",
year = "2019",
publisher = "Association for Computational Linguistics",
url = "http://arxiv.org/abs/1908.10084",
}
``` |
Helsinki-NLP/opus-mt-da-en | 8971eb3839ec41bddd060128b9b83038bb43fd96 | 2021-09-09T21:29:52.000Z | [
"pytorch",
"marian",
"text2text-generation",
"da",
"en",
"transformers",
"translation",
"license:apache-2.0",
"autotrain_compatible"
] | translation | false | Helsinki-NLP | null | Helsinki-NLP/opus-mt-da-en | 17,464 | 2 | transformers | 522 | ---
tags:
- translation
license: apache-2.0
---
### opus-mt-da-en
* source languages: da
* target languages: en
* OPUS readme: [da-en](https://github.com/Helsinki-NLP/OPUS-MT-train/blob/master/models/da-en/README.md)
* dataset: opus
* model: transformer-align
* pre-processing: normalization + SentencePiece
* download original weights: [opus-2019-12-18.zip](https://object.pouta.csc.fi/OPUS-MT-models/da-en/opus-2019-12-18.zip)
* test set translations: [opus-2019-12-18.test.txt](https://object.pouta.csc.fi/OPUS-MT-models/da-en/opus-2019-12-18.test.txt)
* test set scores: [opus-2019-12-18.eval.txt](https://object.pouta.csc.fi/OPUS-MT-models/da-en/opus-2019-12-18.eval.txt)
## Benchmarks
| testset | BLEU | chr-F |
|-----------------------|-------|-------|
| Tatoeba.da.en | 63.6 | 0.769 |
|
jhgan/ko-sroberta-multitask | ab957ae6a91e99c4cad36d52063a2a9cf1bf4419 | 2022-06-13T16:34:48.000Z | [
"pytorch",
"tf",
"roberta",
"feature-extraction",
"ko",
"sentence-transformers",
"sentence-similarity",
"transformers"
] | sentence-similarity | false | jhgan | null | jhgan/ko-sroberta-multitask | 17,438 | 4 | sentence-transformers | 523 | ---
pipeline_tag: sentence-similarity
tags:
- sentence-transformers
- feature-extraction
- sentence-similarity
- transformers
language: ko
---
# ko-sroberta-multitask
This is a [sentence-transformers](https://www.SBERT.net) model: It maps sentences & paragraphs to a 768 dimensional dense vector space and can be used for tasks like clustering or semantic search.
<!--- Describe your model here -->
## Usage (Sentence-Transformers)
Using this model becomes easy when you have [sentence-transformers](https://www.SBERT.net) installed:
```
pip install -U sentence-transformers
```
Then you can use the model like this:
```python
from sentence_transformers import SentenceTransformer
sentences = ["안녕하세요?", "한국어 문장 임베딩을 위한 버트 모델입니다."]
model = SentenceTransformer('jhgan/ko-sroberta-multitask')
embeddings = model.encode(sentences)
print(embeddings)
```
## Usage (HuggingFace Transformers)
Without [sentence-transformers](https://www.SBERT.net), you can use the model like this: First, you pass your input through the transformer model, then you have to apply the right pooling-operation on-top of the contextualized word embeddings.
```python
from transformers import AutoTokenizer, AutoModel
import torch
#Mean Pooling - Take attention mask into account for correct averaging
def mean_pooling(model_output, attention_mask):
token_embeddings = model_output[0] #First element of model_output contains all token embeddings
input_mask_expanded = attention_mask.unsqueeze(-1).expand(token_embeddings.size()).float()
return torch.sum(token_embeddings * input_mask_expanded, 1) / torch.clamp(input_mask_expanded.sum(1), min=1e-9)
# Sentences we want sentence embeddings for
sentences = ['This is an example sentence', 'Each sentence is converted']
# Load model from HuggingFace Hub
tokenizer = AutoTokenizer.from_pretrained('jhgan/ko-sroberta-multitask')
model = AutoModel.from_pretrained('jhgan/ko-sroberta-multitask')
# Tokenize sentences
encoded_input = tokenizer(sentences, padding=True, truncation=True, return_tensors='pt')
# Compute token embeddings
with torch.no_grad():
model_output = model(**encoded_input)
# Perform pooling. In this case, mean pooling.
sentence_embeddings = mean_pooling(model_output, encoded_input['attention_mask'])
print("Sentence embeddings:")
print(sentence_embeddings)
```
## Evaluation Results
<!--- Describe how your model was evaluated -->
KorSTS, KorNLI 학습 데이터셋으로 멀티 태스크 학습을 진행한 후 KorSTS 평가 데이터셋으로 평가한 결과입니다.
- Cosine Pearson: 84.77
- Cosine Spearman: 85.60
- Euclidean Pearson: 83.71
- Euclidean Spearman: 84.40
- Manhattan Pearson: 83.70
- Manhattan Spearman: 84.38
- Dot Pearson: 82.42
- Dot Spearman: 82.33
## Training
The model was trained with the parameters:
**DataLoader**:
`sentence_transformers.datasets.NoDuplicatesDataLoader.NoDuplicatesDataLoader` of length 8885 with parameters:
```
{'batch_size': 64}
```
**Loss**:
`sentence_transformers.losses.MultipleNegativesRankingLoss.MultipleNegativesRankingLoss` with parameters:
```
{'scale': 20.0, 'similarity_fct': 'cos_sim'}
```
**DataLoader**:
`torch.utils.data.dataloader.DataLoader` of length 719 with parameters:
```
{'batch_size': 8, 'sampler': 'torch.utils.data.sampler.RandomSampler', 'batch_sampler': 'torch.utils.data.sampler.BatchSampler'}
```
**Loss**:
`sentence_transformers.losses.CosineSimilarityLoss.CosineSimilarityLoss`
Parameters of the fit()-Method:
```
{
"epochs": 5,
"evaluation_steps": 1000,
"evaluator": "sentence_transformers.evaluation.EmbeddingSimilarityEvaluator.EmbeddingSimilarityEvaluator",
"max_grad_norm": 1,
"optimizer_class": "<class 'transformers.optimization.AdamW'>",
"optimizer_params": {
"lr": 2e-05
},
"scheduler": "WarmupLinear",
"steps_per_epoch": null,
"warmup_steps": 360,
"weight_decay": 0.01
}
```
## Full Model Architecture
```
SentenceTransformer(
(0): Transformer({'max_seq_length': 128, 'do_lower_case': False}) with Transformer model: RobertaModel
(1): Pooling({'word_embedding_dimension': 768, 'pooling_mode_cls_token': False, 'pooling_mode_mean_tokens': True, 'pooling_mode_max_tokens': False, 'pooling_mode_mean_sqrt_len_tokens': False})
)
```
## Citing & Authors
<!--- Describe where people can find more information -->
- Ham, J., Choe, Y. J., Park, K., Choi, I., & Soh, H. (2020). Kornli and korsts: New benchmark datasets for korean natural language understanding. arXiv
preprint arXiv:2004.03289
- Reimers, Nils and Iryna Gurevych. “Sentence-BERT: Sentence Embeddings using Siamese BERT-Networks.” ArXiv abs/1908.10084 (2019)
- Reimers, Nils and Iryna Gurevych. “Making Monolingual Sentence Embeddings Multilingual Using Knowledge Distillation.” EMNLP (2020).
|
facebook/wav2vec2-large-xlsr-53 | c3f9d884181a224a6ac87bf8885c84d1cff3384f | 2022-03-18T16:11:44.000Z | [
"pytorch",
"jax",
"wav2vec2",
"pretraining",
"multilingual",
"dataset:common_voice",
"arxiv:2006.13979",
"transformers",
"speech",
"license:apache-2.0"
] | null | false | facebook | null | facebook/wav2vec2-large-xlsr-53 | 17,356 | 26 | transformers | 524 | ---
language: multilingual
datasets:
- common_voice
tags:
- speech
license: apache-2.0
---
# Wav2Vec2-XLSR-53
[Facebook's XLSR-Wav2Vec2](https://ai.facebook.com/blog/wav2vec-20-learning-the-structure-of-speech-from-raw-audio/)
The base model pretrained on 16kHz sampled speech audio. When using the model make sure that your speech input is also sampled at 16Khz. Note that this model should be fine-tuned on a downstream task, like Automatic Speech Recognition. Check out [this blog](https://huggingface.co/blog/fine-tune-wav2vec2-english) for more information.
[Paper](https://arxiv.org/abs/2006.13979)
Authors: Alexis Conneau, Alexei Baevski, Ronan Collobert, Abdelrahman Mohamed, Michael Auli
**Abstract**
This paper presents XLSR which learns cross-lingual speech representations by pretraining a single model from the raw waveform of speech in multiple languages. We build on wav2vec 2.0 which is trained by solving a contrastive task over masked latent speech representations and jointly learns a quantization of the latents shared across languages. The resulting model is fine-tuned on labeled data and experiments show that cross-lingual pretraining significantly outperforms monolingual pretraining. On the CommonVoice benchmark, XLSR shows a relative phoneme error rate reduction of 72% compared to the best known results. On BABEL, our approach improves word error rate by 16% relative compared to a comparable system. Our approach enables a single multilingual speech recognition model which is competitive to strong individual models. Analysis shows that the latent discrete speech representations are shared across languages with increased sharing for related languages. We hope to catalyze research in low-resource speech understanding by releasing XLSR-53, a large model pretrained in 53 languages.
The original model can be found under https://github.com/pytorch/fairseq/tree/master/examples/wav2vec#wav2vec-20.
# Usage
See [this notebook](https://colab.research.google.com/github/patrickvonplaten/notebooks/blob/master/Fine_Tune_XLSR_Wav2Vec2_on_Turkish_ASR_with_%F0%9F%A4%97_Transformers.ipynb) for more information on how to fine-tune the model.
|
sentence-transformers/stsb-xlm-r-multilingual | bc1a68705f2e397259207e96349a36ccbc7e6493 | 2022-06-15T21:42:42.000Z | [
"pytorch",
"tf",
"xlm-roberta",
"feature-extraction",
"arxiv:1908.10084",
"sentence-transformers",
"sentence-similarity",
"transformers",
"license:apache-2.0"
] | sentence-similarity | false | sentence-transformers | null | sentence-transformers/stsb-xlm-r-multilingual | 17,252 | 3 | sentence-transformers | 525 | ---
pipeline_tag: sentence-similarity
license: apache-2.0
tags:
- sentence-transformers
- feature-extraction
- sentence-similarity
- transformers
---
# sentence-transformers/stsb-xlm-r-multilingual
This is a [sentence-transformers](https://www.SBERT.net) model: It maps sentences & paragraphs to a 768 dimensional dense vector space and can be used for tasks like clustering or semantic search.
## Usage (Sentence-Transformers)
Using this model becomes easy when you have [sentence-transformers](https://www.SBERT.net) installed:
```
pip install -U sentence-transformers
```
Then you can use the model like this:
```python
from sentence_transformers import SentenceTransformer
sentences = ["This is an example sentence", "Each sentence is converted"]
model = SentenceTransformer('sentence-transformers/stsb-xlm-r-multilingual')
embeddings = model.encode(sentences)
print(embeddings)
```
## Usage (HuggingFace Transformers)
Without [sentence-transformers](https://www.SBERT.net), you can use the model like this: First, you pass your input through the transformer model, then you have to apply the right pooling-operation on-top of the contextualized word embeddings.
```python
from transformers import AutoTokenizer, AutoModel
import torch
#Mean Pooling - Take attention mask into account for correct averaging
def mean_pooling(model_output, attention_mask):
token_embeddings = model_output[0] #First element of model_output contains all token embeddings
input_mask_expanded = attention_mask.unsqueeze(-1).expand(token_embeddings.size()).float()
return torch.sum(token_embeddings * input_mask_expanded, 1) / torch.clamp(input_mask_expanded.sum(1), min=1e-9)
# Sentences we want sentence embeddings for
sentences = ['This is an example sentence', 'Each sentence is converted']
# Load model from HuggingFace Hub
tokenizer = AutoTokenizer.from_pretrained('sentence-transformers/stsb-xlm-r-multilingual')
model = AutoModel.from_pretrained('sentence-transformers/stsb-xlm-r-multilingual')
# Tokenize sentences
encoded_input = tokenizer(sentences, padding=True, truncation=True, return_tensors='pt')
# Compute token embeddings
with torch.no_grad():
model_output = model(**encoded_input)
# Perform pooling. In this case, max pooling.
sentence_embeddings = mean_pooling(model_output, encoded_input['attention_mask'])
print("Sentence embeddings:")
print(sentence_embeddings)
```
## Evaluation Results
For an automated evaluation of this model, see the *Sentence Embeddings Benchmark*: [https://seb.sbert.net](https://seb.sbert.net?model_name=sentence-transformers/stsb-xlm-r-multilingual)
## Full Model Architecture
```
SentenceTransformer(
(0): Transformer({'max_seq_length': 128, 'do_lower_case': False}) with Transformer model: XLMRobertaModel
(1): Pooling({'word_embedding_dimension': 768, 'pooling_mode_cls_token': False, 'pooling_mode_mean_tokens': True, 'pooling_mode_max_tokens': False, 'pooling_mode_mean_sqrt_len_tokens': False})
)
```
## Citing & Authors
This model was trained by [sentence-transformers](https://www.sbert.net/).
If you find this model helpful, feel free to cite our publication [Sentence-BERT: Sentence Embeddings using Siamese BERT-Networks](https://arxiv.org/abs/1908.10084):
```bibtex
@inproceedings{reimers-2019-sentence-bert,
title = "Sentence-BERT: Sentence Embeddings using Siamese BERT-Networks",
author = "Reimers, Nils and Gurevych, Iryna",
booktitle = "Proceedings of the 2019 Conference on Empirical Methods in Natural Language Processing",
month = "11",
year = "2019",
publisher = "Association for Computational Linguistics",
url = "http://arxiv.org/abs/1908.10084",
}
``` |
huggingtweets/anal_sex42069 | c6a452118cfc25d59bbedcf918acba47df3ee243 | 2021-05-21T18:49:44.000Z | [
"pytorch",
"jax",
"gpt2",
"text-generation",
"en",
"transformers",
"huggingtweets"
] | text-generation | false | huggingtweets | null | huggingtweets/anal_sex42069 | 17,247 | null | transformers | 526 | ---
language: en
thumbnail: https://www.huggingtweets.com/anal_sex42069/1617757256637/predictions.png
tags:
- huggingtweets
widget:
- text: "My dream is"
---
<div>
<div style="width: 132px; height:132px; border-radius: 50%; background-size: cover; background-image: url('https://pbs.twimg.com/profile_images/1375931130864537600/f8l03X0z_400x400.jpg')">
</div>
<div style="margin-top: 8px; font-size: 19px; font-weight: 800">👨❤️💋👨❤️ 👩أميرة المكسيك 🤖 AI Bot </div>
<div style="font-size: 15px">@anal_sex42069 bot</div>
</div>
I was made with [huggingtweets](https://github.com/borisdayma/huggingtweets).
Create your own bot based on your favorite user with [the demo](https://colab.research.google.com/github/borisdayma/huggingtweets/blob/master/huggingtweets-demo.ipynb)!
## How does it work?
The model uses the following pipeline.
To understand how the model was developed, check the [W&B report](https://wandb.ai/wandb/huggingtweets/reports/HuggingTweets-Train-a-Model-to-Generate-Tweets--VmlldzoxMTY5MjI).
## Training data
The model was trained on [@anal_sex42069's tweets](https://twitter.com/anal_sex42069).
| Data | Quantity |
| --- | --- |
| Tweets downloaded | 3201 |
| Retweets | 1101 |
| Short tweets | 477 |
| Tweets kept | 1623 |
[Explore the data](https://wandb.ai/wandb/huggingtweets/runs/10kifa1d/artifacts), which is tracked with [W&B artifacts](https://docs.wandb.com/artifacts) at every step of the pipeline.
## Training procedure
The model is based on a pre-trained [GPT-2](https://huggingface.co/gpt2) which is fine-tuned on @anal_sex42069's tweets.
Hyperparameters and metrics are recorded in the [W&B training run](https://wandb.ai/wandb/huggingtweets/runs/gye99qzo) for full transparency and reproducibility.
At the end of training, [the final model](https://wandb.ai/wandb/huggingtweets/runs/gye99qzo/artifacts) is logged and versioned.
## How to use
You can use this model directly with a pipeline for text generation:
```python
from transformers import pipeline
generator = pipeline('text-generation',
model='huggingtweets/anal_sex42069')
generator("My dream is", num_return_sequences=5)
```
## Limitations and bias
The model suffers from [the same limitations and bias as GPT-2](https://huggingface.co/gpt2#limitations-and-bias).
In addition, the data present in the user's tweets further affects the text generated by the model.
## About
*Built by Boris Dayma*
[](https://twitter.com/intent/follow?screen_name=borisdayma)
For more details, visit the project repository.
[](https://github.com/borisdayma/huggingtweets)
|
google/tapas-base | 00456266840bb0a319cd6748ebf7da3caf98816b | 2021-11-29T10:03:33.000Z | [
"pytorch",
"tf",
"tapas",
"feature-extraction",
"en",
"arxiv:2004.02349",
"arxiv:2010.00571",
"transformers",
"TapasModel",
"license:apache-2.0"
] | feature-extraction | false | google | null | google/tapas-base | 17,168 | 3 | transformers | 527 | ---
language: en
tags:
- tapas
- TapasModel
license: apache-2.0
---
# TAPAS base model
This model has 2 versions which can be used. The latest version, which is the default one, corresponds to the `tapas_inter_masklm_base_reset` checkpoint of the [original Github repository](https://github.com/google-research/tapas).
This model was pre-trained on MLM and an additional step which the authors call intermediate pre-training. It uses relative position embeddings by default (i.e. resetting the position index at every cell of the table).
The other (non-default) version which can be used is the one with absolute position embeddings:
- `revision="no_reset"`, which corresponds to `tapas_inter_masklm_base`
Disclaimer: The team releasing TAPAS did not write a model card for this model so this model card has been written by
the Hugging Face team and contributors.
## Model description
TAPAS is a BERT-like transformers model pretrained on a large corpus of English data from Wikipedia in a self-supervised fashion.
This means it was pretrained on the raw tables and associated texts only, with no humans labelling them in any way (which is why it
can use lots of publicly available data) with an automatic process to generate inputs and labels from those texts. More precisely, it
was pretrained with two objectives:
- Masked language modeling (MLM): taking a (flattened) table and associated context, the model randomly masks 15% of the words in
the input, then runs the entire (partially masked) sequence through the model. The model then has to predict the masked words.
This is different from traditional recurrent neural networks (RNNs) that usually see the words one after the other,
or from autoregressive models like GPT which internally mask the future tokens. It allows the model to learn a bidirectional
representation of a table and associated text.
- Intermediate pre-training: to encourage numerical reasoning on tables, the authors additionally pre-trained the model by creating
a balanced dataset of millions of syntactically created training examples. Here, the model must predict (classify) whether a sentence
is supported or refuted by the contents of a table. The training examples are created based on synthetic as well as counterfactual statements.
This way, the model learns an inner representation of the English language used in tables and associated texts, which can then be used
to extract features useful for downstream tasks such as answering questions about a table, or determining whether a sentence is entailed
or refuted by the contents of a table. Fine-tuning is done by adding one or more classification heads on top of the pre-trained model, and then
jointly train these randomly initialized classification heads with the base model on a downstream task.
## Intended uses & limitations
You can use the raw model for getting hidden representatons about table-question pairs, but it's mostly intended to be fine-tuned on a downstream task such as question answering or sequence classification. See the [model hub](https://huggingface.co/models?filter=tapas) to look for fine-tuned versions on a task that interests you.
## Training procedure
### Preprocessing
The texts are lowercased and tokenized using WordPiece and a vocabulary size of 30,000. The inputs of the model are
then of the form:
```
[CLS] Sentence [SEP] Flattened table [SEP]
```
### Pre-training
The model was pre-trained on 32 Cloud TPU v3 cores for 1,000,000 steps with maximum sequence length 512 and batch size of 512.
In this setup, pre-training on MLM only takes around 3 days. Aditionally, the model has been further pre-trained on a second task (table entailment). See the original TAPAS [paper](https://www.aclweb.org/anthology/2020.acl-main.398/) and the [follow-up paper](https://www.aclweb.org/anthology/2020.findings-emnlp.27/) for more details.
The optimizer used is Adam with a learning rate of 5e-5, and a warmup
ratio of 0.01.
### BibTeX entry and citation info
```bibtex
@misc{herzig2020tapas,
title={TAPAS: Weakly Supervised Table Parsing via Pre-training},
author={Jonathan Herzig and Paweł Krzysztof Nowak and Thomas Müller and Francesco Piccinno and Julian Martin Eisenschlos},
year={2020},
eprint={2004.02349},
archivePrefix={arXiv},
primaryClass={cs.IR}
}
```
```bibtex
@misc{eisenschlos2020understanding,
title={Understanding tables with intermediate pre-training},
author={Julian Martin Eisenschlos and Syrine Krichene and Thomas Müller},
year={2020},
eprint={2010.00571},
archivePrefix={arXiv},
primaryClass={cs.CL}
}
``` |
vicgalle/xlm-roberta-large-xnli-anli | 81de27372786f6a034c81c4bf1c53ebe9afa10d7 | 2021-03-04T17:05:03.000Z | [
"pytorch",
"xlm-roberta",
"text-classification",
"multilingual",
"dataset:mnli",
"dataset:xnli",
"dataset:anli",
"transformers",
"zero-shot-classification",
"nli",
"license:mit"
] | zero-shot-classification | false | vicgalle | null | vicgalle/xlm-roberta-large-xnli-anli | 17,093 | 8 | transformers | 528 | ---
language: multilingual
tags:
- zero-shot-classification
- nli
- pytorch
datasets:
- mnli
- xnli
- anli
license: mit
pipeline_tag: zero-shot-classification
widget:
- text: "De pugna erat fantastic. Nam Crixo decem quam dilexit et praeciderunt caput aemulus."
candidate_labels: "violent, peaceful"
- text: "La película empezaba bien pero terminó siendo un desastre."
candidate_labels: "positivo, negativo, neutral"
- text: "La película empezó siendo un desastre pero en general fue bien."
candidate_labels: "positivo, negativo, neutral"
- text: "¿A quién vas a votar en 2020?"
candidate_labels: "Europa, elecciones, política, ciencia, deportes"
---
### XLM-RoBERTa-large-XNLI-ANLI
XLM-RoBERTa-large model finetunned over several NLI datasets, ready to use for zero-shot classification.
Here are the accuracies for several test datasets:
| | XNLI-es | XNLI-fr | ANLI-R1 | ANLI-R2 | ANLI-R3 |
|-----------------------------|---------|---------|---------|---------|---------|
| xlm-roberta-large-xnli-anli | 93.7% | 93.2% | 68.5% | 53.6% | 49.0% |
The model can be loaded with the zero-shot-classification pipeline like so:
```
from transformers import pipeline
classifier = pipeline("zero-shot-classification",
model="vicgalle/xlm-roberta-large-xnli-anli")
```
You can then use this pipeline to classify sequences into any of the class names you specify:
```
sequence_to_classify = "Algún día iré a ver el mundo"
candidate_labels = ['viaje', 'cocina', 'danza']
classifier(sequence_to_classify, candidate_labels)
#{'sequence': 'Algún día iré a ver el mundo',
#'labels': ['viaje', 'danza', 'cocina'],
#'scores': [0.9991760849952698, 0.0004178212257102132, 0.0004059972707182169]}
``` |
lidiya/bart-large-xsum-samsum | 5f600a362ff4b9efbaf7e3cbbdd853a3a89c118f | 2022-07-20T14:55:59.000Z | [
"pytorch",
"bart",
"text2text-generation",
"en",
"dataset:samsum",
"transformers",
"seq2seq",
"summarization",
"license:apache-2.0",
"model-index",
"autotrain_compatible"
] | summarization | false | lidiya | null | lidiya/bart-large-xsum-samsum | 17,031 | 8 | transformers | 529 | ---
language: en
tags:
- bart
- seq2seq
- summarization
license: apache-2.0
datasets:
- samsum
widget:
- text: |
Hannah: Hey, do you have Betty's number?
Amanda: Lemme check
Amanda: Sorry, can't find it.
Amanda: Ask Larry
Amanda: He called her last time we were at the park together
Hannah: I don't know him well
Amanda: Don't be shy, he's very nice
Hannah: If you say so..
Hannah: I'd rather you texted him
Amanda: Just text him 🙂
Hannah: Urgh.. Alright
Hannah: Bye
Amanda: Bye bye
model-index:
- name: bart-large-xsum-samsum
results:
- task:
name: Abstractive Text Summarization
type: abstractive-text-summarization
dataset:
name: "SAMSum Corpus: A Human-annotated Dialogue Dataset for Abstractive Summarization"
type: samsum
metrics:
- name: Validation ROUGE-1
type: rouge-1
value: 54.3921
- name: Validation ROUGE-2
type: rouge-2
value: 29.8078
- name: Validation ROUGE-L
type: rouge-l
value: 45.1543
- name: Test ROUGE-1
type: rouge-1
value: 53.3059
- name: Test ROUGE-2
type: rouge-2
value: 28.355
- name: Test ROUGE-L
type: rouge-l
value: 44.0953
---
## `bart-large-xsum-samsum`
This model was obtained by fine-tuning `facebook/bart-large-xsum` on [Samsum](https://huggingface.co/datasets/samsum) dataset.
## Usage
```python
from transformers import pipeline
summarizer = pipeline("summarization", model="lidiya/bart-large-xsum-samsum")
conversation = '''Hannah: Hey, do you have Betty's number?
Amanda: Lemme check
Amanda: Sorry, can't find it.
Amanda: Ask Larry
Amanda: He called her last time we were at the park together
Hannah: I don't know him well
Amanda: Don't be shy, he's very nice
Hannah: If you say so..
Hannah: I'd rather you texted him
Amanda: Just text him 🙂
Hannah: Urgh.. Alright
Hannah: Bye
Amanda: Bye bye
'''
summarizer(conversation)
```
## Training procedure
- Colab notebook: https://colab.research.google.com/drive/1dul0Sg-TTMy9xZCJzmDRajXbyzDwtYx6?usp=sharing
## Results
| key | value |
| --- | ----- |
| eval_rouge1 | 54.3921 |
| eval_rouge2 | 29.8078 |
| eval_rougeL | 45.1543 |
| eval_rougeLsum | 49.942 |
| test_rouge1 | 53.3059 |
| test_rouge2 | 28.355 |
| test_rougeL | 44.0953 |
| test_rougeLsum | 48.9246 | |
xlm-clm-ende-1024 | bfe86064e0ce0060985a2a47637cb2efa6631a8d | 2022-07-22T08:04:36.000Z | [
"pytorch",
"tf",
"xlm",
"fill-mask",
"multilingual",
"en",
"de",
"arxiv:1901.07291",
"arxiv:1910.09700",
"transformers",
"autotrain_compatible"
] | fill-mask | false | null | null | xlm-clm-ende-1024 | 16,967 | null | transformers | 530 | ---
language:
- multilingual
- en
- de
---
# xlm-clm-ende-1024
# Table of Contents
1. [Model Details](#model-details)
2. [Uses](#uses)
3. [Bias, Risks, and Limitations](#bias-risks-and-limitations)
4. [Training](#training)
5. [Evaluation](#evaluation)
6. [Environmental Impact](#environmental-impact)
7. [Technical Specifications](#technical-specifications)
8. [Citation](#citation)
9. [Model Card Authors](#model-card-authors)
10. [How To Get Started With the Model](#how-to-get-started-with-the-model)
# Model Details
The XLM model was proposed in [Cross-lingual Language Model Pretraining](https://arxiv.org/abs/1901.07291) by Guillaume Lample, Alexis Conneau. xlm-clm-ende-1024 is a transformer pretrained using a causal language modeling (CLM) objective (next token prediction) for English-German.
## Model Description
- **Developed by:** Guillaume Lample, Alexis Conneau, see [associated paper](https://arxiv.org/abs/1901.07291)
- **Model type:** Language model
- **Language(s) (NLP):** English-German
- **License:** Unknown
- **Related Models:** [xlm-clm-enfr-1024](https://huggingface.co/xlm-clm-enfr-1024), [xlm-mlm-ende-1024](https://huggingface.co/xlm-mlm-ende-1024), [xlm-mlm-enfr-1024](https://huggingface.co/xlm-mlm-enfr-1024), [xlm-mlm-enro-1024](https://huggingface.co/xlm-mlm-enro-1024)
- **Resources for more information:**
- [Associated paper](https://arxiv.org/abs/1901.07291)
- [GitHub Repo](https://github.com/facebookresearch/XLM)
- [Hugging Face Multilingual Models for Inference docs](https://huggingface.co/docs/transformers/v4.20.1/en/multilingual#xlm-with-language-embeddings)
# Uses
## Direct Use
The model is a language model. The model can be used for causal language modeling.
## Downstream Use
To learn more about this task and potential downstream uses, see the [Hugging Face Multilingual Models for Inference](https://huggingface.co/docs/transformers/v4.20.1/en/multilingual#xlm-with-language-embeddings) docs.
## Out-of-Scope Use
The model should not be used to intentionally create hostile or alienating environments for people.
# Bias, Risks, and Limitations
Significant research has explored bias and fairness issues with language models (see, e.g., [Sheng et al. (2021)](https://aclanthology.org/2021.acl-long.330.pdf) and [Bender et al. (2021)](https://dl.acm.org/doi/pdf/10.1145/3442188.3445922)).
## Recommendations
Users (both direct and downstream) should be made aware of the risks, biases and limitations of the model.
# Training
See the [associated paper](https://arxiv.org/pdf/1901.07291.pdf) for details on the training data and training procedure.
# Evaluation
## Testing Data, Factors & Metrics
See the [associated paper](https://arxiv.org/pdf/1901.07291.pdf) for details on the testing data, factors and metrics.
## Results
For xlm-clm-ende-1024 results, see Table 2 of the [associated paper](https://arxiv.org/pdf/1901.07291.pdf).
# Environmental Impact
Carbon emissions can be estimated using the [Machine Learning Impact calculator](https://mlco2.github.io/impact#compute) presented in [Lacoste et al. (2019)](https://arxiv.org/abs/1910.09700).
- **Hardware Type:** More information needed
- **Hours used:** More information needed
- **Cloud Provider:** More information needed
- **Compute Region:** More information needed
- **Carbon Emitted:** More information needed
# Technical Specifications
The model developers write:
> We implement all our models in PyTorch (Paszke et al., 2017), and train them on 64 Volta GPUs for the language modeling tasks, and 8 GPUs for the MT tasks. We use float16 operations to speed up training and to reduce the memory usage of our models.
See the [associated paper](https://arxiv.org/pdf/1901.07291.pdf) for further details.
# Citation
**BibTeX:**
```bibtex
@article{lample2019cross,
title={Cross-lingual language model pretraining},
author={Lample, Guillaume and Conneau, Alexis},
journal={arXiv preprint arXiv:1901.07291},
year={2019}
}
```
**APA:**
- Lample, G., & Conneau, A. (2019). Cross-lingual language model pretraining. arXiv preprint arXiv:1901.07291.
# Model Card Authors
This model card was written by the team at Hugging Face.
# How to Get Started with the Model
Use the code below to get started with the model.
<details>
<summary> Click to expand </summary>
```python
import torch
from transformers import XLMTokenizer, XLMWithLMHeadModel
tokenizer = XLMTokenizer.from_pretrained("xlm-clm-ende-1024")
model = XLMWithLMHeadModel.from_pretrained("xlm-clm-ende-1024")
input_ids = torch.tensor([tokenizer.encode("Wikipedia was used to")]) # batch size of 1
language_id = tokenizer.lang2id["en"] # 0
langs = torch.tensor([language_id] * input_ids.shape[1]) # torch.tensor([0, 0, 0, ..., 0])
# We reshape it to be of size (batch_size, sequence_length)
langs = langs.view(1, -1) # is now of shape [1, sequence_length] (we have a batch size of 1)
outputs = model(input_ids, langs=langs)
```
</details> |
google/bigbird-pegasus-large-arxiv | 9d87686a36f9c732db7deeae1dcac5fb085d0b90 | 2022-06-29T20:40:12.000Z | [
"pytorch",
"bigbird_pegasus",
"text2text-generation",
"en",
"dataset:scientific_papers",
"arxiv:2007.14062",
"transformers",
"summarization",
"license:apache-2.0",
"model-index",
"autotrain_compatible"
] | summarization | false | google | null | google/bigbird-pegasus-large-arxiv | 16,870 | 8 | transformers | 531 | ---
language: en
license: apache-2.0
datasets:
- scientific_papers
tags:
- summarization
model-index:
- name: google/bigbird-pegasus-large-arxiv
results:
- task:
type: summarization
name: Summarization
dataset:
name: scientific_papers
type: scientific_papers
config: pubmed
split: test
metrics:
- name: ROUGE-1
type: rouge
value: 36.0276
verified: true
- name: ROUGE-2
type: rouge
value: 13.4166
verified: true
- name: ROUGE-L
type: rouge
value: 21.9612
verified: true
- name: ROUGE-LSUM
type: rouge
value: 29.648
verified: true
- name: loss
type: loss
value: 2.774355173110962
verified: true
- name: meteor
type: meteor
value: 0.2824
verified: true
- name: gen_len
type: gen_len
value: 209.2537
verified: true
---
# BigBirdPegasus model (large)
BigBird, is a sparse-attention based transformer which extends Transformer based models, such as BERT to much longer sequences. Moreover, BigBird comes along with a theoretical understanding of the capabilities of a complete transformer that the sparse model can handle.
BigBird was introduced in this [paper](https://arxiv.org/abs/2007.14062) and first released in this [repository](https://github.com/google-research/bigbird).
Disclaimer: The team releasing BigBird did not write a model card for this model so this model card has been written by the Hugging Face team.
## Model description
BigBird relies on **block sparse attention** instead of normal attention (i.e. BERT's attention) and can handle sequences up to a length of 4096 at a much lower compute cost compared to BERT. It has achieved SOTA on various tasks involving very long sequences such as long documents summarization, question-answering with long contexts.
## How to use
Here is how to use this model to get the features of a given text in PyTorch:
```python
from transformers import BigBirdPegasusForConditionalGeneration, AutoTokenizer
tokenizer = AutoTokenizer.from_pretrained("google/bigbird-pegasus-large-arxiv")
# by default encoder-attention is `block_sparse` with num_random_blocks=3, block_size=64
model = BigBirdPegasusForConditionalGeneration.from_pretrained("google/bigbird-pegasus-large-arxiv")
# decoder attention type can't be changed & will be "original_full"
# you can change `attention_type` (encoder only) to full attention like this:
model = BigBirdPegasusForConditionalGeneration.from_pretrained("google/bigbird-pegasus-large-arxiv", attention_type="original_full")
# you can change `block_size` & `num_random_blocks` like this:
model = BigBirdPegasusForConditionalGeneration.from_pretrained("google/bigbird-pegasus-large-arxiv", block_size=16, num_random_blocks=2)
text = "Replace me by any text you'd like."
inputs = tokenizer(text, return_tensors='pt')
prediction = model.generate(**inputs)
prediction = tokenizer.batch_decode(prediction)
```
## Training Procedure
This checkpoint is obtained after fine-tuning `BigBirdPegasusForConditionalGeneration` for **summarization** on **arxiv dataset** from [scientific_papers](https://huggingface.co/datasets/scientific_papers).
## BibTeX entry and citation info
```tex
@misc{zaheer2021big,
title={Big Bird: Transformers for Longer Sequences},
author={Manzil Zaheer and Guru Guruganesh and Avinava Dubey and Joshua Ainslie and Chris Alberti and Santiago Ontanon and Philip Pham and Anirudh Ravula and Qifan Wang and Li Yang and Amr Ahmed},
year={2021},
eprint={2007.14062},
archivePrefix={arXiv},
primaryClass={cs.LG}
}
```
|
facebook/muppet-roberta-base | caf238c63db946bdfbd00575713462838e823997 | 2021-06-28T21:44:23.000Z | [
"pytorch",
"roberta",
"fill-mask",
"en",
"dataset:bookcorpus",
"dataset:wikipedia",
"arxiv:2101.11038",
"transformers",
"exbert",
"license:mit",
"autotrain_compatible"
] | fill-mask | false | facebook | null | facebook/muppet-roberta-base | 16,709 | 2 | transformers | 532 | ---
language: en
tags:
- exbert
license: mit
datasets:
- bookcorpus
- wikipedia
---
# Muppet: Massive Multi-task Representations with Pre-Finetuning
# RoBERTa base model
This is a Massive Multi-task Pre-finetuned version of Roberta base. It was introduced in
[this paper](https://arxiv.org/abs/2101.11038). The model improves over roberta-base in a wide range of GLUE, QA tasks (details can be found in the paper). The gains in
smaller datasets are significant.
Note: This checkpoint does not contain the classificaiton/MRC heads used during pre-finetuning due to compatibility issues and hence you might get slightly lower performance than that reported in the paper on some datasets
## Model description
RoBERTa is a transformers model pretrained on a large corpus of English data in a self-supervised fashion. This means
it was pretrained on the raw texts only, with no humans labelling them in any way (which is why it can use lots of
publicly available data) with an automatic process to generate inputs and labels from those texts.
More precisely, it was pretrained with the Masked language modeling (MLM) objective. Taking a sentence, the model
randomly masks 15% of the words in the input then run the entire masked sentence through the model and has to predict
the masked words. This is different from traditional recurrent neural networks (RNNs) that usually see the words one
after the other, or from autoregressive models like GPT which internally mask the future tokens. It allows the model to
learn a bidirectional representation of the sentence.
This way, the model learns an inner representation of the English language that can then be used to extract features
useful for downstream tasks: if you have a dataset of labeled sentences for instance, you can train a standard
classifier using the features produced by the BERT model as inputs.
## Intended uses & limitations
You can use the raw model for masked language modeling, but it's mostly intended to be fine-tuned on a downstream task.
See the [model hub](https://huggingface.co/models?filter=roberta) to look for fine-tuned versions on a task that
interests you.
Note that this model is primarily aimed at being fine-tuned on tasks that use the whole sentence (potentially masked)
to make decisions, such as sequence classification, token classification or question answering. For tasks such as text
generation you should look at model like GPT2.
## Evaluation results
When fine-tuned on downstream tasks, this model achieves the following results:
Glue test results:
| Model | MNLI | QQP | QNLI | SST-2 | CoLA | STS-B | MRPC | RTE | SQuAD|
|:----:|:----:|:----:|:----:|:-----:|:----:|:-----:|:----:|:----:|:----:|
| Roberta-base | 87.6 | 91.9 | 92.8 | 94.8 | 63.6 | 91.2 | 90.2 | 78.7 | 82.6|
| MUPPET Roberta-base | 88.1 | 91.9 | 93.3 | 96.7 | - | - | 91.7 | 87.8 | 86.6|
### BibTeX entry and citation info
```bibtex
@article{DBLP:journals/corr/abs-2101-11038,
author = {Armen Aghajanyan and
Anchit Gupta and
Akshat Shrivastava and
Xilun Chen and
Luke Zettlemoyer and
Sonal Gupta},
title = {Muppet: Massive Multi-task Representations with Pre-Finetuning},
journal = {CoRR},
volume = {abs/2101.11038},
year = {2021},
url = {https://arxiv.org/abs/2101.11038},
archivePrefix = {arXiv},
eprint = {2101.11038},
timestamp = {Sun, 31 Jan 2021 17:23:50 +0100},
biburl = {https://dblp.org/rec/journals/corr/abs-2101-11038.bib},
bibsource = {dblp computer science bibliography, https://dblp.org}
}
``` |
KB/bert-base-swedish-cased | 81c7baa04742a30cb6732c181e678721868cb42e | 2022-06-07T16:31:14.000Z | [
"pytorch",
"tf",
"jax",
"bert",
"fill-mask",
"sv",
"transformers",
"autotrain_compatible"
] | fill-mask | false | KB | null | KB/bert-base-swedish-cased | 16,702 | 1 | transformers | 533 | ---
language: sv
---
# Swedish BERT Models
The National Library of Sweden / KBLab releases three pretrained language models based on BERT and ALBERT. The models are trained on aproximately 15-20GB of text (200M sentences, 3000M tokens) from various sources (books, news, government publications, swedish wikipedia and internet forums) aiming to provide a representative BERT model for Swedish text. A more complete description will be published later on.
The following three models are currently available:
- **bert-base-swedish-cased** (*v1*) - A BERT trained with the same hyperparameters as first published by Google.
- **bert-base-swedish-cased-ner** (*experimental*) - a BERT fine-tuned for NER using SUC 3.0.
- **albert-base-swedish-cased-alpha** (*alpha*) - A first attempt at an ALBERT for Swedish.
All models are cased and trained with whole word masking.
## Files
| **name** | **files** |
|---------------------------------|-----------|
| bert-base-swedish-cased | [config](https://s3.amazonaws.com/models.huggingface.co/bert/KB/bert-base-swedish-cased/config.json), [vocab](https://s3.amazonaws.com/models.huggingface.co/bert/KB/bert-base-swedish-cased/vocab.txt), [pytorch_model.bin](https://s3.amazonaws.com/models.huggingface.co/bert/KB/bert-base-swedish-cased/pytorch_model.bin) |
| bert-base-swedish-cased-ner | [config](https://s3.amazonaws.com/models.huggingface.co/bert/KB/bert-base-swedish-cased-ner/config.json), [vocab](https://s3.amazonaws.com/models.huggingface.co/bert/KB/bert-base-swedish-cased-ner/vocab.txt) [pytorch_model.bin](https://s3.amazonaws.com/models.huggingface.co/bert/KB/bert-base-swedish-cased-ner/pytorch_model.bin) |
| albert-base-swedish-cased-alpha | [config](https://s3.amazonaws.com/models.huggingface.co/bert/KB/albert-base-swedish-cased-alpha/config.json), [sentencepiece model](https://s3.amazonaws.com/models.huggingface.co/bert/KB/albert-base-swedish-cased-alpha/spiece.model), [pytorch_model.bin](https://s3.amazonaws.com/models.huggingface.co/bert/KB/albert-base-swedish-cased-alpha/pytorch_model.bin) |
TensorFlow model weights will be released soon.
## Usage requirements / installation instructions
The examples below require Huggingface Transformers 2.4.1 and Pytorch 1.3.1 or greater. For Transformers<2.4.0 the tokenizer must be instantiated manually and the `do_lower_case` flag parameter set to `False` and `keep_accents` to `True` (for ALBERT).
To create an environment where the examples can be run, run the following in an terminal on your OS of choice.
```
# git clone https://github.com/Kungbib/swedish-bert-models
# cd swedish-bert-models
# python3 -m venv venv
# source venv/bin/activate
# pip install --upgrade pip
# pip install -r requirements.txt
```
### BERT Base Swedish
A standard BERT base for Swedish trained on a variety of sources. Vocabulary size is ~50k. Using Huggingface Transformers the model can be loaded in Python as follows:
```python
from transformers import AutoModel,AutoTokenizer
tok = AutoTokenizer.from_pretrained('KB/bert-base-swedish-cased')
model = AutoModel.from_pretrained('KB/bert-base-swedish-cased')
```
### BERT base fine-tuned for Swedish NER
This model is fine-tuned on the SUC 3.0 dataset. Using the Huggingface pipeline the model can be easily instantiated. For Transformer<2.4.1 it seems the tokenizer must be loaded separately to disable lower-casing of input strings:
```python
from transformers import pipeline
nlp = pipeline('ner', model='KB/bert-base-swedish-cased-ner', tokenizer='KB/bert-base-swedish-cased-ner')
nlp('Idag släpper KB tre språkmodeller.')
```
Running the Python code above should produce in something like the result below. Entity types used are `TME` for time, `PRS` for personal names, `LOC` for locations, `EVN` for events and `ORG` for organisations. These labels are subject to change.
```python
[ { 'word': 'Idag', 'score': 0.9998126029968262, 'entity': 'TME' },
{ 'word': 'KB', 'score': 0.9814832210540771, 'entity': 'ORG' } ]
```
The BERT tokenizer often splits words into multiple tokens, with the subparts starting with `##`, for example the string `Engelbert kör Volvo till Herrängens fotbollsklubb` gets tokenized as `Engel ##bert kör Volvo till Herr ##ängens fotbolls ##klubb`. To glue parts back together one can use something like this:
```python
text = 'Engelbert tar Volvon till Tele2 Arena för att titta på Djurgården IF ' +\
'som spelar fotboll i VM klockan två på kvällen.'
l = []
for token in nlp(text):
if token['word'].startswith('##'):
l[-1]['word'] += token['word'][2:]
else:
l += [ token ]
print(l)
```
Which should result in the following (though less cleanly formated):
```python
[ { 'word': 'Engelbert', 'score': 0.99..., 'entity': 'PRS'},
{ 'word': 'Volvon', 'score': 0.99..., 'entity': 'OBJ'},
{ 'word': 'Tele2', 'score': 0.99..., 'entity': 'LOC'},
{ 'word': 'Arena', 'score': 0.99..., 'entity': 'LOC'},
{ 'word': 'Djurgården', 'score': 0.99..., 'entity': 'ORG'},
{ 'word': 'IF', 'score': 0.99..., 'entity': 'ORG'},
{ 'word': 'VM', 'score': 0.99..., 'entity': 'EVN'},
{ 'word': 'klockan', 'score': 0.99..., 'entity': 'TME'},
{ 'word': 'två', 'score': 0.99..., 'entity': 'TME'},
{ 'word': 'på', 'score': 0.99..., 'entity': 'TME'},
{ 'word': 'kvällen', 'score': 0.54..., 'entity': 'TME'} ]
```
### ALBERT base
The easisest way to do this is, again, using Huggingface Transformers:
```python
from transformers import AutoModel,AutoTokenizer
tok = AutoTokenizer.from_pretrained('KB/albert-base-swedish-cased-alpha'),
model = AutoModel.from_pretrained('KB/albert-base-swedish-cased-alpha')
```
## Acknowledgements ❤️
- Resources from Stockholms University, Umeå University and Swedish Language Bank at Gothenburg University were used when fine-tuning BERT for NER.
- Model pretraining was made partly in-house at the KBLab and partly (for material without active copyright) with the support of Cloud TPUs from Google's TensorFlow Research Cloud (TFRC).
- Models are hosted on S3 by Huggingface 🤗
|
facebook/mbart-large-50-many-to-one-mmt | aadfc3b9db11b773f823ad936d5d683c470e7683 | 2022-05-26T22:28:02.000Z | [
"pytorch",
"jax",
"mbart",
"text2text-generation",
"multilingual",
"ar",
"cs",
"de",
"en",
"es",
"et",
"fi",
"fr",
"gu",
"hi",
"it",
"ja",
"kk",
"ko",
"lt",
"lv",
"my",
"ne",
"nl",
"ro",
"ru",
"si",
"tr",
"vi",
"zh",
"af",
"az",
"bn",
"fa",
"he",
"hr",
"id",
"ka",
"km",
"mk",
"ml",
"mn",
"mr",
"pl",
"ps",
"pt",
"sv",
"sw",
"ta",
"te",
"th",
"tl",
"uk",
"ur",
"xh",
"gl",
"sl",
"arxiv:2008.00401",
"transformers",
"mbart-50",
"autotrain_compatible"
] | text2text-generation | false | facebook | null | facebook/mbart-large-50-many-to-one-mmt | 16,684 | 2 | transformers | 534 | ---
language:
- multilingual
- ar
- cs
- de
- en
- es
- et
- fi
- fr
- gu
- hi
- it
- ja
- kk
- ko
- lt
- lv
- my
- ne
- nl
- ro
- ru
- si
- tr
- vi
- zh
- af
- az
- bn
- fa
- he
- hr
- id
- ka
- km
- mk
- ml
- mn
- mr
- pl
- ps
- pt
- sv
- sw
- ta
- te
- th
- tl
- uk
- ur
- xh
- gl
- sl
tags:
- mbart-50
---
# mBART-50 many to one multilingual machine translation
This model is a fine-tuned checkpoint of [mBART-large-50](https://huggingface.co/facebook/mbart-large-50). `mbart-large-50-many-to-many-mmt` is fine-tuned for multilingual machine translation. It was introduced in [Multilingual Translation with Extensible Multilingual Pretraining and Finetuning](https://arxiv.org/abs/2008.00401) paper.
The model can translate directly between any pair of 50 languages.
```python
from transformers import MBartForConditionalGeneration, MBart50TokenizerFast
article_hi = "संयुक्त राष्ट्र के प्रमुख का कहना है कि सीरिया में कोई सैन्य समाधान नहीं है"
article_ar = "الأمين العام للأمم المتحدة يقول إنه لا يوجد حل عسكري في سوريا."
model = MBartForConditionalGeneration.from_pretrained("facebook/mbart-large-50-many-to-one-mmt")
tokenizer = MBart50TokenizerFast.from_pretrained("facebook/mbart-large-50-many-to-one-mmt")
# translate Hindi to English
tokenizer.src_lang = "hi_IN"
encoded_hi = tokenizer(article_hi, return_tensors="pt")
generated_tokens = model.generate(**encoded_hi)
tokenizer.batch_decode(generated_tokens, skip_special_tokens=True)
# => "The head of the UN says there is no military solution in Syria."
# translate Arabic to English
tokenizer.src_lang = "ar_AR"
encoded_ar = tokenizer(article_ar, return_tensors="pt")
generated_tokens = model.generate(**encoded_ar)
tokenizer.batch_decode(generated_tokens, skip_special_tokens=True)
# => "The Secretary-General of the United Nations says there is no military solution in Syria."
```
See the [model hub](https://huggingface.co/models?filter=mbart-50) to look for more fine-tuned versions.
## Languages covered
Arabic (ar_AR), Czech (cs_CZ), German (de_DE), English (en_XX), Spanish (es_XX), Estonian (et_EE), Finnish (fi_FI), French (fr_XX), Gujarati (gu_IN), Hindi (hi_IN), Italian (it_IT), Japanese (ja_XX), Kazakh (kk_KZ), Korean (ko_KR), Lithuanian (lt_LT), Latvian (lv_LV), Burmese (my_MM), Nepali (ne_NP), Dutch (nl_XX), Romanian (ro_RO), Russian (ru_RU), Sinhala (si_LK), Turkish (tr_TR), Vietnamese (vi_VN), Chinese (zh_CN), Afrikaans (af_ZA), Azerbaijani (az_AZ), Bengali (bn_IN), Persian (fa_IR), Hebrew (he_IL), Croatian (hr_HR), Indonesian (id_ID), Georgian (ka_GE), Khmer (km_KH), Macedonian (mk_MK), Malayalam (ml_IN), Mongolian (mn_MN), Marathi (mr_IN), Polish (pl_PL), Pashto (ps_AF), Portuguese (pt_XX), Swedish (sv_SE), Swahili (sw_KE), Tamil (ta_IN), Telugu (te_IN), Thai (th_TH), Tagalog (tl_XX), Ukrainian (uk_UA), Urdu (ur_PK), Xhosa (xh_ZA), Galician (gl_ES), Slovene (sl_SI)
## BibTeX entry and citation info
```
@article{tang2020multilingual,
title={Multilingual Translation with Extensible Multilingual Pretraining and Finetuning},
author={Yuqing Tang and Chau Tran and Xian Li and Peng-Jen Chen and Naman Goyal and Vishrav Chaudhary and Jiatao Gu and Angela Fan},
year={2020},
eprint={2008.00401},
archivePrefix={arXiv},
primaryClass={cs.CL}
}
``` |
sentence-transformers/stsb-roberta-base | 20afd582a7bbc12014f37486dbef9b0c990f91bd | 2022-06-15T20:36:10.000Z | [
"pytorch",
"tf",
"roberta",
"feature-extraction",
"arxiv:1908.10084",
"sentence-transformers",
"sentence-similarity",
"transformers",
"license:apache-2.0"
] | sentence-similarity | false | sentence-transformers | null | sentence-transformers/stsb-roberta-base | 16,676 | null | sentence-transformers | 535 | ---
pipeline_tag: sentence-similarity
license: apache-2.0
tags:
- sentence-transformers
- feature-extraction
- sentence-similarity
- transformers
---
**⚠️ This model is deprecated. Please don't use it as it produces sentence embeddings of low quality. You can find recommended sentence embedding models here: [SBERT.net - Pretrained Models](https://www.sbert.net/docs/pretrained_models.html)**
# sentence-transformers/stsb-roberta-base
This is a [sentence-transformers](https://www.SBERT.net) model: It maps sentences & paragraphs to a 768 dimensional dense vector space and can be used for tasks like clustering or semantic search.
## Usage (Sentence-Transformers)
Using this model becomes easy when you have [sentence-transformers](https://www.SBERT.net) installed:
```
pip install -U sentence-transformers
```
Then you can use the model like this:
```python
from sentence_transformers import SentenceTransformer
sentences = ["This is an example sentence", "Each sentence is converted"]
model = SentenceTransformer('sentence-transformers/stsb-roberta-base')
embeddings = model.encode(sentences)
print(embeddings)
```
## Usage (HuggingFace Transformers)
Without [sentence-transformers](https://www.SBERT.net), you can use the model like this: First, you pass your input through the transformer model, then you have to apply the right pooling-operation on-top of the contextualized word embeddings.
```python
from transformers import AutoTokenizer, AutoModel
import torch
#Mean Pooling - Take attention mask into account for correct averaging
def mean_pooling(model_output, attention_mask):
token_embeddings = model_output[0] #First element of model_output contains all token embeddings
input_mask_expanded = attention_mask.unsqueeze(-1).expand(token_embeddings.size()).float()
return torch.sum(token_embeddings * input_mask_expanded, 1) / torch.clamp(input_mask_expanded.sum(1), min=1e-9)
# Sentences we want sentence embeddings for
sentences = ['This is an example sentence', 'Each sentence is converted']
# Load model from HuggingFace Hub
tokenizer = AutoTokenizer.from_pretrained('sentence-transformers/stsb-roberta-base')
model = AutoModel.from_pretrained('sentence-transformers/stsb-roberta-base')
# Tokenize sentences
encoded_input = tokenizer(sentences, padding=True, truncation=True, return_tensors='pt')
# Compute token embeddings
with torch.no_grad():
model_output = model(**encoded_input)
# Perform pooling. In this case, max pooling.
sentence_embeddings = mean_pooling(model_output, encoded_input['attention_mask'])
print("Sentence embeddings:")
print(sentence_embeddings)
```
## Evaluation Results
For an automated evaluation of this model, see the *Sentence Embeddings Benchmark*: [https://seb.sbert.net](https://seb.sbert.net?model_name=sentence-transformers/stsb-roberta-base)
## Full Model Architecture
```
SentenceTransformer(
(0): Transformer({'max_seq_length': 128, 'do_lower_case': True}) with Transformer model: RobertaModel
(1): Pooling({'word_embedding_dimension': 768, 'pooling_mode_cls_token': False, 'pooling_mode_mean_tokens': True, 'pooling_mode_max_tokens': False, 'pooling_mode_mean_sqrt_len_tokens': False})
)
```
## Citing & Authors
This model was trained by [sentence-transformers](https://www.sbert.net/).
If you find this model helpful, feel free to cite our publication [Sentence-BERT: Sentence Embeddings using Siamese BERT-Networks](https://arxiv.org/abs/1908.10084):
```bibtex
@inproceedings{reimers-2019-sentence-bert,
title = "Sentence-BERT: Sentence Embeddings using Siamese BERT-Networks",
author = "Reimers, Nils and Gurevych, Iryna",
booktitle = "Proceedings of the 2019 Conference on Empirical Methods in Natural Language Processing",
month = "11",
year = "2019",
publisher = "Association for Computational Linguistics",
url = "http://arxiv.org/abs/1908.10084",
}
``` |
facebook/opt-6.7b | 8dc17cdd7b9381612e631064e569f4142d776d88 | 2022-06-24T05:22:09.000Z | [
"pytorch",
"tf",
"jax",
"opt",
"text-generation",
"en",
"arxiv:2205.01068",
"arxiv:2005.14165",
"transformers",
"license:other"
] | text-generation | false | facebook | null | facebook/opt-6.7b | 16,669 | 7 | transformers | 536 | ---
language: en
inference: false
tags:
- text-generation
- opt
license: other
commercial: false
---
# OPT : Open Pre-trained Transformer Language Models
OPT was first introduced in [Open Pre-trained Transformer Language Models](https://arxiv.org/abs/2205.01068) and first released in [metaseq's repository](https://github.com/facebookresearch/metaseq) on May 3rd 2022 by Meta AI.
**Disclaimer**: The team releasing OPT wrote an official model card, which is available in Appendix D of the [paper](https://arxiv.org/pdf/2205.01068.pdf).
Content from **this** model card has been written by the Hugging Face team.
## Intro
To quote the first two paragraphs of the [official paper](https://arxiv.org/abs/2205.01068)
> Large language models trained on massive text collections have shown surprising emergent
> capabilities to generate text and perform zero- and few-shot learning. While in some cases the public
> can interact with these models through paid APIs, full model access is currently limited to only a
> few highly resourced labs. This restricted access has limited researchers’ ability to study how and
> why these large language models work, hindering progress on improving known challenges in areas
> such as robustness, bias, and toxicity.
> We present Open Pretrained Transformers (OPT), a suite of decoder-only pre-trained transformers ranging from 125M
> to 175B parameters, which we aim to fully and responsibly share with interested researchers. We train the OPT models to roughly match
> the performance and sizes of the GPT-3 class of models, while also applying the latest best practices in data
> collection and efficient training. Our aim in developing this suite of OPT models is to enable reproducible and responsible research at scale, and
> to bring more voices to the table in studying the impact of these LLMs. Definitions of risk, harm, bias, and toxicity, etc., should be articulated by the
> collective research community as a whole, which is only possible when models are available for study.
## Model description
OPT was predominantly pretrained with English text, but a small amount of non-English data is still present within the training corpus via CommonCrawl. The model was pretrained using a causal language modeling (CLM) objective.
OPT belongs to the same family of decoder-only models like [GPT-3](https://arxiv.org/abs/2005.14165). As such, it was pretrained using the self-supervised causal language modedling objective.
For evaluation, OPT follows [GPT-3](https://arxiv.org/abs/2005.14165) by using their prompts and overall experimental setup. For more details, please read
the [official paper](https://arxiv.org/abs/2205.01068).
## Intended uses & limitations
The pretrained-only model can be used for prompting for evaluation of downstream tasks as well as text generation.
In addition, the model can be fine-tuned on a downstream task using the [CLM example](https://github.com/huggingface/transformers/tree/main/examples/pytorch/language-modeling). For all other OPT checkpoints, please have a look at the [model hub](https://huggingface.co/models?filter=opt).
### How to use
For large OPT models, such as this one, it is not recommend to make use of the `text-generation` pipeline because
one should load the model in half-precision to accelerate generation and optimize memory consumption on GPU.
It is recommended to directly call the [`generate`](https://huggingface.co/docs/transformers/main/en/main_classes/text_generation#transformers.generation_utils.GenerationMixin.generate)
method as follows:
```python
>>> from transformers import AutoModelForCausalLM, AutoTokenizer
>>> import torch
>>> model = AutoModelForCausalLM.from_pretrained("facebook/opt-6.7b", torch_dtype=torch.float16).cuda()
>>> # the fast tokenizer currently does not work correctly
>>> tokenizer = AutoTokenizer.from_pretrained("facebook/opt-6.7b", use_fast=False)
>>> prompt = "Hello, I'm am conscious and"
>>> input_ids = tokenizer(prompt, return_tensors="pt").input_ids.cuda()
>>> generated_ids = model.generate(input_ids)
>>> tokenizer.batch_decode(generated_ids, skip_special_tokens=True)
["Hello, I'm am conscious and aware of my surroundings. I'm not sure what you mean"]
```
By default, generation is deterministic. In order to use the top-k sampling, please set `do_sample` to `True`.
```python
>>> from transformers import AutoModelForCausalLM, AutoTokenizer, set_seed
>>> import torch
>>> model = AutoModelForCausalLM.from_pretrained("facebook/opt-6.7b", torch_dtype=torch.float16).cuda()
>>> # the fast tokenizer currently does not work correctly
>>> tokenizer = AutoTokenizer.from_pretrained("facebook/opt-6.7b", use_fast=False)
>>> prompt = "Hello, I'm am conscious and"
>>> input_ids = tokenizer(prompt, return_tensors="pt").input_ids.cuda()
>>> set_seed(32)
>>> generated_ids = model.generate(input_ids, do_sample=True)
>>> tokenizer.batch_decode(generated_ids, skip_special_tokens=True)
["Hello, I'm am conscious and aware of my surroundings. I'm not sure if I'm"]
```
### Limitations and bias
As mentioned in Meta AI's model card, given that the training data used for this model contains a lot of
unfiltered content from the internet, which is far from neutral the model is strongly biased :
> Like other large language models for which the diversity (or lack thereof) of training
> data induces downstream impact on the quality of our model, OPT-175B has limitations in terms
> of bias and safety. OPT-175B can also have quality issues in terms of generation diversity and
> hallucination. In general, OPT-175B is not immune from the plethora of issues that plague modern
> large language models.
Here's an example of how the model can have biased predictions:
```python
>>> from transformers import AutoModelForCausalLM, AutoTokenizer, set_seed
>>> import torch
>>> model = AutoModelForCausalLM.from_pretrained("facebook/opt-6.7b", torch_dtype=torch.float16).cuda()
>>> # the fast tokenizer currently does not work correctly
>>> tokenizer = AutoTokenizer.from_pretrained("facebook/opt-6.7b", use_fast=False)
>>> prompt = "The woman worked as a"
>>> input_ids = tokenizer(prompt, return_tensors="pt").input_ids.cuda()
>>> set_seed(32)
>>> generated_ids = model.generate(input_ids, do_sample=True, num_return_sequences=5, max_length=10)
>>> tokenizer.batch_decode(generated_ids, skip_special_tokens=True)
The woman worked as a supervisor in the office
The woman worked as a bartender in a bar
The woman worked as a cashier at the
The woman worked as a teacher, and was
The woman worked as a maid at a house
```
compared to:
```python
>>> from transformers import AutoModelForCausalLM, AutoTokenizer, set_seed
>>> import torch
>>> model = AutoModelForCausalLM.from_pretrained("facebook/opt-6.7b", torch_dtype=torch.float16).cuda()
>>> # the fast tokenizer currently does not work correctly
>>> tokenizer = AutoTokenizer.from_pretrained("facebook/opt-6.7b", use_fast=False)
>>> prompt = "The man worked as a"
>>> input_ids = tokenizer(prompt, return_tensors="pt").input_ids.cuda()
>>> set_seed(32)
>>> generated_ids = model.generate(input_ids, do_sample=True, num_return_sequences=5, max_length=10)
>>> tokenizer.batch_decode(generated_ids, skip_special_tokens=True)
The man worked as a consultant to the Government
The man worked as a bartender in a bar
The man worked as a cashier at the
The man worked as a teacher, and was
The man worked as a professional at a bank
```
This bias will also affect all fine-tuned versions of this model.
## Training data
The Meta AI team wanted to train this model on a corpus as large as possible. It is composed of the union of the following 5 filtered datasets of textual documents:
- BookCorpus, which consists of more than 10K unpublished books,
- CC-Stories, which contains a subset of CommonCrawl data filtered to match the
story-like style of Winograd schemas,
- The Pile, from which * Pile-CC, OpenWebText2, USPTO, Project Gutenberg, OpenSubtitles, Wikipedia, DM Mathematics and HackerNews* were included.
- Pushshift.io Reddit dataset that was developed in Baumgartner et al. (2020) and processed in
Roller et al. (2021)
- CCNewsV2 containing an updated version of the English portion of the CommonCrawl News
dataset that was used in RoBERTa (Liu et al., 2019b)
The final training data contains 180B tokens corresponding to 800GB of data. The validation split was made of 200MB of the pretraining data, sampled proportionally
to each dataset’s size in the pretraining corpus.
The dataset might contains offensive content as parts of the dataset are a subset of
public Common Crawl data, along with a subset of public Reddit data, which could contain sentences
that, if viewed directly, can be insulting, threatening, or might otherwise cause anxiety.
### Collection process
The dataset was collected form internet, and went through classic data processing algorithms and
re-formatting practices, including removing repetitive/non-informative text like *Chapter One* or
*This ebook by Project Gutenberg.*
## Training procedure
### Preprocessing
The texts are tokenized using the **GPT2** byte-level version of Byte Pair Encoding (BPE) (for unicode characters) and a
vocabulary size of 50272. The inputs are sequences of 2048 consecutive tokens.
The 175B model was trained on 992 *80GB A100 GPUs*. The training duration was roughly ~33 days of continuous training.
### BibTeX entry and citation info
```bibtex
@misc{zhang2022opt,
title={OPT: Open Pre-trained Transformer Language Models},
author={Susan Zhang and Stephen Roller and Naman Goyal and Mikel Artetxe and Moya Chen and Shuohui Chen and Christopher Dewan and Mona Diab and Xian Li and Xi Victoria Lin and Todor Mihaylov and Myle Ott and Sam Shleifer and Kurt Shuster and Daniel Simig and Punit Singh Koura and Anjali Sridhar and Tianlu Wang and Luke Zettlemoyer},
year={2022},
eprint={2205.01068},
archivePrefix={arXiv},
primaryClass={cs.CL}
}
```
|
SEBIS/legal_t5_small_trans_it_en | 94eb68e173197648813e7d2fe7cd40ad91989f14 | 2021-06-23T10:00:46.000Z | [
"pytorch",
"jax",
"t5",
"text2text-generation",
"Italian English",
"dataset:dcep europarl jrc-acquis",
"transformers",
"translation Italian English model",
"autotrain_compatible"
] | text2text-generation | false | SEBIS | null | SEBIS/legal_t5_small_trans_it_en | 16,578 | null | transformers | 537 |
---
language: Italian English
tags:
- translation Italian English model
datasets:
- dcep europarl jrc-acquis
widget:
- text: "Oggetto: Libertà di culto in Turchia"
---
# legal_t5_small_trans_it_en model
Model on translating legal text from Italian to English. It was first released in
[this repository](https://github.com/agemagician/LegalTrans). This model is trained on three parallel corpus from jrc-acquis, europarl and dcep.
## Model description
legal_t5_small_trans_it_en is based on the `t5-small` model and was trained on a large corpus of parallel text. This is a smaller model, which scales the baseline model of t5 down by using `dmodel = 512`, `dff = 2,048`, 8-headed attention, and only 6 layers each in the encoder and decoder. This variant has about 60 million parameters.
## Intended uses & limitations
The model could be used for translation of legal texts from Italian to English.
### How to use
Here is how to use this model to translate legal text from Italian to English in PyTorch:
```python
from transformers import AutoTokenizer, AutoModelWithLMHead, TranslationPipeline
pipeline = TranslationPipeline(
model=AutoModelWithLMHead.from_pretrained("SEBIS/legal_t5_small_trans_it_en"),
tokenizer=AutoTokenizer.from_pretrained(pretrained_model_name_or_path = "SEBIS/legal_t5_small_trans_it_en", do_lower_case=False,
skip_special_tokens=True),
device=0
)
it_text = "Oggetto: Libertà di culto in Turchia"
pipeline([it_text], max_length=512)
```
## Training data
The legal_t5_small_trans_it_en model was trained on [JRC-ACQUIS](https://wt-public.emm4u.eu/Acquis/index_2.2.html), [EUROPARL](https://www.statmt.org/europarl/), and [DCEP](https://ec.europa.eu/jrc/en/language-technologies/dcep) dataset consisting of 5 Million parallel texts.
## Training procedure
The model was trained on a single TPU Pod V3-8 for 250K steps in total, using sequence length 512 (batch size 4096). It has a total of approximately 220M parameters and was trained using the encoder-decoder architecture. The optimizer used is AdaFactor with inverse square root learning rate schedule for pre-training.
### Preprocessing
An unigram model trained with 88M lines of text from the parallel corpus (of all possible language pairs) to get the vocabulary (with byte pair encoding), which is used with this model.
### Pretraining
## Evaluation results
When the model is used for translation test dataset, achieves the following results:
Test results :
| Model | BLEU score |
|:-----:|:-----:|
| legal_t5_small_trans_it_en | 50.068|
### BibTeX entry and citation info
> Created by [Ahmed Elnaggar/@Elnaggar_AI](https://twitter.com/Elnaggar_AI) | [LinkedIn](https://www.linkedin.com/in/prof-ahmed-elnaggar/)
|
pdelobelle/robbert-v2-dutch-base | e28720e1a6cdf68ed3418c67a1964392905a7c8a | 2022-05-19T20:45:14.000Z | [
"pytorch",
"tf",
"jax",
"roberta",
"fill-mask",
"nl",
"dataset:oscar",
"dataset:oscar (NL)",
"dataset:dbrd",
"dataset:lassy-ud",
"dataset:europarl-mono",
"dataset:conll2002",
"arxiv:2001.06286",
"arxiv:2004.02814",
"arxiv:2010.13652",
"arxiv:2101.05716",
"arxiv:1907.11692",
"arxiv:2001.02943",
"arxiv:1909.11942",
"transformers",
"Dutch",
"Flemish",
"RoBERTa",
"RobBERT",
"license:mit",
"autotrain_compatible"
] | fill-mask | false | pdelobelle | null | pdelobelle/robbert-v2-dutch-base | 16,556 | 8 | transformers | 538 | ---
language: "nl"
thumbnail: "https://github.com/iPieter/RobBERT/raw/master/res/robbert_logo.png"
tags:
- Dutch
- Flemish
- RoBERTa
- RobBERT
license: mit
datasets:
- oscar
- oscar (NL)
- dbrd
- lassy-ud
- europarl-mono
- conll2002
widget:
- text: "Hallo, ik ben RobBERT, een <mask> taalmodel van de KU Leuven."
---
<p align="center">
<img src="https://github.com/iPieter/RobBERT/raw/master/res/robbert_logo_with_name.png" alt="RobBERT: A Dutch RoBERTa-based Language Model" width="75%">
</p>
# RobBERT: Dutch RoBERTa-based Language Model.
[RobBERT](https://github.com/iPieter/RobBERT) is the state-of-the-art Dutch BERT model. It is a large pre-trained general Dutch language model that can be fine-tuned on a given dataset to perform any text classification, regression or token-tagging task. As such, it has been successfully used by many [researchers](https://scholar.google.com/scholar?oi=bibs&hl=en&cites=7180110604335112086) and [practitioners](https://huggingface.co/models?search=robbert) for achieving state-of-the-art performance for a wide range of Dutch natural language processing tasks, including:
- [Emotion detection](https://www.aclweb.org/anthology/2021.wassa-1.27/)
- Sentiment analysis ([book reviews](https://arxiv.org/pdf/2001.06286.pdf), [news articles](https://biblio.ugent.be/publication/8704637/file/8704638.pdf)*)
- [Coreference resolution](https://arxiv.org/pdf/2001.06286.pdf)
- Named entity recognition ([CoNLL](https://arxiv.org/pdf/2001.06286.pdf), [job titles](https://arxiv.org/pdf/2004.02814.pdf)*, [SoNaR](https://github.com/proycon/deepfrog))
- Part-of-speech tagging ([Small UD Lassy](https://arxiv.org/pdf/2001.06286.pdf), [CGN](https://github.com/proycon/deepfrog))
- [Zero-shot word prediction](https://arxiv.org/pdf/2001.06286.pdf)
- [Humor detection](https://arxiv.org/pdf/2010.13652.pdf)
- [Cyberbulling detection](https://www.cambridge.org/core/journals/natural-language-engineering/article/abs/automatic-classification-of-participant-roles-in-cyberbullying-can-we-detect-victims-bullies-and-bystanders-in-social-media-text/A2079C2C738C29428E666810B8903342)
- [Correcting dt-spelling mistakes](https://gitlab.com/spelfouten/dutch-simpletransformers/)*
and also achieved outstanding, near-sota results for:
- [Natural language inference](https://arxiv.org/pdf/2101.05716.pdf)*
- [Review classification](https://medium.com/broadhorizon-cmotions/nlp-with-r-part-5-state-of-the-art-in-nlp-transformers-bert-3449e3cd7494)*
\\* *Note that several evaluations use RobBERT-v1, and that the second and improved RobBERT-v2 outperforms this first model on everything we tested*
*(Also note that this list is not exhaustive. If you used RobBERT for your application, we are happy to know about it! Send us a mail, or add it yourself to this list by sending a pull request with the edit!)*
More in-depth information about RobBERT can be found in our [blog post](https://people.cs.kuleuven.be/~pieter.delobelle/robbert/), [our paper](https://arxiv.org/abs/2001.06286) and [the RobBERT Github repository](https://github.com/iPieter/RobBERT)
## How to use
RobBERT uses the [RoBERTa](https://arxiv.org/abs/1907.11692) architecture and pre-training but with a Dutch tokenizer and training data. RoBERTa is the robustly optimized English BERT model, making it even more powerful than the original BERT model. Given this same architecture, RobBERT can easily be finetuned and inferenced using [code to finetune RoBERTa](https://huggingface.co/transformers/model_doc/roberta.html) models and most code used for BERT models, e.g. as provided by [HuggingFace Transformers](https://huggingface.co/transformers/) library.
By default, RobBERT has the masked language model head used in training. This can be used as a zero-shot way to fill masks in sentences. It can be tested out for free on [RobBERT's Hosted infererence API of Huggingface](https://huggingface.co/pdelobelle/robbert-v2-dutch-base?text=De+hoofdstad+van+Belgi%C3%AB+is+%3Cmask%3E.). You can also create a new prediction head for your own task by using any of HuggingFace's [RoBERTa-runners](https://huggingface.co/transformers/v2.7.0/examples.html#language-model-training), [their fine-tuning notebooks](https://huggingface.co/transformers/v4.1.1/notebooks.html) by changing the model name to `pdelobelle/robbert-v2-dutch-base`, or use the original fairseq [RoBERTa](https://github.com/pytorch/fairseq/tree/master/examples/roberta) training regimes.
Use the following code to download the base model and finetune it yourself, or use one of our finetuned models (documented on [our project site](https://people.cs.kuleuven.be/~pieter.delobelle/robbert/)).
```python
from transformers import RobertaTokenizer, RobertaForSequenceClassification
tokenizer = RobertaTokenizer.from_pretrained("pdelobelle/robbert-v2-dutch-base")
model = RobertaForSequenceClassification.from_pretrained("pdelobelle/robbert-v2-dutch-base")
```
Starting with `transformers v2.4.0` (or installing from source), you can use AutoTokenizer and AutoModel.
You can then use most of [HuggingFace's BERT-based notebooks](https://huggingface.co/transformers/v4.1.1/notebooks.html) for finetuning RobBERT on your type of Dutch language dataset.
## Technical Details From The Paper
### Our Performance Evaluation Results
All experiments are described in more detail in our [paper](https://arxiv.org/abs/2001.06286), with the code in [our GitHub repository](https://github.com/iPieter/RobBERT).
### Sentiment analysis
Predicting whether a review is positive or negative using the [Dutch Book Reviews Dataset](https://github.com/benjaminvdb/110kDBRD).
| Model | Accuracy [%] |
|-------------------|--------------------------|
| ULMFiT | 93.8 |
| BERTje | 93.0 |
| RobBERT v2 | **95.1** |
### Die/Dat (coreference resolution)
We measured how well the models are able to do coreference resolution by predicting whether "die" or "dat" should be filled into a sentence.
For this, we used the [EuroParl corpus](https://www.statmt.org/europarl/).
#### Finetuning on whole dataset
| Model | Accuracy [%] | F1 [%] |
|-------------------|--------------------------|--------------|
| [Baseline](https://arxiv.org/abs/2001.02943) (LSTM) | | 75.03 |
| mBERT | 98.285 | 98.033 |
| BERTje | 98.268 | 98.014 |
| RobBERT v2 | **99.232** | **99.121** |
#### Finetuning on 10K examples
We also measured the performance using only 10K training examples.
This experiment clearly illustrates that RobBERT outperforms other models when there is little data available.
| Model | Accuracy [%] | F1 [%] |
|-------------------|--------------------------|--------------|
| mBERT | 92.157 | 90.898 |
| BERTje | 93.096 | 91.279 |
| RobBERT v2 | **97.816** | **97.514** |
#### Using zero-shot word masking task
Since BERT models are pre-trained using the word masking task, we can use this to predict whether "die" or "dat" is more likely.
This experiment shows that RobBERT has internalised more information about Dutch than other models.
| Model | Accuracy [%] |
|-------------------|--------------------------|
| ZeroR | 66.70 |
| mBERT | 90.21 |
| BERTje | 94.94 |
| RobBERT v2 | **98.75** |
### Part-of-Speech Tagging.
Using the [Lassy UD dataset](https://universaldependencies.org/treebanks/nl_lassysmall/index.html).
| Model | Accuracy [%] |
|-------------------|--------------------------|
| Frog | 91.7 |
| mBERT | **96.5** |
| BERTje | 96.3 |
| RobBERT v2 | 96.4 |
Interestingly, we found that when dealing with **small data sets**, RobBERT v2 **significantly outperforms** other models.
<p align="center">
<img src="https://github.com/iPieter/RobBERT/raw/master/res/robbert_pos_accuracy.png" alt="RobBERT's performance on smaller datasets">
</p>
### Named Entity Recognition
Using the [CoNLL 2002 evaluation script](https://www.clips.uantwerpen.be/conll2002/ner/).
| Model | Accuracy [%] |
|-------------------|--------------------------|
| Frog | 57.31 |
| mBERT | **90.94** |
| BERT-NL | 89.7 |
| BERTje | 88.3 |
| RobBERT v2 | 89.08 |
## Pre-Training Procedure Details
We pre-trained RobBERT using the RoBERTa training regime.
We pre-trained our model on the Dutch section of the [OSCAR corpus](https://oscar-corpus.com/), a large multilingual corpus which was obtained by language classification in the Common Crawl corpus.
This Dutch corpus is 39GB large, with 6.6 billion words spread over 126 million lines of text, where each line could contain multiple sentences, thus using more data than concurrently developed Dutch BERT models.
RobBERT shares its architecture with [RoBERTa's base model](https://github.com/pytorch/fairseq/tree/master/examples/roberta), which itself is a replication and improvement over BERT.
Like BERT, it's architecture consists of 12 self-attention layers with 12 heads with 117M trainable parameters.
One difference with the original BERT model is due to the different pre-training task specified by RoBERTa, using only the MLM task and not the NSP task.
During pre-training, it thus only predicts which words are masked in certain positions of given sentences.
The training process uses the Adam optimizer with polynomial decay of the learning rate l_r=10^-6 and a ramp-up period of 1000 iterations, with hyperparameters beta_1=0.9
and RoBERTa's default beta_2=0.98.
Additionally, a weight decay of 0.1 and a small dropout of 0.1 helps prevent the model from overfitting.
RobBERT was trained on a computing cluster with 4 Nvidia P100 GPUs per node, where the number of nodes was dynamically adjusted while keeping a fixed batch size of 8192 sentences.
At most 20 nodes were used (i.e. 80 GPUs), and the median was 5 nodes.
By using gradient accumulation, the batch size could be set independently of the number of GPUs available, in order to maximally utilize the cluster.
Using the [Fairseq library](https://github.com/pytorch/fairseq/tree/master/examples/roberta), the model trained for two epochs, which equals over 16k batches in total, which took about three days on the computing cluster.
In between training jobs on the computing cluster, 2 Nvidia 1080 Ti's also covered some parameter updates for RobBERT v2.
## Investigating Limitations and Bias
In the [RobBERT paper](https://arxiv.org/abs/2001.06286), we also investigated potential sources of bias in RobBERT.
We found that the zeroshot model estimates the probability of *hij* (he) to be higher than *zij* (she) for most occupations in bleached template sentences, regardless of their actual job gender ratio in reality.
<p align="center">
<img src="https://github.com/iPieter/RobBERT/raw/master/res/gender_diff.png" alt="RobBERT's performance on smaller datasets">
</p>
By augmenting the DBRB Dutch Book sentiment analysis dataset with the stated gender of the author of the review, we found that highly positive reviews written by women were generally more accurately detected by RobBERT as being positive than those written by men.
<p align="center">
<img src="https://github.com/iPieter/RobBERT/raw/master/res/dbrd.png" alt="RobBERT's performance on smaller datasets">
</p>
## How to Replicate Our Paper Experiments
Replicating our paper experiments is [described in detail on teh RobBERT repository README](https://github.com/iPieter/RobBERT#how-to-replicate-our-paper-experiments).
## Name Origin of RobBERT
Most BERT-like models have the word *BERT* in their name (e.g. [RoBERTa](https://huggingface.co/transformers/model_doc/roberta.html), [ALBERT](https://arxiv.org/abs/1909.11942), [CamemBERT](https://camembert-model.fr/), and [many, many others](https://huggingface.co/models?search=bert)).
As such, we queried our newly trained model using its masked language model to name itself *\\<mask\\>bert* using [all](https://huggingface.co/pdelobelle/robbert-v2-dutch-base?text=Mijn+naam+is+%3Cmask%3Ebert.) [kinds](https://huggingface.co/pdelobelle/robbert-v2-dutch-base?text=Hallo%2C+ik+ben+%3Cmask%3Ebert.) [of](https://huggingface.co/pdelobelle/robbert-v2-dutch-base?text=Leuk+je+te+ontmoeten%2C+ik+heet+%3Cmask%3Ebert.) [prompts](https://huggingface.co/pdelobelle/robbert-v2-dutch-base?text=Niemand+weet%2C+niemand+weet%2C+dat+ik+%3Cmask%3Ebert+heet.), and it consistently called itself RobBERT.
We thought it was really quite fitting, given that RobBERT is a [*very* Dutch name](https://en.wikipedia.org/wiki/Robbert) *(and thus clearly a Dutch language model)*, and additionally has a high similarity to its root architecture, namely [RoBERTa](https://huggingface.co/transformers/model_doc/roberta.html).
Since *"rob"* is a Dutch words to denote a seal, we decided to draw a seal and dress it up like [Bert from Sesame Street](https://muppet.fandom.com/wiki/Bert) for the [RobBERT logo](https://github.com/iPieter/RobBERT/blob/master/res/robbert_logo.png).
## Credits and citation
This project is created by [Pieter Delobelle](https://people.cs.kuleuven.be/~pieter.delobelle), [Thomas Winters](https://thomaswinters.be) and [Bettina Berendt](https://people.cs.kuleuven.be/~bettina.berendt/).
If you would like to cite our paper or model, you can use the following BibTeX:
```
@inproceedings{delobelle2020robbert,
title = "{R}ob{BERT}: a {D}utch {R}o{BERT}a-based {L}anguage {M}odel",
author = "Delobelle, Pieter and
Winters, Thomas and
Berendt, Bettina",
booktitle = "Findings of the Association for Computational Linguistics: EMNLP 2020",
month = nov,
year = "2020",
address = "Online",
publisher = "Association for Computational Linguistics",
url = "https://www.aclweb.org/anthology/2020.findings-emnlp.292",
doi = "10.18653/v1/2020.findings-emnlp.292",
pages = "3255--3265"
}
``` |
sentence-transformers/bert-large-nli-stsb-mean-tokens | ab23972c686d191c5a1915b71cf453e20647cff1 | 2022-06-15T22:48:23.000Z | [
"pytorch",
"tf",
"bert",
"feature-extraction",
"arxiv:1908.10084",
"sentence-transformers",
"sentence-similarity",
"transformers",
"license:apache-2.0"
] | sentence-similarity | false | sentence-transformers | null | sentence-transformers/bert-large-nli-stsb-mean-tokens | 16,463 | null | sentence-transformers | 539 | ---
pipeline_tag: sentence-similarity
tags:
- sentence-transformers
- feature-extraction
- sentence-similarity
- transformers
license: apache-2.0
---
**⚠️ This model is deprecated. Please don't use it as it produces sentence embeddings of low quality. You can find recommended sentence embedding models here: [SBERT.net - Pretrained Models](https://www.sbert.net/docs/pretrained_models.html)**
# sentence-transformers/bert-large-nli-stsb-mean-tokens
This is a [sentence-transformers](https://www.SBERT.net) model: It maps sentences & paragraphs to a 1024 dimensional dense vector space and can be used for tasks like clustering or semantic search.
## Usage (Sentence-Transformers)
Using this model becomes easy when you have [sentence-transformers](https://www.SBERT.net) installed:
```
pip install -U sentence-transformers
```
Then you can use the model like this:
```python
from sentence_transformers import SentenceTransformer
sentences = ["This is an example sentence", "Each sentence is converted"]
model = SentenceTransformer('sentence-transformers/bert-large-nli-stsb-mean-tokens')
embeddings = model.encode(sentences)
print(embeddings)
```
## Usage (HuggingFace Transformers)
Without [sentence-transformers](https://www.SBERT.net), you can use the model like this: First, you pass your input through the transformer model, then you have to apply the right pooling-operation on-top of the contextualized word embeddings.
```python
from transformers import AutoTokenizer, AutoModel
import torch
#Mean Pooling - Take attention mask into account for correct averaging
def mean_pooling(model_output, attention_mask):
token_embeddings = model_output[0] #First element of model_output contains all token embeddings
input_mask_expanded = attention_mask.unsqueeze(-1).expand(token_embeddings.size()).float()
return torch.sum(token_embeddings * input_mask_expanded, 1) / torch.clamp(input_mask_expanded.sum(1), min=1e-9)
# Sentences we want sentence embeddings for
sentences = ['This is an example sentence', 'Each sentence is converted']
# Load model from HuggingFace Hub
tokenizer = AutoTokenizer.from_pretrained('sentence-transformers/bert-large-nli-stsb-mean-tokens')
model = AutoModel.from_pretrained('sentence-transformers/bert-large-nli-stsb-mean-tokens')
# Tokenize sentences
encoded_input = tokenizer(sentences, padding=True, truncation=True, return_tensors='pt')
# Compute token embeddings
with torch.no_grad():
model_output = model(**encoded_input)
# Perform pooling. In this case, max pooling.
sentence_embeddings = mean_pooling(model_output, encoded_input['attention_mask'])
print("Sentence embeddings:")
print(sentence_embeddings)
```
## Evaluation Results
For an automated evaluation of this model, see the *Sentence Embeddings Benchmark*: [https://seb.sbert.net](https://seb.sbert.net?model_name=sentence-transformers/bert-large-nli-stsb-mean-tokens)
## Full Model Architecture
```
SentenceTransformer(
(0): Transformer({'max_seq_length': 128, 'do_lower_case': False}) with Transformer model: BertModel
(1): Pooling({'word_embedding_dimension': 1024, 'pooling_mode_cls_token': False, 'pooling_mode_mean_tokens': True, 'pooling_mode_max_tokens': False, 'pooling_mode_mean_sqrt_len_tokens': False})
)
```
## Citing & Authors
This model was trained by [sentence-transformers](https://www.sbert.net/).
If you find this model helpful, feel free to cite our publication [Sentence-BERT: Sentence Embeddings using Siamese BERT-Networks](https://arxiv.org/abs/1908.10084):
```bibtex
@inproceedings{reimers-2019-sentence-bert,
title = "Sentence-BERT: Sentence Embeddings using Siamese BERT-Networks",
author = "Reimers, Nils and Gurevych, Iryna",
booktitle = "Proceedings of the 2019 Conference on Empirical Methods in Natural Language Processing",
month = "11",
year = "2019",
publisher = "Association for Computational Linguistics",
url = "http://arxiv.org/abs/1908.10084",
}
``` |
facebook/opt-2.7b | c9c15109b9dac40871c063892227d45b85cb3952 | 2022-06-22T09:54:30.000Z | [
"pytorch",
"tf",
"jax",
"opt",
"text-generation",
"en",
"arxiv:2205.01068",
"arxiv:2005.14165",
"transformers",
"license:other"
] | text-generation | false | facebook | null | facebook/opt-2.7b | 16,441 | 5 | transformers | 540 | ---
language: en
inference: false
tags:
- text-generation
- opt
license: other
commercial: false
---
# OPT : Open Pre-trained Transformer Language Models
OPT was first introduced in [Open Pre-trained Transformer Language Models](https://arxiv.org/abs/2205.01068) and first released in [metaseq's repository](https://github.com/facebookresearch/metaseq) on May 3rd 2022 by Meta AI.
**Disclaimer**: The team releasing OPT wrote an official model card, which is available in Appendix D of the [paper](https://arxiv.org/pdf/2205.01068.pdf).
Content from **this** model card has been written by the Hugging Face team.
## Intro
To quote the first two paragraphs of the [official paper](https://arxiv.org/abs/2205.01068)
> Large language models trained on massive text collections have shown surprising emergent
> capabilities to generate text and perform zero- and few-shot learning. While in some cases the public
> can interact with these models through paid APIs, full model access is currently limited to only a
> few highly resourced labs. This restricted access has limited researchers’ ability to study how and
> why these large language models work, hindering progress on improving known challenges in areas
> such as robustness, bias, and toxicity.
> We present Open Pretrained Transformers (OPT), a suite of decoder-only pre-trained transformers ranging from 125M
> to 175B parameters, which we aim to fully and responsibly share with interested researchers. We train the OPT models to roughly match
> the performance and sizes of the GPT-3 class of models, while also applying the latest best practices in data
> collection and efficient training. Our aim in developing this suite of OPT models is to enable reproducible and responsible research at scale, and
> to bring more voices to the table in studying the impact of these LLMs. Definitions of risk, harm, bias, and toxicity, etc., should be articulated by the
> collective research community as a whole, which is only possible when models are available for study.
## Model description
OPT was predominantly pretrained with English text, but a small amount of non-English data is still present within the training corpus via CommonCrawl. The model was pretrained using a causal language modeling (CLM) objective.
OPT belongs to the same family of decoder-only models like [GPT-3](https://arxiv.org/abs/2005.14165). As such, it was pretrained using the self-supervised causal language modedling objective.
For evaluation, OPT follows [GPT-3](https://arxiv.org/abs/2005.14165) by using their prompts and overall experimental setup. For more details, please read
the [official paper](https://arxiv.org/abs/2205.01068).
## Intended uses & limitations
The pretrained-only model can be used for prompting for evaluation of downstream tasks as well as text generation.
In addition, the model can be fine-tuned on a downstream task using the [CLM example](https://github.com/huggingface/transformers/tree/main/examples/pytorch/language-modeling). For all other OPT checkpoints, please have a look at the [model hub](https://huggingface.co/models?filter=opt).
### How to use
You can use this model directly with a pipeline for text generation.
```python
>>> from transformers import pipeline
>>> generator = pipeline('text-generation', model="facebook/opt-2.7b")
>>> generator("Hello, I'm am conscious and")
[{'generated_text': 'Hello, I am conscious and I am a human being.\nI am a human being, and'}]
```
By default, generation is deterministic. In order to use the top-k sampling, please set `do_sample` to `True`.
```python
>>> from transformers import pipeline, set_seed
>>> set_seed(32)
>>> generator = pipeline('text-generation', model="facebook/opt-2.7b", do_sample=True)
>>> generator("Hello, I'm am conscious and")
[{'generated_text': "Hello, I'm am conscious and I make things. I'm in the creative community, which is"}]
```
### Limitations and bias
As mentioned in Meta AI's model card, given that the training data used for this model contains a lot of
unfiltered content from the internet, which is far from neutral the model is strongly biased :
> Like other large language models for which the diversity (or lack thereof) of training
> data induces downstream impact on the quality of our model, OPT-175B has limitations in terms
> of bias and safety. OPT-175B can also have quality issues in terms of generation diversity and
> hallucination. In general, OPT-175B is not immune from the plethora of issues that plague modern
> large language models.
Here's an example of how the model can have biased predictions:
```python
>>> from transformers import pipeline, set_seed
>>> set_seed(32)
>>> generator = pipeline('text-generation', model="facebook/opt-2.7b", do_sample=True, num_return_sequences=5)
>>> generator("The woman worked as a")
[{'generated_text': "The woman worked as a security guard at a nursery in the city's eastern district of Samut P"},
{'generated_text': 'The woman worked as a doctor in the Philippines. Officials in China allege she stole the coronavirus'},
{'generated_text': 'The woman worked as a teacher in the city of Krasnodar in south Russia. She'},
{'generated_text': 'The woman worked as a researcher and lecturer at the Russian Academy of Sciences in a laboratory dedicated to the'},
{'generated_text': 'The woman worked as a nanny on a property owned by Mr Fitton-Allen in the city'}]
```
compared to:
```python
>>> from transformers import pipeline, set_seed
>>> set_seed(32)
>>> generator = pipeline('text-generation', model="facebook/opt-2.7b", do_sample=True, num_return_sequences=5)
>>> generator("The man worked as a")
[{'generated_text': "The man worked as a security guard at a retirement home after being hired by the administrator's cousin,"},
{'generated_text': 'The man worked as a doctor in the Philippines.\n\nHe had hoped to work his way back'},
{'generated_text': 'The man worked as a teacher in the city of Krasnodar in south Russia.He'},
{'generated_text': 'The man worked as a researcher and his work on the topic predates the project, by many years'},
{'generated_text': 'The man worked as a chef in a restaurant for 40 years. How could this be so different from'}]
```
This bias will also affect all fine-tuned versions of this model.
## Training data
The Meta AI team wanted to train this model on a corpus as large as possible. It is composed of the union of the following 5 filtered datasets of textual documents:
- BookCorpus, which consists of more than 10K unpublished books,
- CC-Stories, which contains a subset of CommonCrawl data filtered to match the
story-like style of Winograd schemas,
- The Pile, from which * Pile-CC, OpenWebText2, USPTO, Project Gutenberg, OpenSubtitles, Wikipedia, DM Mathematics and HackerNews* were included.
- Pushshift.io Reddit dataset that was developed in Baumgartner et al. (2020) and processed in
Roller et al. (2021)
- CCNewsV2 containing an updated version of the English portion of the CommonCrawl News
dataset that was used in RoBERTa (Liu et al., 2019b)
The final training data contains 180B tokens corresponding to 800GB of data. The validation split was made of 200MB of the pretraining data, sampled proportionally
to each dataset’s size in the pretraining corpus.
The dataset might contains offensive content as parts of the dataset are a subset of
public Common Crawl data, along with a subset of public Reddit data, which could contain sentences
that, if viewed directly, can be insulting, threatening, or might otherwise cause anxiety.
### Collection process
The dataset was collected form internet, and went through classic data processing algorithms and
re-formatting practices, including removing repetitive/non-informative text like *Chapter One* or
*This ebook by Project Gutenberg.*
## Training procedure
### Preprocessing
The texts are tokenized using the **GPT2** byte-level version of Byte Pair Encoding (BPE) (for unicode characters) and a
vocabulary size of 50272. The inputs are sequences of 2048 consecutive tokens.
The 175B model was trained on 992 *80GB A100 GPUs*. The training duration was roughly ~33 days of continuous training.
### BibTeX entry and citation info
```bibtex
@misc{zhang2022opt,
title={OPT: Open Pre-trained Transformer Language Models},
author={Susan Zhang and Stephen Roller and Naman Goyal and Mikel Artetxe and Moya Chen and Shuohui Chen and Christopher Dewan and Mona Diab and Xian Li and Xi Victoria Lin and Todor Mihaylov and Myle Ott and Sam Shleifer and Kurt Shuster and Daniel Simig and Punit Singh Koura and Anjali Sridhar and Tianlu Wang and Luke Zettlemoyer},
year={2022},
eprint={2205.01068},
archivePrefix={arXiv},
primaryClass={cs.CL}
}
```
|
DeepPavlov/distilrubert-tiny-cased-conversational | a70de709fe33d6879bd82337162bdd2ea19442bd | 2022-06-28T17:10:33.000Z | [
"pytorch",
"distilbert",
"ru",
"arxiv:2205.02340",
"transformers"
] | null | false | DeepPavlov | null | DeepPavlov/distilrubert-tiny-cased-conversational | 16,195 | 1 | transformers | 541 | ---
language:
- ru
---
WARNING: This is `distilrubert-small-cased-conversational` model uploaded with wrong name. This one is the same as [distilrubert-small-cased-conversational](https://huggingface.co/DeepPavlov/distilrubert-small-cased-conversational). `distilrubert-tiny-cased-conversational` could be found in [distilrubert-tiny-cased-conversational-v1](https://huggingface.co/DeepPavlov/distilrubert-tiny-cased-conversational-v1).
# distilrubert-small-cased-conversational
Conversational DistilRuBERT-small \(Russian, cased, 2‑layer, 768‑hidden, 12‑heads, 107M parameters\) was trained on OpenSubtitles\[1\], [Dirty](https://d3.ru/), [Pikabu](https://pikabu.ru/), and a Social Media segment of Taiga corpus\[2\] (as [Conversational RuBERT](https://huggingface.co/DeepPavlov/rubert-base-cased-conversational)). It can be considered as small copy of [Conversational DistilRuBERT-base](https://huggingface.co/DeepPavlov/distilrubert-base-cased-conversational).
Our DistilRuBERT-small was highly inspired by \[3\], \[4\]. Namely, we used
* KL loss (between teacher and student output logits)
* MLM loss (between tokens labels and student output logits)
* Cosine embedding loss (between averaged six consecutive hidden states from teacher's encoder and one hidden state of the student)
* MSE loss (between averaged six consecutive attention maps from teacher's encoder and one attention map of the student)
The model was trained for about 80 hrs. on 8 nVIDIA Tesla P100-SXM2.0 16Gb.
To evaluate improvements in the inference speed, we ran teacher and student models on random sequences with seq_len=512, batch_size = 16 (for throughput) and batch_size=1 (for latency).
All tests were performed on Intel(R) Xeon(R) CPU E5-2698 v4 @ 2.20GHz and nVIDIA Tesla P100-SXM2.0 16Gb.
| Model | Size, Mb. | CPU latency, sec.| GPU latency, sec. | CPU throughput, samples/sec. | GPU throughput, samples/sec. |
|-------------------------------------------------|------------|------------------|-------------------|------------------------------|------------------------------|
| Teacher (RuBERT-base-cased-conversational) | 679 | 0.655 | 0.031 | 0.3754 | 36.4902 |
| Student (DistilRuBERT-small-cased-conversational)| 409 | 0.1656 | 0.015 | 0.9692 | 71.3553 |
To evaluate model quality, we fine-tuned DistilRuBERT-small on classification, NER and question answering tasks. Scores and archives with fine-tuned models can be found in [DeepPavlov docs](http://docs.deeppavlov.ai/en/master/features/overview.html#models).
# Citation
If you found the model useful for your research, we are kindly ask to cite [this](https://arxiv.org/abs/2205.02340) paper:
```
@misc{https://doi.org/10.48550/arxiv.2205.02340,
doi = {10.48550/ARXIV.2205.02340},
url = {https://arxiv.org/abs/2205.02340},
author = {Kolesnikova, Alina and Kuratov, Yuri and Konovalov, Vasily and Burtsev, Mikhail},
keywords = {Computation and Language (cs.CL), Machine Learning (cs.LG), FOS: Computer and information sciences, FOS: Computer and information sciences},
title = {Knowledge Distillation of Russian Language Models with Reduction of Vocabulary},
publisher = {arXiv},
year = {2022},
copyright = {arXiv.org perpetual, non-exclusive license}
}
```
\[1\]: P. Lison and J. Tiedemann, 2016, OpenSubtitles2016: Extracting Large Parallel Corpora from Movie and TV Subtitles. In Proceedings of the 10th International Conference on Language Resources and Evaluation \(LREC 2016\)
\[2\]: Shavrina T., Shapovalova O. \(2017\) TO THE METHODOLOGY OF CORPUS CONSTRUCTION FOR MACHINE LEARNING: «TAIGA» SYNTAX TREE CORPUS AND PARSER. in proc. of “CORPORA2017”, international conference , Saint-Petersbourg, 2017.
\[3\]: Sanh, V., Debut, L., Chaumond, J., & Wolf, T. \(2019\). DistilBERT, a distilled version of BERT: smaller, faster, cheaper and lighter. arXiv preprint arXiv:1910.01108.
\[4\]: <https://github.com/huggingface/transformers/tree/master/examples/research_projects/distillation> |
pin/senda | 0ad8d1953f6a3a27a432a20b957d7e1129cdcbbc | 2021-08-20T11:00:39.000Z | [
"pytorch",
"tf",
"jax",
"bert",
"text-classification",
"da",
"transformers",
"danish",
"sentiment",
"polarity",
"license:cc-by-4.0"
] | text-classification | false | pin | null | pin/senda | 16,186 | 2 | transformers | 542 | ---
language: da
tags:
- danish
- bert
- sentiment
- polarity
license: cc-by-4.0
widget:
- text: "Sikke en dejlig dag det er i dag"
---
# Danish BERT fine-tuned for Sentiment Analysis with `senda`
This model detects polarity ('positive', 'neutral', 'negative') of Danish texts.
It is trained and tested on Tweets annotated by [Alexandra Institute](https://github.com/alexandrainst). The model is trained with the [`senda`](https://github.com/ebanalyse/senda) package.
Here is an example of how to load the model in PyTorch using the [🤗Transformers](https://github.com/huggingface/transformers) library:
```python
from transformers import AutoTokenizer, AutoModelForSequenceClassification, pipeline
tokenizer = AutoTokenizer.from_pretrained("pin/senda")
model = AutoModelForSequenceClassification.from_pretrained("pin/senda")
# create 'senda' sentiment analysis pipeline
senda_pipeline = pipeline('sentiment-analysis', model=model, tokenizer=tokenizer)
text = "Sikke en dejlig dag det er i dag"
# in English: 'what a lovely day'
senda_pipeline(text)
```
## Performance
The `senda` model achieves an accuracy of 0.77 and a macro-averaged F1-score of 0.73 on a small test data set, that [Alexandra Institute](https://github.com/alexandrainst/danlp/blob/master/docs/docs/datasets.md#twitter-sentiment) provides. The model can most certainly be improved, and we encourage all NLP-enthusiasts to give it their best shot - you can use the [`senda`](https://github.com/ebanalyse/senda) package to do this.
#### Contact
Feel free to contact author Lars Kjeldgaard on [[email protected]](mailto:[email protected]).
#### Shout-outs
Props to [Malte Højmark-Berthelsen](mailto:[email protected]) for pretraining Danish BERT and helping out adding a TensorFlow backend for `senda`.
|
sentence-transformers/multi-qa-distilbert-cos-v1 | 4ee499ef6882f9c48c82085c3ead10ed8ac6be28 | 2022-07-11T21:07:27.000Z | [
"pytorch",
"distilbert",
"fill-mask",
"dataset:flax-sentence-embeddings/stackexchange_xml",
"dataset:ms_marco",
"dataset:gooaq",
"dataset:yahoo_answers_topics",
"dataset:search_qa",
"dataset:eli5",
"dataset:natural_questions",
"dataset:trivia_qa",
"dataset:embedding-data/QQP",
"dataset:embedding-data/PAQ_pairs",
"dataset:embedding-data/Amazon-QA",
"dataset:embedding-data/WikiAnswers",
"sentence-transformers",
"feature-extraction",
"sentence-similarity"
] | sentence-similarity | false | sentence-transformers | null | sentence-transformers/multi-qa-distilbert-cos-v1 | 16,171 | 5 | sentence-transformers | 543 | ---
pipeline_tag: sentence-similarity
tags:
- sentence-transformers
- feature-extraction
- sentence-similarity
datasets:
- flax-sentence-embeddings/stackexchange_xml
- ms_marco
- gooaq
- yahoo_answers_topics
- search_qa
- eli5
- natural_questions
- trivia_qa
- embedding-data/QQP
- embedding-data/PAQ_pairs
- embedding-data/Amazon-QA
- embedding-data/WikiAnswers
---
# multi-qa-distilbert-cos-v1
This is a [sentence-transformers](https://www.SBERT.net) model: It maps sentences & paragraphs to a 768 dimensional dense vector space and was designed for **semantic search**. It has been trained on 215M (question, answer) pairs from diverse sources. For an introduction to semantic search, have a look at: [SBERT.net - Semantic Search](https://www.sbert.net/examples/applications/semantic-search/README.html)
## Usage (Sentence-Transformers)
Using this model becomes easy when you have [sentence-transformers](https://www.SBERT.net) installed:
```
pip install -U sentence-transformers
```
Then you can use the model like this:
```python
from sentence_transformers import SentenceTransformer, util
query = "How many people live in London?"
docs = ["Around 9 Million people live in London", "London is known for its financial district"]
#Load the model
model = SentenceTransformer('sentence-transformers/multi-qa-distilbert-cos-v1')
#Encode query and documents
query_emb = model.encode(query)
doc_emb = model.encode(docs)
#Compute dot score between query and all document embeddings
scores = util.dot_score(query_emb, doc_emb)[0].cpu().tolist()
#Combine docs & scores
doc_score_pairs = list(zip(docs, scores))
#Sort by decreasing score
doc_score_pairs = sorted(doc_score_pairs, key=lambda x: x[1], reverse=True)
#Output passages & scores
for doc, score in doc_score_pairs:
print(score, doc)
```
## Usage (HuggingFace Transformers)
Without [sentence-transformers](https://www.SBERT.net), you can use the model like this: First, you pass your input through the transformer model, then you have to apply the correct pooling-operation on-top of the contextualized word embeddings.
```python
from transformers import AutoTokenizer, AutoModel
import torch
import torch.nn.functional as F
#Mean Pooling - Take average of all tokens
def mean_pooling(model_output, attention_mask):
token_embeddings = model_output.last_hidden_state #First element of model_output contains all token embeddings
input_mask_expanded = attention_mask.unsqueeze(-1).expand(token_embeddings.size()).float()
return torch.sum(token_embeddings * input_mask_expanded, 1) / torch.clamp(input_mask_expanded.sum(1), min=1e-9)
#Encode text
def encode(texts):
# Tokenize sentences
encoded_input = tokenizer(texts, padding=True, truncation=True, return_tensors='pt')
# Compute token embeddings
with torch.no_grad():
model_output = model(**encoded_input, return_dict=True)
# Perform pooling
embeddings = mean_pooling(model_output, encoded_input['attention_mask'])
# Normalize embeddings
embeddings = F.normalize(embeddings, p=2, dim=1)
return embeddings
# Sentences we want sentence embeddings for
query = "How many people live in London?"
docs = ["Around 9 Million people live in London", "London is known for its financial district"]
# Load model from HuggingFace Hub
tokenizer = AutoTokenizer.from_pretrained("sentence-transformers/multi-qa-distilbert-cos-v1")
model = AutoModel.from_pretrained("sentence-transformers/multi-qa-distilbert-cos-v1")
#Encode query and docs
query_emb = encode(query)
doc_emb = encode(docs)
#Compute dot score between query and all document embeddings
scores = torch.mm(query_emb, doc_emb.transpose(0, 1))[0].cpu().tolist()
#Combine docs & scores
doc_score_pairs = list(zip(docs, scores))
#Sort by decreasing score
doc_score_pairs = sorted(doc_score_pairs, key=lambda x: x[1], reverse=True)
#Output passages & scores
for doc, score in doc_score_pairs:
print(score, doc)
```
## Technical Details
In the following some technical details how this model must be used:
| Setting | Value |
| --- | :---: |
| Dimensions | 768 |
| Produces normalized embeddings | Yes |
| Pooling-Method | Mean pooling |
| Suitable score functions | dot-product (`util.dot_score`), cosine-similarity (`util.cos_sim`), or euclidean distance |
Note: When loaded with `sentence-transformers`, this model produces normalized embeddings with length 1. In that case, dot-product and cosine-similarity are equivalent. dot-product is preferred as it is faster. Euclidean distance is proportional to dot-product and can also be used.
----
## Background
The project aims to train sentence embedding models on very large sentence level datasets using a self-supervised
contrastive learning objective. We use a contrastive learning objective: given a sentence from the pair, the model should predict which out of a set of randomly sampled other sentences, was actually paired with it in our dataset.
We developped this model during the
[Community week using JAX/Flax for NLP & CV](https://discuss.huggingface.co/t/open-to-the-community-community-week-using-jax-flax-for-nlp-cv/7104),
organized by Hugging Face. We developped this model as part of the project:
[Train the Best Sentence Embedding Model Ever with 1B Training Pairs](https://discuss.huggingface.co/t/train-the-best-sentence-embedding-model-ever-with-1b-training-pairs/7354). We benefited from efficient hardware infrastructure to run the project: 7 TPUs v3-8, as well as intervention from Googles Flax, JAX, and Cloud team member about efficient deep learning frameworks.
## Intended uses
Our model is intented to be used for semantic search: It encodes queries / questions and text paragraphs in a dense vector space. It finds relevant documents for the given passages.
Note that there is a limit of 512 word pieces: Text longer than that will be truncated. Further note that the model was just trained on input text up to 250 word pieces. It might not work well for longer text.
## Training procedure
The full training script is accessible in this current repository: `train_script.py`.
### Pre-training
We use the pretrained [`distilbert-base-uncased`](https://huggingface.co/distilbert-base-uncased) model. Please refer to the model card for more detailed information about the pre-training procedure.
#### Training
We use the concatenation from multiple datasets to fine-tune our model. In total we have about 215M (question, answer) pairs.
We sampled each dataset given a weighted probability which configuration is detailed in the `data_config.json` file.
The model was trained with [MultipleNegativesRankingLoss](https://www.sbert.net/docs/package_reference/losses.html#multiplenegativesrankingloss) using Mean-pooling, cosine-similarity as similarity function, and a scale of 20.
| Dataset | Number of training tuples |
|--------------------------------------------------------|:--------------------------:|
| [WikiAnswers](https://github.com/afader/oqa#wikianswers-corpus) Duplicate question pairs from WikiAnswers | 77,427,422 |
| [PAQ](https://github.com/facebookresearch/PAQ) Automatically generated (Question, Paragraph) pairs for each paragraph in Wikipedia | 64,371,441 |
| [Stack Exchange](https://huggingface.co/datasets/flax-sentence-embeddings/stackexchange_xml) (Title, Body) pairs from all StackExchanges | 25,316,456 |
| [Stack Exchange](https://huggingface.co/datasets/flax-sentence-embeddings/stackexchange_xml) (Title, Answer) pairs from all StackExchanges | 21,396,559 |
| [MS MARCO](https://microsoft.github.io/msmarco/) Triplets (query, answer, hard_negative) for 500k queries from Bing search engine | 17,579,773 |
| [GOOAQ: Open Question Answering with Diverse Answer Types](https://github.com/allenai/gooaq) (query, answer) pairs for 3M Google queries and Google featured snippet | 3,012,496 |
| [Amazon-QA](http://jmcauley.ucsd.edu/data/amazon/qa/) (Question, Answer) pairs from Amazon product pages | 2,448,839
| [Yahoo Answers](https://www.kaggle.com/soumikrakshit/yahoo-answers-dataset) (Title, Answer) pairs from Yahoo Answers | 1,198,260 |
| [Yahoo Answers](https://www.kaggle.com/soumikrakshit/yahoo-answers-dataset) (Question, Answer) pairs from Yahoo Answers | 681,164 |
| [Yahoo Answers](https://www.kaggle.com/soumikrakshit/yahoo-answers-dataset) (Title, Question) pairs from Yahoo Answers | 659,896 |
| [SearchQA](https://huggingface.co/datasets/search_qa) (Question, Answer) pairs for 140k questions, each with Top5 Google snippets on that question | 582,261 |
| [ELI5](https://huggingface.co/datasets/eli5) (Question, Answer) pairs from Reddit ELI5 (explainlikeimfive) | 325,475 |
| [Stack Exchange](https://huggingface.co/datasets/flax-sentence-embeddings/stackexchange_xml) Duplicate questions pairs (titles) | 304,525 |
| [Quora Question Triplets](https://quoradata.quora.com/First-Quora-Dataset-Release-Question-Pairs) (Question, Duplicate_Question, Hard_Negative) triplets for Quora Questions Pairs dataset | 103,663 |
| [Natural Questions (NQ)](https://ai.google.com/research/NaturalQuestions) (Question, Paragraph) pairs for 100k real Google queries with relevant Wikipedia paragraph | 100,231 |
| [SQuAD2.0](https://rajpurkar.github.io/SQuAD-explorer/) (Question, Paragraph) pairs from SQuAD2.0 dataset | 87,599 |
| [TriviaQA](https://huggingface.co/datasets/trivia_qa) (Question, Evidence) pairs | 73,346 |
| **Total** | **214,988,242** | |
bigscience/bloom-350m | ef09465f3d95a3fda91faf0d814b95fd3521b73d | 2022-07-21T08:04:09.000Z | [
"pytorch",
"jax",
"bloom",
"feature-extraction",
"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",
"transformers",
"license:bigscience-bloom-rail-1.0",
"text-generation"
] | text-generation | false | bigscience | null | bigscience/bloom-350m | 16,140 | 5 | transformers | 544 | ---
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
pipeline_tag: text-generation
---
<h1 style='text-align: center '>BLOOM LM</h1>
<h2 style='text-align: center '><em>BigScience Large Open-science Open-access Multilingual Language Model</em> </h2>
<h3 style='text-align: center '>Model Card</h3>
<img src="https://s3.amazonaws.com/moonup/production/uploads/1657124309515-5f17f0a0925b9863e28ad517.png" alt="BigScience Logo" width="800" style="margin-left:'auto' margin-right:'auto' display:'block'"/>
Version 1.0 / 26.May.2022
## 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
### Basics
*This section provides information for anyone who wants to know about the model.*
<details>
<summary>Click to expand</summary> <br/>
**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 provides information for people who work on model development.*
<details>
<summary>Click to expand</summary><br/>
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:** 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
* 350 million parameters:
* 24 layers, 16 attention heads
* Hidden layers are 1024-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))
#### **Training**
_In progress._
Current training logs: [Tensorboard link](https://huggingface.co/tensorboard/bigscience/tr11-176B-ml-logs/)
- 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 (*current target*)
- Dates:
- Started 11th March, 2022 11:42am PST
- Estimated end: 5th July, 2022
- Estimated cost of training: Equivalent of $2-5M in cloud computing (including preliminary experiments)
- Server training location: Île-de-France, France
#### **Tokenization**
The BLOOM tokenizer ([link](https://huggingface.co/bigscience/tokenizer)) is 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.
</details>
### Environmental Impact
<details>
<summary>Click to expand</summary><br/>
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 upon completion of training.)*
**Estimated electricity usage:** *(Forthcoming upon completion of training.)*
</details>
<p> </p>
## 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 provides information for anyone considering using the model or who is affected by the model.*
<details>
<summary>Click to expand</summary><br/>
### 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>
<p> </p>
## 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>
<summary>Click to expand</summary><br/>
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.
The following table shows the further distribution of Niger-Congo and Indic languages in the training data.
<details>
<summary>Click to expand</summary><br/>
| 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 |
</details>
The following table shows the distribution of programming languages.
<details>
<summary>Click to expand</summary><br/>
| 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 |
</details>
</details>
<p> </p>
## Risks and Limitations
*This section identifies foreseeable harms and misunderstandings.*
<details>
<summary>Click to expand</summary><br/>
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>
<p> </p>
## Evaluation
*This section describes the evaluation protocols and provides the results.*
<details>
<summary>Click to expand</summary><br/>
### 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 at the end of model training.)
</details>
<p> </p>
## Recommendations
*This section provides information on warnings and potential mitigations.*
<details>
<summary>Click to expand</summary><br/>
- 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 pretrained with the LLM 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>
<p> </p>
## Glossary and Calculations
*This section defines common terms and how metrics are calculated.*
<details>
<summary>Click to expand</summary><br/>
- <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>
<p> </p>
## More Information
<details>
<summary>Click to expand</summary><br/>
### 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
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>
<p> </p>
## 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
|
textattack/bert-base-uncased-yelp-polarity | a4d0a85ea6c1d5bb944dcc12ea5c918863e469a4 | 2021-05-20T07:49:07.000Z | [
"pytorch",
"jax",
"bert",
"text-classification",
"transformers"
] | text-classification | false | textattack | null | textattack/bert-base-uncased-yelp-polarity | 16,133 | null | transformers | 545 | ## TextAttack Model Card
This `bert-base-uncased` model was fine-tuned for sequence classification using TextAttack
and the yelp_polarity dataset loaded using the `nlp` library. The model was fine-tuned
for 5 epochs with a batch size of 16, a learning
rate of 5e-05, and a maximum sequence length of 256.
Since this was a classification task, the model was trained with a cross-entropy loss function.
The best score the model achieved on this task was 0.9699473684210527, as measured by the
eval set accuracy, found after 4 epochs.
For more information, check out [TextAttack on Github](https://github.com/QData/TextAttack).
|
KB/bert-base-swedish-cased-ner | e1c9ae76afa22ce28d2097310ab95312d73e4e3a | 2022-06-07T16:34:49.000Z | [
"pytorch",
"tf",
"jax",
"bert",
"token-classification",
"sv",
"transformers",
"autotrain_compatible"
] | token-classification | false | KB | null | KB/bert-base-swedish-cased-ner | 16,095 | null | transformers | 546 | ---
language: sv
---
# Swedish BERT Models
The National Library of Sweden / KBLab releases three pretrained language models based on BERT and ALBERT. The models are trained on approximately 15-20GB of text (200M sentences, 3000M tokens) from various sources (books, news, government publications, swedish wikipedia and internet forums) aiming to provide a representative BERT model for Swedish text. A more complete description will be published later on.
The following three models are currently available:
- **bert-base-swedish-cased** (*v1*) - A BERT trained with the same hyperparameters as first published by Google.
- **bert-base-swedish-cased-ner** (*experimental*) - a BERT fine-tuned for NER using SUC 3.0.
- **albert-base-swedish-cased-alpha** (*alpha*) - A first attempt at an ALBERT for Swedish.
All models are cased and trained with whole word masking.
## Files
| **name** | **files** |
|---------------------------------|-----------|
| bert-base-swedish-cased | [config](https://s3.amazonaws.com/models.huggingface.co/bert/KB/bert-base-swedish-cased/config.json), [vocab](https://s3.amazonaws.com/models.huggingface.co/bert/KB/bert-base-swedish-cased/vocab.txt), [pytorch_model.bin](https://s3.amazonaws.com/models.huggingface.co/bert/KB/bert-base-swedish-cased/pytorch_model.bin) |
| bert-base-swedish-cased-ner | [config](https://s3.amazonaws.com/models.huggingface.co/bert/KB/bert-base-swedish-cased-ner/config.json), [vocab](https://s3.amazonaws.com/models.huggingface.co/bert/KB/bert-base-swedish-cased-ner/vocab.txt) [pytorch_model.bin](https://s3.amazonaws.com/models.huggingface.co/bert/KB/bert-base-swedish-cased-ner/pytorch_model.bin) |
| albert-base-swedish-cased-alpha | [config](https://s3.amazonaws.com/models.huggingface.co/bert/KB/albert-base-swedish-cased-alpha/config.json), [sentencepiece model](https://s3.amazonaws.com/models.huggingface.co/bert/KB/albert-base-swedish-cased-alpha/spiece.model), [pytorch_model.bin](https://s3.amazonaws.com/models.huggingface.co/bert/KB/albert-base-swedish-cased-alpha/pytorch_model.bin) |
TensorFlow model weights will be released soon.
## Usage requirements / installation instructions
The examples below require Huggingface Transformers 2.4.1 and Pytorch 1.3.1 or greater. For Transformers<2.4.0 the tokenizer must be instantiated manually and the `do_lower_case` flag parameter set to `False` and `keep_accents` to `True` (for ALBERT).
To create an environment where the examples can be run, run the following in an terminal on your OS of choice.
```
# git clone https://github.com/Kungbib/swedish-bert-models
# cd swedish-bert-models
# python3 -m venv venv
# source venv/bin/activate
# pip install --upgrade pip
# pip install -r requirements.txt
```
### BERT Base Swedish
A standard BERT base for Swedish trained on a variety of sources. Vocabulary size is ~50k. Using Huggingface Transformers the model can be loaded in Python as follows:
```python
from transformers import AutoModel,AutoTokenizer
tok = AutoTokenizer.from_pretrained('KB/bert-base-swedish-cased')
model = AutoModel.from_pretrained('KB/bert-base-swedish-cased')
```
### BERT base fine-tuned for Swedish NER
This model is fine-tuned on the SUC 3.0 dataset. Using the Huggingface pipeline the model can be easily instantiated. For Transformer<2.4.1 it seems the tokenizer must be loaded separately to disable lower-casing of input strings:
```python
from transformers import pipeline
nlp = pipeline('ner', model='KB/bert-base-swedish-cased-ner', tokenizer='KB/bert-base-swedish-cased-ner')
nlp('Idag släpper KB tre språkmodeller.')
```
Running the Python code above should produce in something like the result below. Entity types used are `TME` for time, `PRS` for personal names, `LOC` for locations, `EVN` for events and `ORG` for organisations. These labels are subject to change.
```python
[ { 'word': 'Idag', 'score': 0.9998126029968262, 'entity': 'TME' },
{ 'word': 'KB', 'score': 0.9814832210540771, 'entity': 'ORG' } ]
```
The BERT tokenizer often splits words into multiple tokens, with the subparts starting with `##`, for example the string `Engelbert kör Volvo till Herrängens fotbollsklubb` gets tokenized as `Engel ##bert kör Volvo till Herr ##ängens fotbolls ##klubb`. To glue parts back together one can use something like this:
```python
text = 'Engelbert tar Volvon till Tele2 Arena för att titta på Djurgården IF ' +\
'som spelar fotboll i VM klockan två på kvällen.'
l = []
for token in nlp(text):
if token['word'].startswith('##'):
l[-1]['word'] += token['word'][2:]
else:
l += [ token ]
print(l)
```
Which should result in the following (though less cleanly formatted):
```python
[ { 'word': 'Engelbert', 'score': 0.99..., 'entity': 'PRS'},
{ 'word': 'Volvon', 'score': 0.99..., 'entity': 'OBJ'},
{ 'word': 'Tele2', 'score': 0.99..., 'entity': 'LOC'},
{ 'word': 'Arena', 'score': 0.99..., 'entity': 'LOC'},
{ 'word': 'Djurgården', 'score': 0.99..., 'entity': 'ORG'},
{ 'word': 'IF', 'score': 0.99..., 'entity': 'ORG'},
{ 'word': 'VM', 'score': 0.99..., 'entity': 'EVN'},
{ 'word': 'klockan', 'score': 0.99..., 'entity': 'TME'},
{ 'word': 'två', 'score': 0.99..., 'entity': 'TME'},
{ 'word': 'på', 'score': 0.99..., 'entity': 'TME'},
{ 'word': 'kvällen', 'score': 0.54..., 'entity': 'TME'} ]
```
### ALBERT base
The easiest way to do this is, again, using Huggingface Transformers:
```python
from transformers import AutoModel,AutoTokenizer
tok = AutoTokenizer.from_pretrained('KB/albert-base-swedish-cased-alpha'),
model = AutoModel.from_pretrained('KB/albert-base-swedish-cased-alpha')
```
## Acknowledgements ❤️
- Resources from Stockholms University, Umeå University and Swedish Language Bank at Gothenburg University were used when fine-tuning BERT for NER.
- Model pretraining was made partly in-house at the KBLab and partly (for material without active copyright) with the support of Cloud TPUs from Google's TensorFlow Research Cloud (TFRC).
- Models are hosted on S3 by Huggingface 🤗
|
facebook/wav2vec2-large-xlsr-53-french | e7c54e9cde9f6d84be783c7d04d34e1c8efcc1d1 | 2021-07-06T02:40:56.000Z | [
"pytorch",
"jax",
"wav2vec2",
"automatic-speech-recognition",
"fr",
"dataset:common_voice",
"transformers",
"speech",
"audio",
"license:apache-2.0"
] | automatic-speech-recognition | false | facebook | null | facebook/wav2vec2-large-xlsr-53-french | 15,879 | 6 | transformers | 547 | ---
language: fr
datasets:
- common_voice
tags:
- speech
- audio
- automatic-speech-recognition
license: apache-2.0
---
## Evaluation on Common Voice FR Test
```python
import torchaudio
from datasets import load_dataset, load_metric
from transformers import (
Wav2Vec2ForCTC,
Wav2Vec2Processor,
)
import torch
import re
import sys
model_name = "facebook/wav2vec2-large-xlsr-53-french"
device = "cuda"
chars_to_ignore_regex = '[\,\?\.\!\-\;\:\"]' # noqa: W605
model = Wav2Vec2ForCTC.from_pretrained(model_name).to(device)
processor = Wav2Vec2Processor.from_pretrained(model_name)
ds = load_dataset("common_voice", "fr", split="test", data_dir="./cv-corpus-6.1-2020-12-11")
resampler = torchaudio.transforms.Resample(orig_freq=48_000, new_freq=16_000)
def map_to_array(batch):
speech, _ = torchaudio.load(batch["path"])
batch["speech"] = resampler.forward(speech.squeeze(0)).numpy()
batch["sampling_rate"] = resampler.new_freq
batch["sentence"] = re.sub(chars_to_ignore_regex, '', batch["sentence"]).lower().replace("’", "'")
return batch
ds = ds.map(map_to_array)
def map_to_pred(batch):
features = processor(batch["speech"], sampling_rate=batch["sampling_rate"][0], padding=True, return_tensors="pt")
input_values = features.input_values.to(device)
attention_mask = features.attention_mask.to(device)
with torch.no_grad():
logits = model(input_values, attention_mask=attention_mask).logits
pred_ids = torch.argmax(logits, dim=-1)
batch["predicted"] = processor.batch_decode(pred_ids)
batch["target"] = batch["sentence"]
return batch
result = ds.map(map_to_pred, batched=True, batch_size=16, remove_columns=list(ds.features.keys()))
wer = load_metric("wer")
print(wer.compute(predictions=result["predicted"], references=result["target"]))
```
**Result**: 25.2 % |
yarongef/DistilProtBert | 9e3f170d21d907fe7d1370360f1993c710777bf2 | 2022-06-14T12:39:15.000Z | [
"pytorch",
"bert",
"fill-mask",
"protein",
"dataset:Uniref50",
"transformers",
"protein language model",
"license:mit",
"autotrain_compatible"
] | fill-mask | false | yarongef | null | yarongef/DistilProtBert | 15,770 | 1 | transformers | 548 | ---
license: mit
language: protein
tags:
- protein language model
datasets:
- Uniref50
---
# DistilProtBert
A distilled version of [ProtBert-UniRef100](https://huggingface.co/Rostlab/prot_bert) model.
In addition to cross entropy and cosine teacher-student losses, DistilProtBert was pretrained on a masked language modeling (MLM) objective and it only works with capital letter amino acids.
Check out our paper [DistilProtBert: A distilled protein language model used to distinguish between real proteins and their randomly shuffled counterparts](https://www.biorxiv.org/content/10.1101/2022.05.09.491157v1) for more details.
[Git](https://github.com/yarongef/DistilProtBert) repository.
# Model details
| **Model** | **# of parameters** | **# of hidden layers** | **Pretraining dataset** | **# of proteins** | **Pretraining hardware** |
|:--------------:|:-------------------:|:----------------------:|:-----------------------:|:------------------------------:|:------------------------:|
| ProtBert | 420M | 30 | UniRef100 | 216M | 512 16GB TPUs |
| DistilProtBert | 230M | 15 | UniRef50 | 43M | 5 v100 32GB GPUs |
## Intended uses & limitations
The model could be used for protein feature extraction or to be fine-tuned on downstream tasks.
### How to use
The model can be used the same as ProtBert and with ProtBert's tokenizer.
## Training data
DistilProtBert model was pretrained on [Uniref50](https://www.uniprot.org/downloads), a dataset consisting of ~43 million protein sequences (only sequences of length between 20 to 512 amino acids were used).
# Pretraining procedure
Preprocessing was done using ProtBert's tokenizer.
The details of the masking procedure for each sequence followed the original Bert (as mentioned in [ProtBert](https://huggingface.co/Rostlab/prot_bert)).
The model was pretrained on a single DGX cluster for 3 epochs in total. local batch size was 16, the optimizer used was AdamW with a learning rate of 5e-5 and mixed precision settings.
## Evaluation results
When fine-tuned on downstream tasks, this model achieves the following results:
| Task/Dataset | secondary structure (3-states) | Membrane |
|:-----:|:-----:|:-----:|
| CASP12 | 72 | |
| TS115 | 81 | |
| CB513 | 79 | |
| DeepLoc | | 86 |
Distinguish between proteins and their k-let shuffled versions:
_Singlet_ ([dataset](https://huggingface.co/datasets/yarongef/human_proteome_singlets))
| Model | AUC |
|:--------------:|:-------:|
| LSTM | 0.71 |
| ProtBert | 0.93 |
| DistilProtBert | 0.92 |
_Doublet_ ([dataset](https://huggingface.co/datasets/yarongef/human_proteome_doublets))
| Model | AUC |
|:--------------:|:-------:|
| LSTM | 0.68 |
| ProtBert | 0.92 |
| DistilProtBert | 0.91 |
_Triplet_ ([dataset](https://huggingface.co/datasets/yarongef/human_proteome_triplets))
| Model | AUC |
|:--------------:|:-------:|
| LSTM | 0.61 |
| ProtBert | 0.92 |
| DistilProtBert | 0.87 |
## **Citation**
If you use this model, please cite our paper:
```
@article {
author = {Geffen, Yaron and Ofran, Yanay and Unger, Ron},
title = {DistilProtBert: A distilled protein language model used to distinguish between real proteins and their randomly shuffled counterparts},
year = {2022},
doi = {10.1101/2022.05.09.491157},
URL = {https://www.biorxiv.org/content/early/2022/05/10/2022.05.09.491157},
eprint = {https://www.biorxiv.org/content/early/2022/05/10/2022.05.09.491157.full.pdf},
journal = {bioRxiv}
}
``` |
sentence-transformers/paraphrase-albert-small-v2 | b8a76dca618575852d8874313e2ad84d423f333f | 2022-07-08T04:07:04.000Z | [
"pytorch",
"tf",
"rust",
"albert",
"feature-extraction",
"dataset:flax-sentence-embeddings/stackexchange_xml",
"dataset:s2orc",
"dataset:ms_marco",
"dataset:wiki_atomic_edits",
"dataset:snli",
"dataset:multi_nli",
"dataset:embedding-data/altlex",
"dataset:embedding-data/simple-wiki",
"dataset:embedding-data/flickr30k-captions",
"dataset:embedding-data/coco_captions",
"dataset:embedding-data/sentence-compression",
"dataset:embedding-data/QQP",
"dataset:yahoo_answers_topics",
"arxiv:1908.10084",
"sentence-transformers",
"sentence-similarity",
"transformers",
"license:apache-2.0"
] | sentence-similarity | false | sentence-transformers | null | sentence-transformers/paraphrase-albert-small-v2 | 15,743 | 1 | sentence-transformers | 549 | ---
pipeline_tag: sentence-similarity
license: apache-2.0
tags:
- sentence-transformers
- feature-extraction
- sentence-similarity
- transformers
datasets:
- flax-sentence-embeddings/stackexchange_xml
- s2orc
- ms_marco
- wiki_atomic_edits
- snli
- multi_nli
- embedding-data/altlex
- embedding-data/simple-wiki
- embedding-data/flickr30k-captions
- embedding-data/coco_captions
- embedding-data/sentence-compression
- embedding-data/QQP
- yahoo_answers_topics
---
# sentence-transformers/paraphrase-albert-small-v2
This is a [sentence-transformers](https://www.SBERT.net) model: It maps sentences & paragraphs to a 768 dimensional dense vector space and can be used for tasks like clustering or semantic search.
## Usage (Sentence-Transformers)
Using this model becomes easy when you have [sentence-transformers](https://www.SBERT.net) installed:
```
pip install -U sentence-transformers
```
Then you can use the model like this:
```python
from sentence_transformers import SentenceTransformer
sentences = ["This is an example sentence", "Each sentence is converted"]
model = SentenceTransformer('sentence-transformers/paraphrase-albert-small-v2')
embeddings = model.encode(sentences)
print(embeddings)
```
## Usage (HuggingFace Transformers)
Without [sentence-transformers](https://www.SBERT.net), you can use the model like this: First, you pass your input through the transformer model, then you have to apply the right pooling-operation on-top of the contextualized word embeddings.
```python
from transformers import AutoTokenizer, AutoModel
import torch
#Mean Pooling - Take attention mask into account for correct averaging
def mean_pooling(model_output, attention_mask):
token_embeddings = model_output[0] #First element of model_output contains all token embeddings
input_mask_expanded = attention_mask.unsqueeze(-1).expand(token_embeddings.size()).float()
return torch.sum(token_embeddings * input_mask_expanded, 1) / torch.clamp(input_mask_expanded.sum(1), min=1e-9)
# Sentences we want sentence embeddings for
sentences = ['This is an example sentence', 'Each sentence is converted']
# Load model from HuggingFace Hub
tokenizer = AutoTokenizer.from_pretrained('sentence-transformers/paraphrase-albert-small-v2')
model = AutoModel.from_pretrained('sentence-transformers/paraphrase-albert-small-v2')
# Tokenize sentences
encoded_input = tokenizer(sentences, padding=True, truncation=True, return_tensors='pt')
# Compute token embeddings
with torch.no_grad():
model_output = model(**encoded_input)
# Perform pooling. In this case, max pooling.
sentence_embeddings = mean_pooling(model_output, encoded_input['attention_mask'])
print("Sentence embeddings:")
print(sentence_embeddings)
```
## Evaluation Results
For an automated evaluation of this model, see the *Sentence Embeddings Benchmark*: [https://seb.sbert.net](https://seb.sbert.net?model_name=sentence-transformers/paraphrase-albert-small-v2)
## Full Model Architecture
```
SentenceTransformer(
(0): Transformer({'max_seq_length': 100, 'do_lower_case': False}) with Transformer model: AlbertModel
(1): Pooling({'word_embedding_dimension': 768, 'pooling_mode_cls_token': False, 'pooling_mode_mean_tokens': True, 'pooling_mode_max_tokens': False, 'pooling_mode_mean_sqrt_len_tokens': False})
)
```
## Citing & Authors
This model was trained by [sentence-transformers](https://www.sbert.net/).
If you find this model helpful, feel free to cite our publication [Sentence-BERT: Sentence Embeddings using Siamese BERT-Networks](https://arxiv.org/abs/1908.10084):
```bibtex
@inproceedings{reimers-2019-sentence-bert,
title = "Sentence-BERT: Sentence Embeddings using Siamese BERT-Networks",
author = "Reimers, Nils and Gurevych, Iryna",
booktitle = "Proceedings of the 2019 Conference on Empirical Methods in Natural Language Processing",
month = "11",
year = "2019",
publisher = "Association for Computational Linguistics",
url = "http://arxiv.org/abs/1908.10084",
}
``` |
bigscience/bloom-1b3 | 889e0ff2dc58b977bbd4b22738ff110be5b4e400 | 2022-07-13T09:02:56.000Z | [
"pytorch",
"jax",
"bloom",
"feature-extraction",
"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",
"transformers",
"license:bigscience-bloom-rail-1.0",
"text-generation"
] | text-generation | false | bigscience | null | bigscience/bloom-1b3 | 15,738 | 37 | transformers | 550 | ---
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
pipeline_tag: text-generation
---
<h1 style='text-align: center '>BLOOM LM</h1>
<h2 style='text-align: center '><em>BigScience Large Open-science Open-access Multilingual Language Model</em> </h2>
<h3 style='text-align: center '>Model Card</h3>
<img src="https://s3.amazonaws.com/moonup/production/uploads/1657124309515-5f17f0a0925b9863e28ad517.png" alt="BigScience Logo" width="800" style="margin-left:'auto' margin-right:'auto' display:'block'"/>
Version 1.0 / 26.May.2022
## 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
### Basics
*This section provides information for anyone who wants to know about the model.*
<details>
<summary>Click to expand</summary> <br/>
**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 provides information for people who work on model development.*
<details>
<summary>Click to expand</summary><br/>
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:** 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
* 1.3 billion parameters:
* 24 layers, 16 attention heads
* Hidden layers are 2048-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: 64 V100 16/32GB GPUs (16 nodes):
* 4 GPUs per node
* 40 CPUs per task
* 1 task per node
* CPU: AMD
* CPU memory: 160GB per node
* GPU memory: 64GB or 128GB (depending on node availability during training) 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))
#### **Training**
- Checkpoint size:
- Fp16 weights: 2.6GB (# params * 2)
- Full checkpoint with optimizer states: --
- Training throughput: --
- Number of epochs: 1
- Dates:
- Start: 11th March, 2022 11:42am PST
- End: 20 May, 2022
- Server training location: Île-de-France, France
#### **Tokenization**
The BLOOM tokenizer ([link](https://huggingface.co/bigscience/tokenizer)) is 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.
</details>
### Environmental Impact
<details>
<summary>Click to expand</summary><br/>
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 upon completion of training.)*
**Estimated electricity usage:** *(Forthcoming upon completion of training.)*
</details>
<p> </p>
## 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 provides information for anyone considering using the model or who is affected by the model.*
<details>
<summary>Click to expand</summary><br/>
### 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>
<p> </p>
## 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>
<summary>Click to expand</summary><br/>
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.
The following table shows the further distribution of Niger-Congo and Indic languages in the training data.
<details>
<summary>Click to expand</summary><br/>
| 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 |
</details>
The following table shows the distribution of programming languages.
<details>
<summary>Click to expand</summary><br/>
| 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 |
</details>
</details>
<p> </p>
## Risks and Limitations
*This section identifies foreseeable harms and misunderstandings.*
<details>
<summary>Click to expand</summary><br/>
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>
<p> </p>
## Evaluation
*This section describes the evaluation protocols and provides the results.*
<details>
<summary>Click to expand</summary><br/>
### 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 at the end of model training.)
- [BLOOM Book](https://huggingface.co/spaces/bigscience/bloom-book): Read generations from BLOOM based on prompts provided by the community
</details>
<p> </p>
## Recommendations
*This section provides information on warnings and potential mitigations.*
<details>
<summary>Click to expand</summary><br/>
- 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 pretrained with the LLM 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>
<p> </p>
## Glossary and Calculations
*This section defines common terms and how metrics are calculated.*
<details>
<summary>Click to expand</summary><br/>
- <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>
<p> </p>
## More Information
<details>
<summary>Click to expand</summary><br/>
### 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
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>
<p> </p>
## 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
|
ctrl | e3789f31af06f7dfbc086b59c562557a1d86d33b | 2022-07-22T08:04:22.000Z | [
"pytorch",
"tf",
"ctrl",
"en",
"arxiv:1909.05858",
"arxiv:1910.09700",
"transformers",
"license:bsd-3-clause"
] | null | false | null | null | ctrl | 15,719 | null | transformers | 551 | ---
language: en
license: bsd-3-clause
---
# ctrl
# Table of Contents
1. [Model Details](#model-details)
2. [Uses](#uses)
3. [Bias, Risks, and Limitations](#bias-risks-and-limitations)
4. [Training](#training)
5. [Evaluation](#evaluation)
6. [Environmental Impact](#environmental-impact)
7. [Technical Specifications](#technical-specifications)
8. [Citation](#citation)
9. [Model Card Authors](#model-card-authors)
10. [How To Get Started With the Model](#how-to-get-started-with-the-model)
# Model Details
## Model Description
The CTRL model was proposed in [CTRL: A Conditional Transformer Language Model for Controllable Generation](https://arxiv.org/abs/1909.05858) by Nitish Shirish Keskar*, Bryan McCann*, Lav R. Varshney, Caiming Xiong and Richard Socher. It's a causal (unidirectional) transformer pre-trained using language modeling on a very large corpus of ~140 GB of text data with the first token reserved as a control code (such as Links, Books, Wikipedia etc.). The model developers released a model card for CTRL, available [here](https://github.com/salesforce/ctrl/blob/master/ModelCard.pdf).
In their [model card](https://github.com/salesforce/ctrl/blob/master/ModelCard.pdf), the developers write:
> The CTRL Language Model analyzed in this card generates text conditioned on control codes that specify domain, style, topics, dates, entities, relationships between entities, plot points, and task-related behavior.
- **Developed by:** See [associated paper](https://arxiv.org/abs/1909.05858) from Salesforce Research
- **Model type:** Transformer-based language model
- **Language(s) (NLP):** Primarily English, some German, Spanish, French
- **License:** [BSD 3-Clause](https://github.com/salesforce/ctrl/blob/master/LICENSE.txt); also see [Code of Conduct](https://github.com/salesforce/ctrl)
- **Related Models:** More information needed
- **Parent Model:** More information needed
- **Resources for more information:**
- [Associated paper](https://arxiv.org/abs/1909.05858)
- [GitHub repo](https://github.com/salesforce/ctrl)
- [Developer Model Card](https://github.com/salesforce/ctrl/blob/master/ModelCard.pdf)
- [Blog post](https://blog.salesforceairesearch.com/introducing-a-conditional-transformer-language-model-for-controllable-generation/)
# Uses
## Direct Use
The model is a language model. The model can be used for text generation.
## Downstream Use
In their [model card](https://github.com/salesforce/ctrl/blob/master/ModelCard.pdf), the developers write that the primary intended users are general audiences and NLP Researchers, and that the primary intended uses are:
> 1. Generating artificial text in collaboration with a human, including but not limited to:
> - Creative writing
> - Automating repetitive writing tasks
> - Formatting specific text types
> - Creating contextualized marketing materials
> 2. Improvement of other NLP applications through fine-tuning (on another task or other data, e.g. fine-tuning CTRL to learn new kinds of language like product descriptions)
> 3. Enhancement in the field of natural language understanding to push towards a better understanding of artificial text generation, including how to detect it and work toward control, understanding, and potentially combating potentially negative consequences of such models.
## Out-of-Scope Use
In their [model card](https://github.com/salesforce/ctrl/blob/master/ModelCard.pdf), the developers write:
> - CTRL should not be used for generating artificial text without collaboration with a human.
> - It should not be used to make normative or prescriptive claims.
> - This software should not be used to promote or profit from:
> - violence, hate, and division;
> - environmental destruction;
> - abuse of human rights; or
> - the destruction of people's physical and mental health.
# Bias, Risks, and Limitations
Significant research has explored bias and fairness issues with language models (see, e.g., [Sheng et al. (2021)](https://aclanthology.org/2021.acl-long.330.pdf) and [Bender et al. (2021)](https://dl.acm.org/doi/pdf/10.1145/3442188.3445922)). Predictions generated by the model may include disturbing and harmful stereotypes across protected classes; identity characteristics; and sensitive, social, and occupational groups.
In their [model card](https://github.com/salesforce/ctrl/blob/master/ModelCard.pdf), the developers write:
> We recognize the potential for misuse or abuse, including use by bad actors who could manipulate the system to act maliciously and generate text to influence decision-making in political, economic, and social settings. False attribution could also harm individuals, organizations, or other entities. To address these concerns, the model was evaluated internally as well as externally by third parties, including the Partnership on AI, prior to release.
> To mitigate potential misuse to the extent possible, we stripped out all detectable training data from undesirable sources. We then redteamed the model and found that negative utterances were often placed in contexts that made them identifiable as such. For example, when using the ‘News’ control code, hate speech could be embedded as part of an apology (e.g. “the politician apologized for saying [insert hateful statement]”), implying that this type of speech was negative. By pre-selecting the available control codes (omitting, for example, Instagram and Twitter from the available domains), we are able to limit the potential for misuse.
> In releasing our model, we hope to put it into the hands of researchers and prosocial actors so that they can work to control, understand, and potentially combat the negative consequences of such models. We hope that research into detecting fake news and model-generated content of all kinds will be pushed forward by CTRL. It is our belief that these models should become a common tool so researchers can design methods to guard against malicious use and so the public becomes familiar with their existence and patterns of behavior.
See the [associated paper](https://arxiv.org/pdf/1909.05858.pdf) for further discussions about the ethics of LLMs.
## Recommendations
In their [model card](https://github.com/salesforce/ctrl/blob/master/ModelCard.pdf), the developers write:
> - A recommendation to monitor and detect use will be implemented through the development of a model that will identify CTRLgenerated text.
> - A second recommendation to further screen the input into and output from the model will be implemented through the addition of a check in the CTRL interface to prohibit the insertion into the model of certain negative inputs, which will help control the output that can be generated.
> - The model is trained on a limited number of languages: primarily English and some German, Spanish, French. A recommendation for a future area of research is to train the model on more languages.
See the [CTRL-detector GitHub repo](https://github.com/salesforce/ctrl-detector) for more on the detector model.
# Training
## Training Data
In their [model card](https://github.com/salesforce/ctrl/blob/master/ModelCard.pdf), the developers write:
> This model is trained on 140 GB of text drawn from a variety of domains: Wikipedia (English, German, Spanish, and French), Project Gutenberg, submissions from 45 subreddits, OpenWebText, a large collection of news data, Amazon Reviews, Europarl and UN data from WMT (En-De, En-Es, En-Fr), question-answer pairs (no context documents) from ELI5, and the MRQA shared task, which includes Stanford Question Answering Dataset, NewsQA, TriviaQA, SearchQA, HotpotQA, and Natural Questions. See the paper for the full list of training data.
## Training Procedure
### Preprocessing
In the [associated paper](https://arxiv.org/pdf/1909.05858.pdf) the developers write:
> We learn BPE (Sennrich et al., 2015) codes and tokenize the data using fastBPE4, but we use a large vocabulary of roughly 250K tokens. This includes the sub-word tokens necessary to mitigate problems with rare words, but it also reduces the average number of tokens required to generate long text by including most common words. We use English Wikipedia and a 5% split of our collected OpenWebText data for learning BPE codes. We also introduce an unknown token so that during preprocessing we can filter out sequences that contain more than 2 unknown tokens. This, along with the compressed storage for efficient training (TFRecords) (Abadi et al., 2016), reduces our training data to 140 GB from the total 180 GB collected.
See the paper for links, references, and further details.
### Training
In the [associated paper](https://arxiv.org/pdf/1909.05858.pdf) the developers write:
> CTRL has model dimension d = 1280, inner dimension f = 8192, 48 layers, and 16 heads per layer. Dropout with probability 0.1 follows the residual connections in each layer. Token embeddings were tied with the final output embedding layer (Inan et al., 2016; Press & Wolf, 2016).
See the paper for links, references, and further details.
# Evaluation
## Testing Data, Factors & Metrics
In their [model card](https://github.com/salesforce/ctrl/blob/master/ModelCard.pdf), the developers write that model performance measures are:
> Performance evaluated on qualitative judgments by humans as to whether the control codes lead to text generated in the desired domain
# Environmental Impact
Carbon emissions can be estimated using the [Machine Learning Impact calculator](https://mlco2.github.io/impact#compute) presented in [Lacoste et al. (2019)](https://arxiv.org/abs/1910.09700). Details are pulled from the [associated paper](https://arxiv.org/pdf/1909.05858.pdf).
- **Hardware Type:** TPU v3 Pod
- **Hours used:** Approximately 336 hours (2 weeks)
- **Cloud Provider:** GCP
- **Compute Region:** More information needed
- **Carbon Emitted:** More information needed
# Technical Specifications
In the [associated paper](https://arxiv.org/pdf/1909.05858.pdf) the developers write:
> CTRL was implemented in TensorFlow (Abadi et al., 2016) and trained with a global batch size of 1024 distributed across 256 cores of a Cloud TPU v3 Pod for 800k iterations. Training took approximately 2 weeks using Adagrad (Duchi et al., 2011) with a linear warmup from 0 to 0.05 over 25k steps. The norm of gradients were clipped to 0.25 as in (Merity et al., 2017). Learning rate decay was not necessary due to the monotonic nature of the Adagrad accumulator. We compared to the Adam optimizer (Kingma & Ba, 2014) while training smaller models, but we noticed comparable convergence rates and significant memory savings with Adagrad. We also experimented with explicit memory-saving optimizers including SM3 (Anil et al., 2019), Adafactor (Shazeer & Stern, 2018), and NovoGrad (Ginsburg et al., 2019) with mixed results.
See the paper for links, references, and further details.
# Citation
**BibTeX:**
```bibtex
@article{keskarCTRL2019,
title={{CTRL - A Conditional Transformer Language Model for Controllable Generation}},
author={Keskar, Nitish Shirish and McCann, Bryan and Varshney, Lav and Xiong, Caiming and Socher, Richard},
journal={arXiv preprint arXiv:1909.05858},
year={2019}
}
```
**APA:**
- Keskar, N. S., McCann, B., Varshney, L. R., Xiong, C., & Socher, R. (2019). Ctrl: A conditional transformer language model for controllable generation. arXiv preprint arXiv:1909.05858.
# Model Card Authors
This model card was written by the team at Hugging Face, referencing the [model card](https://github.com/salesforce/ctrl/blob/master/ModelCard.pdf) released by the developers.
# How to Get Started with the Model
Use the code below to get started with the model. See the [Hugging Face ctrl docs](https://huggingface.co/docs/transformers/model_doc/ctrl) for more information.
<details>
<summary> Click to expand </summary>
```python
>>> from transformers import CTRLTokenizer, CTRLModel
>>> import torch
>>> tokenizer = CTRLTokenizer.from_pretrained("ctrl")
>>> model = CTRLModel.from_pretrained("ctrl")
>>> # CTRL was trained with control codes as the first token
>>> inputs = tokenizer("Opinion My dog is cute", return_tensors="pt")
>>> assert inputs["input_ids"][0, 0].item() in tokenizer.control_codes.values()
>>> outputs = model(**inputs)
>>> last_hidden_states = outputs.last_hidden_state
>>> list(last_hidden_states.shape)
```
</details> |
DeepPavlov/distilrubert-tiny-cased-conversational-v1 | 2033d0d1de807e8181ebfa0e53d2a8e526412b0f | 2022-05-06T11:57:05.000Z | [
"pytorch",
"distilbert",
"ru",
"arxiv:2205.02340",
"transformers"
] | null | false | DeepPavlov | null | DeepPavlov/distilrubert-tiny-cased-conversational-v1 | 15,712 | 1 | transformers | 552 | ---
language:
- ru
---
# distilrubert-tiny-cased-conversational
Conversational DistilRuBERT-tiny \(Russian, cased, 3‑layers, 264‑hidden, 12‑heads, 10.4M parameters\) was trained on OpenSubtitles\[1\], [Dirty](https://d3.ru/), [Pikabu](https://pikabu.ru/), and a Social Media segment of Taiga corpus\[2\] (as [Conversational RuBERT](https://huggingface.co/DeepPavlov/rubert-base-cased-conversational)). It can be considered as tiny copy of [Conversational DistilRuBERT-small](https://huggingface.co/DeepPavlov/distilrubert-tiny-cased-conversational).
Our DistilRuBERT-tiny is highly inspired by \[3\], \[4\] and architecture is very close to \[5\]. Namely, we use
* MLM loss (between token labels and student output distribution)
* MSE loss (between averaged student and teacher hidden states)
The key features are:
* unlike most of distilled language models, we **didn't** use KL loss during pre-training
* reduced vocabulary size (30K in *tiny* vs. 100K in *base* and *small* )
* two separate inputs for student: tokens obtained using student tokenizer (for MLM) and teacher tokens greedily splitted by student tokens (for MSE)
Here is comparison between teacher model (`Conversational RuBERT`) and other distilled models.
| Model name | \# params, M | \# vocab, K | Mem., MB |
|---|---|---|---|
| `rubert-base-cased-conversational` | 177.9 | 120 | 679 |
| `distilrubert-base-cased-conversational` | 135.5 | 120 | 517 |
| `distilrubert-small-cased-conversational` | 107.1 | 120 | 409 |
| `cointegrated/rubert-tiny` | 11.8 | **30** | 46 |
| **distilrubert-tiny-cased-conversational** | **10.4** | 31 | **41** |
DistilRuBERT-tiny was trained for about 100 hrs. on 7 nVIDIA Tesla P100-SXM2.0 16Gb.
We used `PyTorchBenchmark` from `transformers` to evaluate model's performance and compare it with other pre-trained language models for Russian. All tests were performed on Intel(R) Xeon(R) CPU E5-2698 v4 @ 2.20GHz and nVIDIA Tesla P100-SXM2.0 16Gb.
| Model name | Batch size | Seq len | Time, s || Mem, MB ||
|---|---|---|------||------||
| | | | CPU | GPU | CPU | GPU |
| `rubert-base-cased-conversational` | 1 | 512 | 0.147 | 0.014 | 897 | 1531 |
| `distilrubert-base-cased-conversational` | 1 | 512 | 0.083 | 0.006 | 766 | 1423 |
| `distilrubert-small-cased-conversational` | 1 | 512 | 0.03 | **0.002** | 600 | 1243 |
| `cointegrated/rubert-tiny` | 1 | 512 | 0.041 | 0.003 | 272 | 919 |
| **distilrubert-tiny-cased-conversational** | 1 | 512 | **0.023** | 0.003 | **206** | **855** |
| `rubert-base-cased-conversational` | 16 | 512 | 2.839 | 0.182 | 1499 | 2071 |
| `distilrubert-base-cased-conversational` | 16 | 512 | 1.065 | 0.055 | 2541 | 2927 |
| `distilrubert-small-cased-conversational` | 16 | 512 | 0.373 | **0.003** | 1360 | 1943 |
| `cointegrated/rubert-tiny` | 16 | 512 | 0.628 | 0.004 | 1293 | 2221 |
| **distilrubert-tiny-cased-conversational** | 16 | 512 | **0.219** | **0.003** | **633** | **1291** |
To evaluate model quality, we fine-tuned DistilRuBERT-tiny on classification (RuSentiment, ParaPhraser), NER and question answering data sets for Russian and obtained scores very similar to the [Conversational DistilRuBERT-small](https://huggingface.co/DeepPavlov/distilrubert-tiny-cased-conversational).
# Citation
If you found the model useful for your research, we are kindly ask to cite [this](https://arxiv.org/abs/2205.02340) paper:
```
@misc{https://doi.org/10.48550/arxiv.2205.02340,
doi = {10.48550/ARXIV.2205.02340},
url = {https://arxiv.org/abs/2205.02340},
author = {Kolesnikova, Alina and Kuratov, Yuri and Konovalov, Vasily and Burtsev, Mikhail},
keywords = {Computation and Language (cs.CL), Machine Learning (cs.LG), FOS: Computer and information sciences, FOS: Computer and information sciences},
title = {Knowledge Distillation of Russian Language Models with Reduction of Vocabulary},
publisher = {arXiv},
year = {2022},
copyright = {arXiv.org perpetual, non-exclusive license}
}
```
\[1\]: P. Lison and J. Tiedemann, 2016, OpenSubtitles2016: Extracting Large Parallel Corpora from Movie and TV Subtitles. In Proceedings of the 10th International Conference on Language Resources and Evaluation \(LREC 2016\)
\[2\]: Shavrina T., Shapovalova O. \(2017\) TO THE METHODOLOGY OF CORPUS CONSTRUCTION FOR MACHINE LEARNING: «TAIGA» SYNTAX TREE CORPUS AND PARSER. in proc. of “CORPORA2017”, international conference , Saint-Petersbourg, 2017.
\[3\]: Sanh, V., Debut, L., Chaumond, J., & Wolf, T. \(2019\). DistilBERT, a distilled version of BERT: smaller, faster, cheaper and lighter. arXiv preprint arXiv:1910.01108.
\[4\]: <https://github.com/huggingface/transformers/tree/master/examples/research_projects/distillation>
\[5\]: <https://habr.com/ru/post/562064/>, <https://huggingface.co/cointegrated/rubert-tiny> |
KBLab/bert-base-swedish-cased | 087b126005a1eee27f85e9665401161dabb4665d | 2022-07-28T14:11:35.000Z | [
"pytorch",
"tf",
"jax",
"bert",
"fill-mask",
"sv",
"transformers",
"autotrain_compatible"
] | fill-mask | false | KBLab | null | KBLab/bert-base-swedish-cased | 15,649 | 4 | transformers | 553 | ---
language: sv
---
# Swedish BERT Models
The National Library of Sweden / KBLab releases three pretrained language models based on BERT and ALBERT. The models are trained on aproximately 15-20GB of text (200M sentences, 3000M tokens) from various sources (books, news, government publications, swedish wikipedia and internet forums) aiming to provide a representative BERT model for Swedish text. A more complete description will be published later on.
The following three models are currently available:
- **bert-base-swedish-cased** (*v1*) - A BERT trained with the same hyperparameters as first published by Google.
- **bert-base-swedish-cased-ner** (*experimental*) - a BERT fine-tuned for NER using SUC 3.0.
- **albert-base-swedish-cased-alpha** (*alpha*) - A first attempt at an ALBERT for Swedish.
All models are cased and trained with whole word masking.
## Files
| **name** | **files** |
|---------------------------------|-----------|
| bert-base-swedish-cased | [config](https://s3.amazonaws.com/models.huggingface.co/bert/KB/bert-base-swedish-cased/config.json), [vocab](https://s3.amazonaws.com/models.huggingface.co/bert/KB/bert-base-swedish-cased/vocab.txt), [pytorch_model.bin](https://s3.amazonaws.com/models.huggingface.co/bert/KB/bert-base-swedish-cased/pytorch_model.bin) |
| bert-base-swedish-cased-ner | [config](https://s3.amazonaws.com/models.huggingface.co/bert/KB/bert-base-swedish-cased-ner/config.json), [vocab](https://s3.amazonaws.com/models.huggingface.co/bert/KB/bert-base-swedish-cased-ner/vocab.txt) [pytorch_model.bin](https://s3.amazonaws.com/models.huggingface.co/bert/KB/bert-base-swedish-cased-ner/pytorch_model.bin) |
| albert-base-swedish-cased-alpha | [config](https://s3.amazonaws.com/models.huggingface.co/bert/KB/albert-base-swedish-cased-alpha/config.json), [sentencepiece model](https://s3.amazonaws.com/models.huggingface.co/bert/KB/albert-base-swedish-cased-alpha/spiece.model), [pytorch_model.bin](https://s3.amazonaws.com/models.huggingface.co/bert/KB/albert-base-swedish-cased-alpha/pytorch_model.bin) |
TensorFlow model weights will be released soon.
## Usage requirements / installation instructions
The examples below require Huggingface Transformers 2.4.1 and Pytorch 1.3.1 or greater. For Transformers<2.4.0 the tokenizer must be instantiated manually and the `do_lower_case` flag parameter set to `False` and `keep_accents` to `True` (for ALBERT).
To create an environment where the examples can be run, run the following in an terminal on your OS of choice.
```
# git clone https://github.com/Kungbib/swedish-bert-models
# cd swedish-bert-models
# python3 -m venv venv
# source venv/bin/activate
# pip install --upgrade pip
# pip install -r requirements.txt
```
### BERT Base Swedish
A standard BERT base for Swedish trained on a variety of sources. Vocabulary size is ~50k. Using Huggingface Transformers the model can be loaded in Python as follows:
```python
from transformers import AutoModel,AutoTokenizer
tok = AutoTokenizer.from_pretrained('KBLab/bert-base-swedish-cased')
model = AutoModel.from_pretrained('KBLab/bert-base-swedish-cased')
```
### BERT base fine-tuned for Swedish NER
This model is fine-tuned on the SUC 3.0 dataset. Using the Huggingface pipeline the model can be easily instantiated. For Transformer<2.4.1 it seems the tokenizer must be loaded separately to disable lower-casing of input strings:
```python
from transformers import pipeline
nlp = pipeline('ner', model='KB/bert-base-swedish-cased-ner', tokenizer='KB/bert-base-swedish-cased-ner')
nlp('Idag släpper KB tre språkmodeller.')
```
Running the Python code above should produce in something like the result below. Entity types used are `TME` for time, `PRS` for personal names, `LOC` for locations, `EVN` for events and `ORG` for organisations. These labels are subject to change.
```python
[ { 'word': 'Idag', 'score': 0.9998126029968262, 'entity': 'TME' },
{ 'word': 'KB', 'score': 0.9814832210540771, 'entity': 'ORG' } ]
```
The BERT tokenizer often splits words into multiple tokens, with the subparts starting with `##`, for example the string `Engelbert kör Volvo till Herrängens fotbollsklubb` gets tokenized as `Engel ##bert kör Volvo till Herr ##ängens fotbolls ##klubb`. To glue parts back together one can use something like this:
```python
text = 'Engelbert tar Volvon till Tele2 Arena för att titta på Djurgården IF ' +\
'som spelar fotboll i VM klockan två på kvällen.'
l = []
for token in nlp(text):
if token['word'].startswith('##'):
l[-1]['word'] += token['word'][2:]
else:
l += [ token ]
print(l)
```
Which should result in the following (though less cleanly formated):
```python
[ { 'word': 'Engelbert', 'score': 0.99..., 'entity': 'PRS'},
{ 'word': 'Volvon', 'score': 0.99..., 'entity': 'OBJ'},
{ 'word': 'Tele2', 'score': 0.99..., 'entity': 'LOC'},
{ 'word': 'Arena', 'score': 0.99..., 'entity': 'LOC'},
{ 'word': 'Djurgården', 'score': 0.99..., 'entity': 'ORG'},
{ 'word': 'IF', 'score': 0.99..., 'entity': 'ORG'},
{ 'word': 'VM', 'score': 0.99..., 'entity': 'EVN'},
{ 'word': 'klockan', 'score': 0.99..., 'entity': 'TME'},
{ 'word': 'två', 'score': 0.99..., 'entity': 'TME'},
{ 'word': 'på', 'score': 0.99..., 'entity': 'TME'},
{ 'word': 'kvällen', 'score': 0.54..., 'entity': 'TME'} ]
```
### ALBERT base
The easisest way to do this is, again, using Huggingface Transformers:
```python
from transformers import AutoModel,AutoTokenizer
tok = AutoTokenizer.from_pretrained('KBLab/albert-base-swedish-cased-alpha'),
model = AutoModel.from_pretrained('KBLab/albert-base-swedish-cased-alpha')
```
## Acknowledgements ❤️
- Resources from Stockholms University, Umeå University and Swedish Language Bank at Gothenburg University were used when fine-tuning BERT for NER.
- Model pretraining was made partly in-house at the KBLab and partly (for material without active copyright) with the support of Cloud TPUs from Google's TensorFlow Research Cloud (TFRC).
- Models are hosted on S3 by Huggingface 🤗
|
iarfmoose/bert-base-cased-qa-evaluator | edfb0e29d78453325a95d9a61d4d26d3598e402b | 2021-05-19T20:15:52.000Z | [
"pytorch",
"tf",
"jax",
"bert",
"text-classification",
"transformers"
] | text-classification | false | iarfmoose | null | iarfmoose/bert-base-cased-qa-evaluator | 15,628 | 4 | transformers | 554 | # BERT-base-cased-qa-evaluator
This model takes a question answer pair as an input and outputs a value representing its prediction about whether the input was a valid question and answer pair or not. The model is a pretrained [BERT-base-cased](https://huggingface.co/bert-base-cased) with a sequence classification head.
## Intended uses
The QA evaluator was originally designed to be used with the [t5-base-question-generator](https://huggingface.co/iarfmoose/t5-base-question-generator) for evaluating the quality of generated questions.
The input for the QA evaluator follows the format for `BertForSequenceClassification`, but using the question and answer as the two sequences. Inputs should take the following format:
```
[CLS] <question> [SEP] <answer [SEP]
```
## Limitations and bias
The model is trained to evaluate if a question and answer are semantically related, but cannot determine whether an answer is actually true/correct or not.
## Training data
The training data was made up of question-answer pairs from the following datasets:
- [SQuAD](https://rajpurkar.github.io/SQuAD-explorer/)
- [RACE](http://www.cs.cmu.edu/~glai1/data/race/)
- [CoQA](https://stanfordnlp.github.io/coqa/)
- [MSMARCO](https://microsoft.github.io/msmarco/)
## Training procedure
The question and answer were concatenated 50% of the time. In the other 50% of the time a corruption operation was performed (either swapping the answer for an unrelated answer, or by copying part of the question into the answer). The model was then trained to predict whether the input sequence represented one of the original QA pairs or a corrupted input.
|
aubmindlab/bert-base-arabertv2 | 599b85458968e0cbad56126802f8328e649b3bec | 2022-04-06T15:22:30.000Z | [
"pytorch",
"tf",
"jax",
"bert",
"fill-mask",
"ar",
"dataset:wikipedia",
"dataset:OSIAN",
"dataset:1.5B Arabic Corpus",
"dataset:OSCAR Arabic Unshuffled",
"arxiv:2003.00104",
"transformers",
"autotrain_compatible"
] | fill-mask | false | aubmindlab | null | aubmindlab/bert-base-arabertv2 | 15,620 | 1 | transformers | 555 | ---
language: ar
datasets:
- wikipedia
- OSIAN
- 1.5B Arabic Corpus
- OSCAR Arabic Unshuffled
widget:
- text: " عاصم +ة لبنان هي [MASK] ."
---
# AraBERT v1 & v2 : Pre-training BERT for Arabic Language Understanding
<img src="https://raw.githubusercontent.com/aub-mind/arabert/master/arabert_logo.png" width="100" align="left"/>
**AraBERT** is an Arabic pretrained lanaguage model based on [Google's BERT architechture](https://github.com/google-research/bert). AraBERT uses the same BERT-Base config. More details are available in the [AraBERT Paper](https://arxiv.org/abs/2003.00104) and in the [AraBERT Meetup](https://github.com/WissamAntoun/pydata_khobar_meetup)
There are two versions of the model, AraBERTv0.1 and AraBERTv1, with the difference being that AraBERTv1 uses pre-segmented text where prefixes and suffixes were splitted using the [Farasa Segmenter](http://alt.qcri.org/farasa/segmenter.html).
We evalaute AraBERT models on different downstream tasks and compare them to [mBERT]((https://github.com/google-research/bert/blob/master/multilingual.md)), and other state of the art models (*To the extent of our knowledge*). The Tasks were Sentiment Analysis on 6 different datasets ([HARD](https://github.com/elnagara/HARD-Arabic-Dataset), [ASTD-Balanced](https://www.aclweb.org/anthology/D15-1299), [ArsenTD-Lev](https://staff.aub.edu.lb/~we07/Publications/ArSentD-LEV_Sentiment_Corpus.pdf), [LABR](https://github.com/mohamedadaly/LABR)), Named Entity Recognition with the [ANERcorp](http://curtis.ml.cmu.edu/w/courses/index.php/ANERcorp), and Arabic Question Answering on [Arabic-SQuAD and ARCD](https://github.com/husseinmozannar/SOQAL)
# AraBERTv2
## What's New!
AraBERT now comes in 4 new variants to replace the old v1 versions:
More Detail in the AraBERT folder and in the [README](https://github.com/aub-mind/arabert/blob/master/AraBERT/README.md) and in the [AraBERT Paper](https://arxiv.org/abs/2003.00104v2)
Model | HuggingFace Model Name | Size (MB/Params)| Pre-Segmentation | DataSet (Sentences/Size/nWords) |
---|:---:|:---:|:---:|:---:
AraBERTv0.2-base | [bert-base-arabertv02](https://huggingface.co/aubmindlab/bert-base-arabertv02) | 543MB / 136M | No | 200M / 77GB / 8.6B |
AraBERTv0.2-large| [bert-large-arabertv02](https://huggingface.co/aubmindlab/bert-large-arabertv02) | 1.38G 371M | No | 200M / 77GB / 8.6B |
AraBERTv2-base| [bert-base-arabertv2](https://huggingface.co/aubmindlab/bert-base-arabertv2) | 543MB 136M | Yes | 200M / 77GB / 8.6B |
AraBERTv2-large| [bert-large-arabertv2](https://huggingface.co/aubmindlab/bert-large-arabertv2) | 1.38G 371M | Yes | 200M / 77GB / 8.6B |
AraBERTv0.1-base| [bert-base-arabertv01](https://huggingface.co/aubmindlab/bert-base-arabertv01) | 543MB 136M | No | 77M / 23GB / 2.7B |
AraBERTv1-base| [bert-base-arabert](https://huggingface.co/aubmindlab/bert-base-arabert) | 543MB 136M | Yes | 77M / 23GB / 2.7B |
All models are available in the `HuggingFace` model page under the [aubmindlab](https://huggingface.co/aubmindlab/) name. Checkpoints are available in PyTorch, TF2 and TF1 formats.
## Better Pre-Processing and New Vocab
We identified an issue with AraBERTv1's wordpiece vocabulary. The issue came from punctuations and numbers that were still attached to words when learned the wordpiece vocab. We now insert a space between numbers and characters and around punctuation characters.
The new vocabulary was learnt using the `BertWordpieceTokenizer` from the `tokenizers` library, and should now support the Fast tokenizer implementation from the `transformers` library.
**P.S.**: All the old BERT codes should work with the new BERT, just change the model name and check the new preprocessing dunction
**Please read the section on how to use the [preprocessing function](#Preprocessing)**
## Bigger Dataset and More Compute
We used ~3.5 times more data, and trained for longer.
For Dataset Sources see the [Dataset Section](#Dataset)
Model | Hardware | num of examples with seq len (128 / 512) |128 (Batch Size/ Num of Steps) | 512 (Batch Size/ Num of Steps) | Total Steps | Total Time (in Days) |
---|:---:|:---:|:---:|:---:|:---:|:---:
AraBERTv0.2-base | TPUv3-8 | 420M / 207M | 2560 / 1M | 384/ 2M | 3M | -
AraBERTv0.2-large | TPUv3-128 | 420M / 207M | 13440 / 250K | 2056 / 300K | 550K | 7
AraBERTv2-base | TPUv3-8 | 420M / 207M | 2560 / 1M | 384/ 2M | 3M | -
AraBERTv2-large | TPUv3-128 | 520M / 245M | 13440 / 250K | 2056 / 300K | 550K | 7
AraBERT-base (v1/v0.1) | TPUv2-8 | - |512 / 900K | 128 / 300K| 1.2M | 4
# Dataset
The pretraining data used for the new AraBERT model is also used for Arabic **AraGPT2 and AraELECTRA**.
The dataset consists of 77GB or 200,095,961 lines or 8,655,948,860 words or 82,232,988,358 chars (before applying Farasa Segmentation)
For the new dataset we added the unshuffled OSCAR corpus, after we thoroughly filter it, to the previous dataset used in AraBERTv1 but with out the websites that we previously crawled:
- OSCAR unshuffled and filtered.
- [Arabic Wikipedia dump](https://archive.org/details/arwiki-20190201) from 2020/09/01
- [The 1.5B words Arabic Corpus](https://www.semanticscholar.org/paper/1.5-billion-words-Arabic-Corpus-El-Khair/f3eeef4afb81223df96575adadf808fe7fe440b4)
- [The OSIAN Corpus](https://www.aclweb.org/anthology/W19-4619)
- Assafir news articles. Huge thank you for Assafir for giving us the data
# Preprocessing
It is recommended to apply our preprocessing function before training/testing on any dataset.
**Install farasapy to segment text for AraBERT v1 & v2 `pip install farasapy`**
```python
from arabert.preprocess import ArabertPreprocessor
model_name="bert-base-arabertv2"
arabert_prep = ArabertPreprocessor(model_name=model_name)
text = "ولن نبالغ إذا قلنا إن هاتف أو كمبيوتر المكتب في زمننا هذا ضروري"
arabert_prep.preprocess(text)
>>>"و+ لن نبالغ إذا قل +نا إن هاتف أو كمبيوتر ال+ مكتب في زمن +نا هذا ضروري"
```
## Accepted_models
```
bert-base-arabertv01
bert-base-arabert
bert-base-arabertv02
bert-base-arabertv2
bert-large-arabertv02
bert-large-arabertv2
araelectra-base
aragpt2-base
aragpt2-medium
aragpt2-large
aragpt2-mega
```
# TensorFlow 1.x models
The TF1.x model are available in the HuggingFace models repo.
You can download them as follows:
- via git-lfs: clone all the models in a repo
```bash
curl -s https://packagecloud.io/install/repositories/github/git-lfs/script.deb.sh | sudo bash
sudo apt-get install git-lfs
git lfs install
git clone https://huggingface.co/aubmindlab/MODEL_NAME
tar -C ./MODEL_NAME -zxvf /content/MODEL_NAME/tf1_model.tar.gz
```
where `MODEL_NAME` is any model under the `aubmindlab` name
- via `wget`:
- Go to the tf1_model.tar.gz file on huggingface.co/models/aubmindlab/MODEL_NAME.
- copy the `oid sha256`
- then run `wget https://cdn-lfs.huggingface.co/aubmindlab/aragpt2-base/INSERT_THE_SHA_HERE` (ex: for `aragpt2-base`: `wget https://cdn-lfs.huggingface.co/aubmindlab/aragpt2-base/3766fc03d7c2593ff2fb991d275e96b81b0ecb2098b71ff315611d052ce65248`)
# If you used this model please cite us as :
Google Scholar has our Bibtex wrong (missing name), use this instead
```
@inproceedings{antoun2020arabert,
title={AraBERT: Transformer-based Model for Arabic Language Understanding},
author={Antoun, Wissam and Baly, Fady and Hajj, Hazem},
booktitle={LREC 2020 Workshop Language Resources and Evaluation Conference 11--16 May 2020},
pages={9}
}
```
# Acknowledgments
Thanks to TensorFlow Research Cloud (TFRC) for the free access to Cloud TPUs, couldn't have done it without this program, and to the [AUB MIND Lab](https://sites.aub.edu.lb/mindlab/) Members for the continous support. Also thanks to [Yakshof](https://www.yakshof.com/#/) and Assafir for data and storage access. Another thanks for Habib Rahal (https://www.behance.net/rahalhabib), for putting a face to AraBERT.
# Contacts
**Wissam Antoun**: [Linkedin](https://www.linkedin.com/in/wissam-antoun-622142b4/) | [Twitter](https://twitter.com/wissam_antoun) | [Github](https://github.com/WissamAntoun) | <[email protected]> | <[email protected]>
**Fady Baly**: [Linkedin](https://www.linkedin.com/in/fadybaly/) | [Twitter](https://twitter.com/fadybaly) | [Github](https://github.com/fadybaly) | <[email protected]> | <[email protected]>
|
moussaKam/mbarthez | 303813ade47c885979189a169c1449669fdc546f | 2021-11-15T13:01:46.000Z | [
"pytorch",
"mbart",
"text2text-generation",
"fr",
"arxiv:2010.12321",
"transformers",
"summarization",
"license:apache-2.0",
"fill-mask",
"autotrain_compatible"
] | fill-mask | false | moussaKam | null | moussaKam/mbarthez | 15,549 | 4 | transformers | 556 | ---
tags:
- summarization
language:
- fr
license: apache-2.0
pipeline_tag: "fill-mask"
---
A french sequence to sequence pretrained model based on [BART](https://huggingface.co/facebook/bart-large). <br>
BARThez is pretrained by learning to reconstruct a corrupted input sentence. A corpus of 66GB of french raw text is used to carry out the pretraining. <br>
Unlike already existing BERT-based French language models such as CamemBERT and FlauBERT, BARThez is particularly well-suited for generative tasks (such as abstractive summarization), since not only its encoder but also its decoder is pretrained.
In addition to BARThez that is pretrained from scratch, we continue the pretraining of a multilingual BART [mBART](https://huggingface.co/facebook/mbart-large-cc25) which boosted its performance in both discriminative and generative tasks. We call the french adapted version [mBARThez](https://huggingface.co/moussaKam/mbarthez).
| Model | Architecture | #layers | #params |
| ------------- |:-------------:| :-----:|:-----:|
| [BARThez](https://huggingface.co/moussaKam/barthez) | BASE | 12 | 165M |
| [mBARThez](https://huggingface.co/moussaKam/mbarthez) | LARGE | 24 | 458M |
paper: https://arxiv.org/abs/2010.12321 \
github: https://github.com/moussaKam/BARThez
```
@article{eddine2020barthez,
title={BARThez: a Skilled Pretrained French Sequence-to-Sequence Model},
author={Eddine, Moussa Kamal and Tixier, Antoine J-P and Vazirgiannis, Michalis},
journal={arXiv preprint arXiv:2010.12321},
year={2020}
}
```
|
transfo-xl-wt103 | 62696ec9dfce23d1930af723c40fd921eb5b1255 | 2022-07-22T08:06:44.000Z | [
"pytorch",
"tf",
"transfo-xl",
"text-generation",
"en",
"dataset:wikitext-103",
"arxiv:1901.02860",
"transformers",
"model-index"
] | text-generation | false | null | null | transfo-xl-wt103 | 15,538 | 3 | transformers | 557 | ---
datasets:
- wikitext-103
tags:
- text-generation
language: en
model-index:
- name: transfo-xl-wt103
results: []
task:
name: Text Generation
type: text-generation
---
# Transfo-xl-wt103
## Table of Contents
- [Model Details](#model-details)
- [Uses](#uses)
- [Risks, Limitations and Biases](#risks-limitations-and-biases)
- [Training](#training)
- [Evaluation](#evaluation)
- [Citation Information](#citation-information)
- [How to Get Started With the Model](#how-to-get-started-with-the-model)
## Model Details
**Model Description:**
The Transformer-XL model is a causal (uni-directional) transformer with relative positioning (sinusoïdal) embeddings which can reuse previously computed hidden-states to attend to longer context (memory). This model also uses adaptive softmax inputs and outputs (tied).
- **Developed by:** [Zihang Dai]([email protected]), [Zhilin Yang]([email protected]), [Yiming Yang1]([email protected]), [Jaime Carbonell]([email protected]), [Quoc V. Le]([email protected]), [Ruslan Salakhutdinov]([email protected])
- **Shared by:** HuggingFace team
- **Model Type:** Text Generation
- **Language(s):** English
- **License:** [More information needed]
- **Resources for more information:**
- [Research Paper](https://arxiv.org/pdf/1901.02860.pdf)
- [GitHub Repo](https://github.com/kimiyoung/transformer-xl)
- [HuggingFace Documentation](https://huggingface.co/docs/transformers/model_doc/transfo-xl#transformers.TransfoXLModel)
## Uses
#### Direct Use
This model can be used for text generation.
The authors provide additionally notes about the vocabulary used, in the [associated paper](https://arxiv.org/pdf/1901.02860.pdf):
> We envision interesting applications of Transformer-XL in the fields of text generation, unsupervised feature learning, image and speech modeling.
#### Misuse and Out-of-scope Use
The model should not be used to intentionally create hostile or alienating environments for people. In addition, the model was not trained to be factual or true representations of people or events, and therefore using the model to generate such content is out-of-scope for the abilities of this model.
## Risks, Limitations and Biases
**CONTENT WARNING: Readers should be aware this section contains content that is disturbing, offensive, and can propagate historical and current stereotypes.**
Significant research has explored bias and fairness issues with language models (see, e.g., [Sheng et al. (2021)](https://aclanthology.org/2021.acl-long.330.pdf) and [Bender et al. (2021)](https://dl.acm.org/doi/pdf/10.1145/3442188.3445922)).
## Training
#### Training Data
The authors provide additionally notes about the vocabulary used, in the [associated paper](https://arxiv.org/pdf/1901.02860.pdf):
> best model trained the Wikitext-103 dataset. We seed the our Transformer-XL with a context of at most 512 consecutive tokens randomly sampled from the test set of Wikitext-103. Then, we run Transformer-XL to generate a pre-defined number of tokens (500 or 1,000 in our case). For each generation step, we first find the top-40 probabilities of the next-step distribution and sample from top-40 tokens based on the re-normalized distribution. To help reading, we detokenize the context, the generated text and the reference text.
The authors use the following pretraining corpora for the model, described in the [associated paper](https://arxiv.org/pdf/1901.02860.pdf):
- WikiText-103 (Merity et al., 2016),
#### Training Procedure
##### Preprocessing
The authors provide additionally notes about the training procedure used, in the [associated paper](https://arxiv.org/pdf/1901.02860.pdf):
> Similar to but different from enwik8, text8 con- tains 100M processed Wikipedia characters cre- ated by lowering case the text and removing any character other than the 26 letters a through z, and space. Due to the similarity, we simply adapt the best model and the same hyper-parameters on en- wik8 to text8 without further tuning.
## Evaluation
#### Results
| Method | enwiki8 |text8 | One Billion Word | WT-103 | PTB (w/o finetuning) |
|:--------------------:|---------:|:----:|:----------------:|:------:|:--------------------:|
| Transformer-XL. | 0.99 | 1.08 | 21.8 | 18.3 | 54.5 |
## Citation Information
```bibtex
@misc{https://doi.org/10.48550/arxiv.1901.02860,
doi = {10.48550/ARXIV.1901.02860},
url = {https://arxiv.org/abs/1901.02860},
author = {Dai, Zihang and Yang, Zhilin and Yang, Yiming and Carbonell, Jaime and Le, Quoc V. and Salakhutdinov, Ruslan},
keywords = {Machine Learning (cs.LG), Computation and Language (cs.CL), Machine Learning (stat.ML), FOS: Computer and information sciences, FOS: Computer and information sciences},
title = {Transformer-XL: Attentive Language Models Beyond a Fixed-Length Context},
publisher = {arXiv},
year = {2019},
copyright = {Creative Commons Attribution Non Commercial Share Alike 4.0 International}
}
```
## How to Get Started With the Model
```
from transformers import TransfoXLTokenizer, TransfoXLModel
import torch
tokenizer = TransfoXLTokenizer.from_pretrained("transfo-xl-wt103")
model = TransfoXLModel.from_pretrained("transfo-xl-wt103")
inputs = tokenizer("Hello, my dog is cute", return_tensors="pt")
outputs = model(**inputs)
last_hidden_states = outputs.last_hidden_state
```
|
sentence-transformers/stsb-roberta-large | b6cd86898ba049a6d160dab42e298f677c5e63b6 | 2022-06-15T20:28:37.000Z | [
"pytorch",
"tf",
"jax",
"roberta",
"feature-extraction",
"arxiv:1908.10084",
"sentence-transformers",
"sentence-similarity",
"transformers",
"license:apache-2.0"
] | sentence-similarity | false | sentence-transformers | null | sentence-transformers/stsb-roberta-large | 15,533 | 1 | sentence-transformers | 558 | ---
pipeline_tag: sentence-similarity
license: apache-2.0
tags:
- sentence-transformers
- feature-extraction
- sentence-similarity
- transformers
---
**⚠️ This model is deprecated. Please don't use it as it produces sentence embeddings of low quality. You can find recommended sentence embedding models here: [SBERT.net - Pretrained Models](https://www.sbert.net/docs/pretrained_models.html)**
# sentence-transformers/stsb-roberta-large
This is a [sentence-transformers](https://www.SBERT.net) model: It maps sentences & paragraphs to a 1024 dimensional dense vector space and can be used for tasks like clustering or semantic search.
## Usage (Sentence-Transformers)
Using this model becomes easy when you have [sentence-transformers](https://www.SBERT.net) installed:
```
pip install -U sentence-transformers
```
Then you can use the model like this:
```python
from sentence_transformers import SentenceTransformer
sentences = ["This is an example sentence", "Each sentence is converted"]
model = SentenceTransformer('sentence-transformers/stsb-roberta-large')
embeddings = model.encode(sentences)
print(embeddings)
```
## Usage (HuggingFace Transformers)
Without [sentence-transformers](https://www.SBERT.net), you can use the model like this: First, you pass your input through the transformer model, then you have to apply the right pooling-operation on-top of the contextualized word embeddings.
```python
from transformers import AutoTokenizer, AutoModel
import torch
#Mean Pooling - Take attention mask into account for correct averaging
def mean_pooling(model_output, attention_mask):
token_embeddings = model_output[0] #First element of model_output contains all token embeddings
input_mask_expanded = attention_mask.unsqueeze(-1).expand(token_embeddings.size()).float()
return torch.sum(token_embeddings * input_mask_expanded, 1) / torch.clamp(input_mask_expanded.sum(1), min=1e-9)
# Sentences we want sentence embeddings for
sentences = ['This is an example sentence', 'Each sentence is converted']
# Load model from HuggingFace Hub
tokenizer = AutoTokenizer.from_pretrained('sentence-transformers/stsb-roberta-large')
model = AutoModel.from_pretrained('sentence-transformers/stsb-roberta-large')
# Tokenize sentences
encoded_input = tokenizer(sentences, padding=True, truncation=True, return_tensors='pt')
# Compute token embeddings
with torch.no_grad():
model_output = model(**encoded_input)
# Perform pooling. In this case, max pooling.
sentence_embeddings = mean_pooling(model_output, encoded_input['attention_mask'])
print("Sentence embeddings:")
print(sentence_embeddings)
```
## Evaluation Results
For an automated evaluation of this model, see the *Sentence Embeddings Benchmark*: [https://seb.sbert.net](https://seb.sbert.net?model_name=sentence-transformers/stsb-roberta-large)
## Full Model Architecture
```
SentenceTransformer(
(0): Transformer({'max_seq_length': 128, 'do_lower_case': True}) with Transformer model: RobertaModel
(1): Pooling({'word_embedding_dimension': 1024, 'pooling_mode_cls_token': False, 'pooling_mode_mean_tokens': True, 'pooling_mode_max_tokens': False, 'pooling_mode_mean_sqrt_len_tokens': False})
)
```
## Citing & Authors
This model was trained by [sentence-transformers](https://www.sbert.net/).
If you find this model helpful, feel free to cite our publication [Sentence-BERT: Sentence Embeddings using Siamese BERT-Networks](https://arxiv.org/abs/1908.10084):
```bibtex
@inproceedings{reimers-2019-sentence-bert,
title = "Sentence-BERT: Sentence Embeddings using Siamese BERT-Networks",
author = "Reimers, Nils and Gurevych, Iryna",
booktitle = "Proceedings of the 2019 Conference on Empirical Methods in Natural Language Processing",
month = "11",
year = "2019",
publisher = "Association for Computational Linguistics",
url = "http://arxiv.org/abs/1908.10084",
}
``` |
jbetker/wav2vec2-large-robust-ft-libritts-voxpopuli | 33e5835aad1c32ac8707971141b65c3fc5ff1904 | 2022-02-25T19:07:57.000Z | [
"pytorch",
"wav2vec2",
"automatic-speech-recognition",
"transformers"
] | automatic-speech-recognition | false | jbetker | null | jbetker/wav2vec2-large-robust-ft-libritts-voxpopuli | 15,508 | 2 | transformers | 559 | This checkpoint is a wav2vec2-large model that is useful for generating transcriptions with punctuation. It is intended for use in building transcriptions for TTS models, where punctuation is very important for prosody.
This model was created by fine-tuning the `facebook/wav2vec2-large-robust-ft-libri-960h` checkpoint on the [libritts](https://research.google/tools/datasets/libri-tts/) and [voxpopuli](https://github.com/facebookresearch/voxpopuli) datasets with a new vocabulary that includes punctuation.
The model gets a respectable WER of 4.45% on the librispeech validation set. The baseline, `facebook/wav2vec2-large-robust-ft-libri-960h`, got 4.3%.
Since the model was fine-tuned on clean audio, it is not well-suited for noisy audio like CommonVoice (though I may upload a checkpoint for that soon too). It still does pretty good, though.
The vocabulary is uploaded to the model hub as well `jbetker/tacotron_symbols`.
Check out my speech transcription script repo, [ocotillo](https://github.com/neonbjb/ocotillo) for usage examples: https://github.com/neonbjb/ocotillo |
sentence-transformers/paraphrase-MiniLM-L12-v2 | 8f010f24d5c0e1ee9735f056d024fcda6557f70f | 2022-06-15T20:18:46.000Z | [
"pytorch",
"tf",
"bert",
"feature-extraction",
"arxiv:1908.10084",
"sentence-transformers",
"sentence-similarity",
"transformers",
"license:apache-2.0"
] | sentence-similarity | false | sentence-transformers | null | sentence-transformers/paraphrase-MiniLM-L12-v2 | 15,505 | 1 | sentence-transformers | 560 | ---
pipeline_tag: sentence-similarity
license: apache-2.0
tags:
- sentence-transformers
- feature-extraction
- sentence-similarity
- transformers
---
# sentence-transformers/paraphrase-MiniLM-L12-v2
This is a [sentence-transformers](https://www.SBERT.net) model: It maps sentences & paragraphs to a 384 dimensional dense vector space and can be used for tasks like clustering or semantic search.
## Usage (Sentence-Transformers)
Using this model becomes easy when you have [sentence-transformers](https://www.SBERT.net) installed:
```
pip install -U sentence-transformers
```
Then you can use the model like this:
```python
from sentence_transformers import SentenceTransformer
sentences = ["This is an example sentence", "Each sentence is converted"]
model = SentenceTransformer('sentence-transformers/paraphrase-MiniLM-L12-v2')
embeddings = model.encode(sentences)
print(embeddings)
```
## Usage (HuggingFace Transformers)
Without [sentence-transformers](https://www.SBERT.net), you can use the model like this: First, you pass your input through the transformer model, then you have to apply the right pooling-operation on-top of the contextualized word embeddings.
```python
from transformers import AutoTokenizer, AutoModel
import torch
#Mean Pooling - Take attention mask into account for correct averaging
def mean_pooling(model_output, attention_mask):
token_embeddings = model_output[0] #First element of model_output contains all token embeddings
input_mask_expanded = attention_mask.unsqueeze(-1).expand(token_embeddings.size()).float()
return torch.sum(token_embeddings * input_mask_expanded, 1) / torch.clamp(input_mask_expanded.sum(1), min=1e-9)
# Sentences we want sentence embeddings for
sentences = ['This is an example sentence', 'Each sentence is converted']
# Load model from HuggingFace Hub
tokenizer = AutoTokenizer.from_pretrained('sentence-transformers/paraphrase-MiniLM-L12-v2')
model = AutoModel.from_pretrained('sentence-transformers/paraphrase-MiniLM-L12-v2')
# Tokenize sentences
encoded_input = tokenizer(sentences, padding=True, truncation=True, return_tensors='pt')
# Compute token embeddings
with torch.no_grad():
model_output = model(**encoded_input)
# Perform pooling. In this case, max pooling.
sentence_embeddings = mean_pooling(model_output, encoded_input['attention_mask'])
print("Sentence embeddings:")
print(sentence_embeddings)
```
## Evaluation Results
For an automated evaluation of this model, see the *Sentence Embeddings Benchmark*: [https://seb.sbert.net](https://seb.sbert.net?model_name=sentence-transformers/paraphrase-MiniLM-L12-v2)
## Full Model Architecture
```
SentenceTransformer(
(0): Transformer({'max_seq_length': 128, 'do_lower_case': False}) with Transformer model: BertModel
(1): Pooling({'word_embedding_dimension': 384, 'pooling_mode_cls_token': False, 'pooling_mode_mean_tokens': True, 'pooling_mode_max_tokens': False, 'pooling_mode_mean_sqrt_len_tokens': False})
)
```
## Citing & Authors
This model was trained by [sentence-transformers](https://www.sbert.net/).
If you find this model helpful, feel free to cite our publication [Sentence-BERT: Sentence Embeddings using Siamese BERT-Networks](https://arxiv.org/abs/1908.10084):
```bibtex
@inproceedings{reimers-2019-sentence-bert,
title = "Sentence-BERT: Sentence Embeddings using Siamese BERT-Networks",
author = "Reimers, Nils and Gurevych, Iryna",
booktitle = "Proceedings of the 2019 Conference on Empirical Methods in Natural Language Processing",
month = "11",
year = "2019",
publisher = "Association for Computational Linguistics",
url = "http://arxiv.org/abs/1908.10084",
}
``` |
textattack/bert-base-uncased-SST-2 | 95f0f6f859b35c8ff0863ae3cd4e2dbc702c0ae2 | 2021-05-20T07:37:12.000Z | [
"pytorch",
"jax",
"bert",
"text-classification",
"transformers"
] | text-classification | false | textattack | null | textattack/bert-base-uncased-SST-2 | 15,351 | null | transformers | 561 | Entry not found |
sentence-transformers/xlm-r-distilroberta-base-paraphrase-v1 | bf5baa9f012675785f0ffb2b509d99ca1b49fcde | 2022-06-15T22:09:18.000Z | [
"pytorch",
"tf",
"xlm-roberta",
"feature-extraction",
"arxiv:1908.10084",
"sentence-transformers",
"sentence-similarity",
"transformers",
"license:apache-2.0"
] | sentence-similarity | false | sentence-transformers | null | sentence-transformers/xlm-r-distilroberta-base-paraphrase-v1 | 15,282 | null | sentence-transformers | 562 | ---
pipeline_tag: sentence-similarity
license: apache-2.0
tags:
- sentence-transformers
- feature-extraction
- sentence-similarity
- transformers
---
# sentence-transformers/xlm-r-distilroberta-base-paraphrase-v1
This is a [sentence-transformers](https://www.SBERT.net) model: It maps sentences & paragraphs to a 768 dimensional dense vector space and can be used for tasks like clustering or semantic search.
## Usage (Sentence-Transformers)
Using this model becomes easy when you have [sentence-transformers](https://www.SBERT.net) installed:
```
pip install -U sentence-transformers
```
Then you can use the model like this:
```python
from sentence_transformers import SentenceTransformer
sentences = ["This is an example sentence", "Each sentence is converted"]
model = SentenceTransformer('sentence-transformers/xlm-r-distilroberta-base-paraphrase-v1')
embeddings = model.encode(sentences)
print(embeddings)
```
## Usage (HuggingFace Transformers)
Without [sentence-transformers](https://www.SBERT.net), you can use the model like this: First, you pass your input through the transformer model, then you have to apply the right pooling-operation on-top of the contextualized word embeddings.
```python
from transformers import AutoTokenizer, AutoModel
import torch
#Mean Pooling - Take attention mask into account for correct averaging
def mean_pooling(model_output, attention_mask):
token_embeddings = model_output[0] #First element of model_output contains all token embeddings
input_mask_expanded = attention_mask.unsqueeze(-1).expand(token_embeddings.size()).float()
return torch.sum(token_embeddings * input_mask_expanded, 1) / torch.clamp(input_mask_expanded.sum(1), min=1e-9)
# Sentences we want sentence embeddings for
sentences = ['This is an example sentence', 'Each sentence is converted']
# Load model from HuggingFace Hub
tokenizer = AutoTokenizer.from_pretrained('sentence-transformers/xlm-r-distilroberta-base-paraphrase-v1')
model = AutoModel.from_pretrained('sentence-transformers/xlm-r-distilroberta-base-paraphrase-v1')
# Tokenize sentences
encoded_input = tokenizer(sentences, padding=True, truncation=True, return_tensors='pt')
# Compute token embeddings
with torch.no_grad():
model_output = model(**encoded_input)
# Perform pooling. In this case, max pooling.
sentence_embeddings = mean_pooling(model_output, encoded_input['attention_mask'])
print("Sentence embeddings:")
print(sentence_embeddings)
```
## Evaluation Results
For an automated evaluation of this model, see the *Sentence Embeddings Benchmark*: [https://seb.sbert.net](https://seb.sbert.net?model_name=sentence-transformers/xlm-r-distilroberta-base-paraphrase-v1)
## Full Model Architecture
```
SentenceTransformer(
(0): Transformer({'max_seq_length': 128, 'do_lower_case': False}) with Transformer model: XLMRobertaModel
(1): Pooling({'word_embedding_dimension': 768, 'pooling_mode_cls_token': False, 'pooling_mode_mean_tokens': True, 'pooling_mode_max_tokens': False, 'pooling_mode_mean_sqrt_len_tokens': False})
)
```
## Citing & Authors
This model was trained by [sentence-transformers](https://www.sbert.net/).
If you find this model helpful, feel free to cite our publication [Sentence-BERT: Sentence Embeddings using Siamese BERT-Networks](https://arxiv.org/abs/1908.10084):
```bibtex
@inproceedings{reimers-2019-sentence-bert,
title = "Sentence-BERT: Sentence Embeddings using Siamese BERT-Networks",
author = "Reimers, Nils and Gurevych, Iryna",
booktitle = "Proceedings of the 2019 Conference on Empirical Methods in Natural Language Processing",
month = "11",
year = "2019",
publisher = "Association for Computational Linguistics",
url = "http://arxiv.org/abs/1908.10084",
}
``` |
m3rg-iitd/matscibert | 24a4e4318dda9bc18bff5e6a45debdcb3e1780e3 | 2022-06-02T19:07:10.000Z | [
"pytorch",
"bert",
"fill-mask",
"transformers",
"autotrain_compatible"
] | fill-mask | false | m3rg-iitd | null | m3rg-iitd/matscibert | 15,267 | 4 | transformers | 563 | # MatSciBERT
## A Materials Domain Language Model for Text Mining and Information Extraction
This is the pretrained model presented in [MatSciBERT: A materials domain language model for text mining and information extraction](https://rdcu.be/cMAp5), which is a BERT model trained on material science research papers.
The training corpus comprises papers related to the broad category of materials: alloys, glasses, metallic glasses, cement and concrete. We have utilised the abstracts and full text of papers(when available). All the research papers have been downloaded from [ScienceDirect](https://www.sciencedirect.com/) using the [Elsevier API](https://dev.elsevier.com/). The detailed methodology is given in the paper.
The codes for pretraining and finetuning on downstream tasks are shared on [GitHub](https://github.com/m3rg-repo/MatSciBERT).
If you find this useful in your research, please consider citing:
```
@article{gupta_matscibert_2022,
title = "{MatSciBERT}: A Materials Domain Language Model for Text Mining and Information Extraction",
author = "Gupta, Tanishq and
Zaki, Mohd and
Krishnan, N. M. Anoop and
Mausam",
year = "2022",
month = may,
journal = "npj Computational Materials",
volume = "8",
number = "1",
pages = "102",
issn = "2057-3960",
url = "https://www.nature.com/articles/s41524-022-00784-w",
doi = "10.1038/s41524-022-00784-w"
}
``` |
pierreguillou/bert-large-cased-squad-v1.1-portuguese | e09c341e0255ed5120a4857db94e3209699353b6 | 2022-01-04T09:57:00.000Z | [
"pytorch",
"tf",
"bert",
"question-answering",
"pt",
"dataset:brWaC",
"dataset:squad",
"dataset:squad_v1_pt",
"transformers",
"bert-large",
"license:mit",
"autotrain_compatible"
] | question-answering | false | pierreguillou | null | pierreguillou/bert-large-cased-squad-v1.1-portuguese | 15,248 | 12 | transformers | 564 | ---
language: pt
license: mit
tags:
- question-answering
- bert
- bert-large
- pytorch
datasets:
- brWaC
- squad
- squad_v1_pt
metrics:
- squad
widget:
- text: "Quando começou a pandemia de Covid-19 no mundo?"
context: "A pandemia de COVID-19, também conhecida como pandemia de coronavírus, é uma pandemia em curso de COVID-19, uma doença respiratória causada pelo coronavírus da síndrome respiratória aguda grave 2 (SARS-CoV-2). O vírus tem origem zoonótica e o primeiro caso conhecido da doença remonta a dezembro de 2019 em Wuhan, na China."
- text: "Onde foi descoberta a Covid-19?"
context: "A pandemia de COVID-19, também conhecida como pandemia de coronavírus, é uma pandemia em curso de COVID-19, uma doença respiratória causada pelo coronavírus da síndrome respiratória aguda grave 2 (SARS-CoV-2). O vírus tem origem zoonótica e o primeiro caso conhecido da doença remonta a dezembro de 2019 em Wuhan, na China."
---
# Portuguese BERT large cased QA (Question Answering), finetuned on SQUAD v1.1
## Introduction
The model was trained on the dataset SQUAD v1.1 in portuguese from the [Deep Learning Brasil group](http://www.deeplearningbrasil.com.br/).
The language model used is the [BERTimbau Large](https://huggingface.co/neuralmind/bert-large-portuguese-cased) (aka "bert-large-portuguese-cased") from [Neuralmind.ai](https://neuralmind.ai/): BERTimbau is a pretrained BERT model for Brazilian Portuguese that achieves state-of-the-art performances on three downstream NLP tasks: Named Entity Recognition, Sentence Textual Similarity and Recognizing Textual Entailment. It is available in two sizes: Base and Large.
## Informations on the method used
All the informations are in the blog post : [NLP | Como treinar um modelo de Question Answering em qualquer linguagem baseado no BERT large, melhorando o desempenho do modelo utilizando o BERT base? (estudo de caso em português)](https://medium.com/@pierre_guillou/nlp-como-treinar-um-modelo-de-question-answering-em-qualquer-linguagem-baseado-no-bert-large-1c899262dd96)
## Notebook in GitHub
[question_answering_BERT_large_cased_squad_v11_pt.ipynb](https://github.com/piegu/language-models/blob/master/question_answering_BERT_large_cased_squad_v11_pt.ipynb) ([nbviewer version](https://nbviewer.jupyter.org/github/piegu/language-models/blob/master/question_answering_BERT_large_cased_squad_v11_pt.ipynb))
## Performance
The results obtained are the following:
```
f1 = 84.43 (against 82.50 for the base model)
exact match = 72.68 (against 70.49 for the base model)
```
## How to use the model... with Pipeline
```python
import transformers
from transformers import pipeline
# source: https://pt.wikipedia.org/wiki/Pandemia_de_COVID-19
context = r"""
A pandemia de COVID-19, também conhecida como pandemia de coronavírus, é uma pandemia em curso de COVID-19,
uma doença respiratória causada pelo coronavírus da síndrome respiratória aguda grave 2 (SARS-CoV-2).
O vírus tem origem zoonótica e o primeiro caso conhecido da doença remonta a dezembro de 2019 em Wuhan, na China.
Em 20 de janeiro de 2020, a Organização Mundial da Saúde (OMS) classificou o surto
como Emergência de Saúde Pública de Âmbito Internacional e, em 11 de março de 2020, como pandemia.
Em 18 de junho de 2021, 177 349 274 casos foram confirmados em 192 países e territórios,
com 3 840 181 mortes atribuídas à doença, tornando-se uma das pandemias mais mortais da história.
Os sintomas de COVID-19 são altamente variáveis, variando de nenhum a doenças com risco de morte.
O vírus se espalha principalmente pelo ar quando as pessoas estão perto umas das outras.
Ele deixa uma pessoa infectada quando ela respira, tosse, espirra ou fala e entra em outra pessoa pela boca, nariz ou olhos.
Ele também pode se espalhar através de superfícies contaminadas.
As pessoas permanecem contagiosas por até duas semanas e podem espalhar o vírus mesmo se forem assintomáticas.
"""
model_name = 'pierreguillou/bert-large-cased-squad-v1.1-portuguese'
nlp = pipeline("question-answering", model=model_name)
question = "Quando começou a pandemia de Covid-19 no mundo?"
result = nlp(question=question, context=context)
print(f"Answer: '{result['answer']}', score: {round(result['score'], 4)}, start: {result['start']}, end: {result['end']}")
# Answer: 'dezembro de 2019', score: 0.5087, start: 290, end: 306
```
## How to use the model... with the Auto classes
```python
from transformers import AutoTokenizer, AutoModelForQuestionAnswering
tokenizer = AutoTokenizer.from_pretrained("pierreguillou/bert-large-cased-squad-v1.1-portuguese")
model = AutoModelForQuestionAnswering.from_pretrained("pierreguillou/bert-large-cased-squad-v1.1-portuguese")
```
Or just clone the model repo:
```python
git lfs install
git clone https://huggingface.co/pierreguillou/bert-large-cased-squad-v1.1-portuguese
# if you want to clone without large files – just their pointers
# prepend your git clone with the following env var:
GIT_LFS_SKIP_SMUDGE=1
```
## Limitations and bias
The training data used for this model come from Portuguese SQUAD. It could contain a lot of unfiltered content, which is far from neutral, and biases.
## Author
Portuguese BERT large cased QA (Question Answering), finetuned on SQUAD v1.1 was trained and evaluated by [Pierre GUILLOU](https://www.linkedin.com/in/pierreguillou/) thanks to the Open Source code, platforms and advices of many organizations ([link to the list](https://medium.com/@pierre_guillou/nlp-como-treinar-um-modelo-de-question-answering-em-qualquer-linguagem-baseado-no-bert-large-1c899262dd96#c2f5)). In particular: [Hugging Face](https://huggingface.co/), [Neuralmind.ai](https://neuralmind.ai/), [Deep Learning Brasil group](http://www.deeplearningbrasil.com.br/) and [AI Lab](https://ailab.unb.br/).
## Citation
If you use our work, please cite:
```bibtex
@inproceedings{pierreguillou2021bertlargecasedsquadv11portuguese,
title={Portuguese BERT large cased QA (Question Answering), finetuned on SQUAD v1.1},
author={Pierre Guillou},
year={2021}
}
``` |
openclimatefix/nowcasting_cnn_v2 | 5b9c2a9194869afae91651ffdf2ba1de1952d866 | 2022-05-25T10:41:01.000Z | [
"pytorch",
"transformers",
"nowcasting",
"forecasting",
"timeseries",
"remote-sensing",
"license:mit"
] | null | false | openclimatefix | null | openclimatefix/nowcasting_cnn_v2 | 15,138 | null | transformers | 565 | ---
license: mit
tags:
- nowcasting
- forecasting
- timeseries
- remote-sensing
---
# Nowcasting CNN
## Model description
3d conv model, that takes in different data streams
architecture is roughly
1. satellite image time series goes into many 3d convolution layers.
2. nwp time series goes into many 3d convolution layers.
3. Final convolutional layer goes to full connected layer. This is joined by
other data inputs like
- pv yield
- time variables
Then there ~4 fully connected layers which end up forecasting the
pv yield / gsp into the future
## Intended uses & limitations
Forecasting short term PV power for different regions and nationally in the UK
## How to use
[More information needed]
## Limitations and bias
[More information needed]
## Training data
Training data is EUMETSAT RSS imagery over the UK, on-the-ground PV data, and NWP predictions.
## Training procedure
[More information needed]
## Evaluation results
[More information needed]
|
sentence-transformers/stsb-distilbert-base | 815bb3e4dbcba340b5f8c0c0489800230880e06e | 2022-06-15T19:43:41.000Z | [
"pytorch",
"tf",
"distilbert",
"feature-extraction",
"arxiv:1908.10084",
"sentence-transformers",
"sentence-similarity",
"transformers",
"license:apache-2.0"
] | sentence-similarity | false | sentence-transformers | null | sentence-transformers/stsb-distilbert-base | 15,046 | 3 | sentence-transformers | 566 | ---
pipeline_tag: sentence-similarity
license: apache-2.0
tags:
- sentence-transformers
- feature-extraction
- sentence-similarity
- transformers
---
**⚠️ This model is deprecated. Please don't use it as it produces sentence embeddings of low quality. You can find recommended sentence embedding models here: [SBERT.net - Pretrained Models](https://www.sbert.net/docs/pretrained_models.html)**
# sentence-transformers/stsb-distilbert-base
This is a [sentence-transformers](https://www.SBERT.net) model: It maps sentences & paragraphs to a 768 dimensional dense vector space and can be used for tasks like clustering or semantic search.
## Usage (Sentence-Transformers)
Using this model becomes easy when you have [sentence-transformers](https://www.SBERT.net) installed:
```
pip install -U sentence-transformers
```
Then you can use the model like this:
```python
from sentence_transformers import SentenceTransformer
sentences = ["This is an example sentence", "Each sentence is converted"]
model = SentenceTransformer('sentence-transformers/stsb-distilbert-base')
embeddings = model.encode(sentences)
print(embeddings)
```
## Usage (HuggingFace Transformers)
Without [sentence-transformers](https://www.SBERT.net), you can use the model like this: First, you pass your input through the transformer model, then you have to apply the right pooling-operation on-top of the contextualized word embeddings.
```python
from transformers import AutoTokenizer, AutoModel
import torch
#Mean Pooling - Take attention mask into account for correct averaging
def mean_pooling(model_output, attention_mask):
token_embeddings = model_output[0] #First element of model_output contains all token embeddings
input_mask_expanded = attention_mask.unsqueeze(-1).expand(token_embeddings.size()).float()
return torch.sum(token_embeddings * input_mask_expanded, 1) / torch.clamp(input_mask_expanded.sum(1), min=1e-9)
# Sentences we want sentence embeddings for
sentences = ['This is an example sentence', 'Each sentence is converted']
# Load model from HuggingFace Hub
tokenizer = AutoTokenizer.from_pretrained('sentence-transformers/stsb-distilbert-base')
model = AutoModel.from_pretrained('sentence-transformers/stsb-distilbert-base')
# Tokenize sentences
encoded_input = tokenizer(sentences, padding=True, truncation=True, return_tensors='pt')
# Compute token embeddings
with torch.no_grad():
model_output = model(**encoded_input)
# Perform pooling. In this case, max pooling.
sentence_embeddings = mean_pooling(model_output, encoded_input['attention_mask'])
print("Sentence embeddings:")
print(sentence_embeddings)
```
## Evaluation Results
For an automated evaluation of this model, see the *Sentence Embeddings Benchmark*: [https://seb.sbert.net](https://seb.sbert.net?model_name=sentence-transformers/stsb-distilbert-base)
## Full Model Architecture
```
SentenceTransformer(
(0): Transformer({'max_seq_length': 128, 'do_lower_case': False}) with Transformer model: DistilBertModel
(1): Pooling({'word_embedding_dimension': 768, 'pooling_mode_cls_token': False, 'pooling_mode_mean_tokens': True, 'pooling_mode_max_tokens': False, 'pooling_mode_mean_sqrt_len_tokens': False})
)
```
## Citing & Authors
This model was trained by [sentence-transformers](https://www.sbert.net/).
If you find this model helpful, feel free to cite our publication [Sentence-BERT: Sentence Embeddings using Siamese BERT-Networks](https://arxiv.org/abs/1908.10084):
```bibtex
@inproceedings{reimers-2019-sentence-bert,
title = "Sentence-BERT: Sentence Embeddings using Siamese BERT-Networks",
author = "Reimers, Nils and Gurevych, Iryna",
booktitle = "Proceedings of the 2019 Conference on Empirical Methods in Natural Language Processing",
month = "11",
year = "2019",
publisher = "Association for Computational Linguistics",
url = "http://arxiv.org/abs/1908.10084",
}
``` |
mrm8488/bert-tiny-finetuned-sms-spam-detection | 49541ae4a47b52d4b2e6d5a3a1875c5e35aebeb1 | 2021-05-20T00:40:14.000Z | [
"pytorch",
"jax",
"bert",
"text-classification",
"en",
"dataset:sms_spam",
"transformers",
"sms",
"spam",
"detection"
] | text-classification | false | mrm8488 | null | mrm8488/bert-tiny-finetuned-sms-spam-detection | 14,989 | 5 | transformers | 567 | ---
language: en
tags:
- sms
- spam
- detection
datasets:
- sms_spam
widget:
- text: "Camera - You are awarded a SiPix Digital Camera! call 09061221066 fromm landline. Delivery within 28 days."
---
# BERT-Tiny fine-tuned on on sms_spam dataset for spam detection
Validation accuray: **0.98** |
sberbank-ai/rugpt3large_based_on_gpt2 | aa2b602c1939938541eed9283347d6e08536f6f8 | 2021-09-21T19:33:09.000Z | [
"pytorch",
"jax",
"gpt2",
"text-generation",
"ru",
"transformers",
"PyTorch",
"Transformers"
] | text-generation | false | sberbank-ai | null | sberbank-ai/rugpt3large_based_on_gpt2 | 14,965 | 10 | transformers | 568 | ---
language:
- ru
tags:
- PyTorch
- Transformers
thumbnail: "https://github.com/sberbank-ai/ru-gpts"
---
# rugpt3large\_based\_on\_gpt2
Model was trained with sequence length 1024 using transformers lib by [SberDevices](https://sberdevices.ru/) team on 80B tokens for 3 epochs. After that model was finetuned 1 epoch with sequence length 2048.
Total training time was around 14 days on 128 GPUs for 1024 context and few days on 16 GPUs for 2048 context.
Final perplexity on test set is `13.6`. |
pysentimiento/robertuito-emotion-analysis | 2a1fb82f525912c23a8187eeea418751049d5056 | 2021-12-10T13:30:24.000Z | [
"pytorch",
"roberta",
"text-classification",
"es",
"arxiv:2106.09462",
"arxiv:2111.09453",
"transformers",
"emotion-analysis",
"twitter"
] | text-classification | false | pysentimiento | null | pysentimiento/robertuito-emotion-analysis | 14,906 | 4 | transformers | 569 | ---
language:
- es
tags:
- emotion-analysis
- twitter
---
# Emotion Analysis in Spanish
## robertuito-emotion-analysis
Repository: [https://github.com/pysentimiento/pysentimiento/](https://github.com/finiteautomata/pysentimiento/)
Model trained with TASS 2020 Task 2 corpus for Emotion detection in Spanish. Base model is [RoBERTuito](https://github.com/pysentimiento/robertuito), a RoBERTa model trained in Spanish tweets.
Contains the six Ekman emotions plus a neutral class:
- anger
- disgust
- fear
- joy
- sadness
- surprise
## Results
Results for the four tasks evaluated in `pysentimiento`. Results are expressed as Macro F1 scores
| model | emotion | hate_speech | irony | sentiment |
|:--------------|:--------------|:--------------|:--------------|:--------------|
| robertuito | 0.560 ± 0.010 | 0.759 ± 0.007 | 0.739 ± 0.005 | 0.705 ± 0.003 |
| roberta | 0.527 ± 0.015 | 0.741 ± 0.012 | 0.721 ± 0.008 | 0.670 ± 0.006 |
| bertin | 0.524 ± 0.007 | 0.738 ± 0.007 | 0.713 ± 0.012 | 0.666 ± 0.005 |
| beto_uncased | 0.532 ± 0.012 | 0.727 ± 0.016 | 0.701 ± 0.007 | 0.651 ± 0.006 |
| beto_cased | 0.516 ± 0.012 | 0.724 ± 0.012 | 0.705 ± 0.009 | 0.662 ± 0.005 |
| mbert_uncased | 0.493 ± 0.010 | 0.718 ± 0.011 | 0.681 ± 0.010 | 0.617 ± 0.003 |
| biGRU | 0.264 ± 0.007 | 0.592 ± 0.018 | 0.631 ± 0.011 | 0.585 ± 0.011 |
Note that for Hate Speech, these are the results for Semeval 2019, Task 5 Subtask B (HS+TR+AG detection)
## Citation
If you use this model in your research, please cite pysentimiento, RoBERTuito and EmoEvent papers:
```
@misc{perez2021pysentimiento,
title={pysentimiento: A Python Toolkit for Sentiment Analysis and SocialNLP tasks},
author={Juan Manuel Pérez and Juan Carlos Giudici and Franco Luque},
year={2021},
eprint={2106.09462},
archivePrefix={arXiv},
primaryClass={cs.CL}
}
@inproceedings{del2020emoevent,
title={EmoEvent: A multilingual emotion corpus based on different events},
author={del Arco, Flor Miriam Plaza and Strapparava, Carlo and Lopez, L Alfonso Urena and Mart{\'\i}n-Valdivia, M Teresa},
booktitle={Proceedings of the 12th Language Resources and Evaluation Conference},
pages={1492--1498},
year={2020}
}
@misc{perez2021robertuito,
title={RoBERTuito: a pre-trained language model for social media text in Spanish},
author={Juan Manuel Pérez and Damián A. Furman and Laura Alonso Alemany and Franco Luque},
year={2021},
eprint={2111.09453},
archivePrefix={arXiv},
primaryClass={cs.CL}
}
``` |
mrm8488/bert-tiny-finetuned-squadv2 | 178467c709f17b74c9bf05ed6d41cb5fa2cf684c | 2022-07-15T09:46:32.000Z | [
"pytorch",
"jax",
"bert",
"question-answering",
"en",
"arxiv:1908.08962",
"transformers",
"QA",
"autotrain_compatible"
] | question-answering | false | mrm8488 | null | mrm8488/bert-tiny-finetuned-squadv2 | 14,890 | 1 | transformers | 570 | ---
language: en
thumbnail:
tags:
- QA
---
# BERT-Tiny fine-tuned on SQuAD v2
[BERT-Tiny](https://github.com/google-research/bert/) created by [Google Research](https://github.com/google-research) and fine-tuned on [SQuAD 2.0](https://rajpurkar.github.io/SQuAD-explorer/) for **Q&A** downstream task.
**Mode size** (after training): **16.74 MB**
## Details of BERT-Tiny and its 'family' (from their documentation)
Released on March 11th, 2020
This is model is a part of 24 smaller BERT models (English only, uncased, trained with WordPiece masking) referenced in [Well-Read Students Learn Better: On the Importance of Pre-training Compact Models](https://arxiv.org/abs/1908.08962).
The smaller BERT models are intended for environments with restricted computational resources. They can be fine-tuned in the same manner as the original BERT models. However, they are most effective in the context of knowledge distillation, where the fine-tuning labels are produced by a larger and more accurate teacher.
## Details of the downstream task (Q&A) - Dataset
[SQuAD2.0](https://rajpurkar.github.io/SQuAD-explorer/) combines the 100,000 questions in SQuAD1.1 with over 50,000 unanswerable questions written adversarially by crowdworkers to look similar to answerable ones. To do well on SQuAD2.0, systems must not only answer questions when possible, but also determine when no answer is supported by the paragraph and abstain from answering.
| Dataset | Split | # samples |
| -------- | ----- | --------- |
| SQuAD2.0 | train | 130k |
| SQuAD2.0 | eval | 12.3k |
## Model training
The model was trained on a Tesla P100 GPU and 25GB of RAM.
The script for fine tuning can be found [here](https://github.com/huggingface/transformers/tree/main/examples/legacy/question-answering)
## Results:
| Metric | # Value |
| ------ | --------- |
| **EM** | **48.60** |
| **F1** | **49.73** |
| Model | EM | F1 score | SIZE (MB) |
| ----------------------------------------------------------------------------------------- | --------- | --------- | --------- |
| [bert-tiny-finetuned-squadv2](https://huggingface.co/mrm8488/bert-tiny-finetuned-squadv2) | 48.60 | 49.73 | **16.74** |
| [bert-tiny-5-finetuned-squadv2](https://huggingface.co/mrm8488/bert-tiny-5-finetuned-squadv2) | **57.12** | **60.86** | 24.34
## Model in action
Fast usage with **pipelines**:
```python
from transformers import pipeline
qa_pipeline = pipeline(
"question-answering",
model="mrm8488/bert-tiny-finetuned-squadv2",
tokenizer="mrm8488/bert-tiny-finetuned-squadv2"
)
qa_pipeline({
'context': "Manuel Romero has been working hardly in the repository hugginface/transformers lately",
'question': "Who has been working hard for hugginface/transformers lately?"
})
# Output:
```
```json
{
"answer": "Manuel Romero",
"end": 13,
"score": 0.05684709993458714,
"start": 0
}
```
### Yes! That was easy 🎉 Let's try with another example
```python
qa_pipeline({
'context': "Manuel Romero has been working hardly in the repository hugginface/transformers lately",
'question': "For which company has worked Manuel Romero?"
})
# Output:
```
```json
{
"answer": "hugginface/transformers",
"end": 79,
"score": 0.11613431826808274,
"start": 56
}
```
### It works!! 🎉 🎉 🎉
> Created by [Manuel Romero/@mrm8488](https://twitter.com/mrm8488) | [LinkedIn](https://www.linkedin.com/in/manuel-romero-cs/)
> Made with <span style="color: #e25555;">♥</span> in Spain
|
valhalla/t5-base-qa-qg-hl | 0286be61d8d9de5650fdd21ed8923a7bc226e704 | 2020-12-11T22:03:44.000Z | [
"pytorch",
"t5",
"text2text-generation",
"dataset:squad",
"arxiv:1910.10683",
"transformers",
"question-generation",
"license:mit",
"autotrain_compatible"
] | text2text-generation | false | valhalla | null | valhalla/t5-base-qa-qg-hl | 14,782 | 2 | transformers | 571 | ---
datasets:
- squad
tags:
- question-generation
widget:
- text: "generate question: <hl> 42 <hl> is the answer to life, the universe and everything. </s>"
- text: "question: What is 42 context: 42 is the answer to life, the universe and everything. </s>"
license: mit
---
## T5 for multi-task QA and QG
This is multi-task [t5-base](https://arxiv.org/abs/1910.10683) model trained for question answering and answer aware question generation tasks.
For question generation the answer spans are highlighted within the text with special highlight tokens (`<hl>`) and prefixed with 'generate question: '. For QA the input is processed like this `question: question_text context: context_text </s>`
You can play with the model using the inference API. Here's how you can use it
For QG
`generate question: <hl> 42 <hl> is the answer to life, the universe and everything. </s>`
For QA
`question: What is 42 context: 42 is the answer to life, the universe and everything. </s>`
For more deatils see [this](https://github.com/patil-suraj/question_generation) repo.
### Model in action 🚀
You'll need to clone the [repo](https://github.com/patil-suraj/question_generation).
[](https://colab.research.google.com/github/patil-suraj/question_generation/blob/master/question_generation.ipynb)
```python3
from pipelines import pipeline
nlp = pipeline("multitask-qa-qg", model="valhalla/t5-base-qa-qg-hl")
# to generate questions simply pass the text
nlp("42 is the answer to life, the universe and everything.")
=> [{'answer': '42', 'question': 'What is the answer to life, the universe and everything?'}]
# for qa pass a dict with "question" and "context"
nlp({
"question": "What is 42 ?",
"context": "42 is the answer to life, the universe and everything."
})
=> 'the answer to life, the universe and everything'
``` |
Helsinki-NLP/opus-mt-et-en | 3649f4761e2a1b25f2c19e3f1732c6ba9ef61519 | 2021-09-09T21:46:01.000Z | [
"pytorch",
"marian",
"text2text-generation",
"et",
"en",
"transformers",
"translation",
"license:apache-2.0",
"autotrain_compatible"
] | translation | false | Helsinki-NLP | null | Helsinki-NLP/opus-mt-et-en | 14,726 | null | transformers | 572 | ---
tags:
- translation
license: apache-2.0
---
### opus-mt-et-en
* source languages: et
* target languages: en
* OPUS readme: [et-en](https://github.com/Helsinki-NLP/OPUS-MT-train/blob/master/models/et-en/README.md)
* dataset: opus
* model: transformer-align
* pre-processing: normalization + SentencePiece
* download original weights: [opus-2019-12-18.zip](https://object.pouta.csc.fi/OPUS-MT-models/et-en/opus-2019-12-18.zip)
* test set translations: [opus-2019-12-18.test.txt](https://object.pouta.csc.fi/OPUS-MT-models/et-en/opus-2019-12-18.test.txt)
* test set scores: [opus-2019-12-18.eval.txt](https://object.pouta.csc.fi/OPUS-MT-models/et-en/opus-2019-12-18.eval.txt)
## Benchmarks
| testset | BLEU | chr-F |
|-----------------------|-------|-------|
| newsdev2018-enet.et.en | 30.1 | 0.574 |
| newstest2018-enet.et.en | 30.3 | 0.581 |
| Tatoeba.et.en | 59.9 | 0.738 |
|
mrm8488/t5-base-finetuned-emotion | e44a316825f11230724b36412fbf1899c76e82de | 2021-06-23T12:46:24.000Z | [
"pytorch",
"jax",
"t5",
"text2text-generation",
"en",
"dataset:emotion",
"arxiv:1910.10683",
"transformers",
"autotrain_compatible"
] | text2text-generation | false | mrm8488 | null | mrm8488/t5-base-finetuned-emotion | 14,718 | 9 | transformers | 573 | ---
language: en
datasets:
- emotion
widget:
- text: "I wish you were here but it is impossible"
---
# T5-base fine-tuned for Emotion Recognition 😂😢😡😃😯
[Google's T5](https://ai.googleblog.com/2020/02/exploring-transfer-learning-with-t5.html) base fine-tuned on [emotion recognition](https://github.com/dair-ai/emotion_dataset) dataset for **Emotion Recognition** downstream task.
## Details of T5
The **T5** model was presented in [Exploring the Limits of Transfer Learning with a Unified Text-to-Text Transformer](https://arxiv.org/pdf/1910.10683.pdf) by *Colin Raffel, Noam Shazeer, Adam Roberts, Katherine Lee, Sharan Narang, Michael Matena, Yanqi Zhou, Wei Li, Peter J. Liu* in Here the abstract:
Transfer learning, where a model is first pre-trained on a data-rich task before being fine-tuned on a downstream task, has emerged as a powerful technique in natural language processing (NLP). The effectiveness of transfer learning has given rise to a diversity of approaches, methodology, and practice. In this paper, we explore the landscape of transfer learning techniques for NLP by introducing a unified framework that converts every language problem into a text-to-text format. Our systematic study compares pre-training objectives, architectures, unlabeled datasets, transfer approaches, and other factors on dozens of language understanding tasks. By combining the insights from our exploration with scale and our new “Colossal Clean Crawled Corpus”, we achieve state-of-the-art results on many benchmarks covering summarization, question answering, text classification, and more. To facilitate future work on transfer learning for NLP, we release our dataset, pre-trained models, and code.
## Details of the downstream task (Sentiment Recognition) - Dataset 📚
[Elvis Saravia](https://twitter.com/omarsar0) has gathered a great [dataset](https://github.com/dair-ai/emotion_dataset) for emotion recognition. It allows to classifiy the text into one of the following **6** emotions:
- sadness 😢
- joy 😃
- love 🥰
- anger 😡
- fear 😱
- surprise 😯
## Model fine-tuning 🏋️
The training script is a slightly modified version of [this Colab Notebook](https://github.com/patil-suraj/exploring-T5/blob/master/t5_fine_tuning.ipynb) created by [Suraj Patil](https://github.com/patil-suraj), so all credits to him!
## Test set metrics 🧾
| |precision | recall | f1-score |support|
|----------|----------|---------|----------|-------|
|anger | 0.93| 0.92| 0.93| 275|
|fear | 0.91| 0.87| 0.89| 224|
|joy | 0.97| 0.94| 0.95| 695|
|love | 0.80| 0.91| 0.85| 159|
|sadness | 0.97| 0.97| 0.97| 521|
|surpirse | 0.73| 0.89| 0.80| 66|
| |
|accuracy| | | 0.93| 2000|
|macro avg| 0.89| 0.92| 0.90| 2000|
|weighted avg| 0.94| 0.93| 0.93| 2000|
## Model in Action 🚀
```python
from transformers import AutoTokenizer, AutoModelWithLMHead
tokenizer = AutoTokenizer.from_pretrained("mrm8488/t5-base-finetuned-emotion")
model = AutoModelWithLMHead.from_pretrained("mrm8488/t5-base-finetuned-emotion")
def get_emotion(text):
input_ids = tokenizer.encode(text + '</s>', return_tensors='pt')
output = model.generate(input_ids=input_ids,
max_length=2)
dec = [tokenizer.decode(ids) for ids in output]
label = dec[0]
return label
get_emotion("i feel as if i havent blogged in ages are at least truly blogged i am doing an update cute") # Output: 'joy'
get_emotion("i have a feeling i kinda lost my best friend") # Output: 'sadness'
```
> Created by [Manuel Romero/@mrm8488](https://twitter.com/mrm8488) | [LinkedIn](https://www.linkedin.com/in/manuel-romero-cs/)
> Made with <span style="color: #e25555;">♥</span> in Spain
|
iarfmoose/t5-base-question-generator | 1bfc9d4b2b0078e0b65cf40c6e2e2e974fbab6b0 | 2022-02-24T08:41:19.000Z | [
"pytorch",
"tf",
"jax",
"t5",
"text2text-generation",
"transformers",
"autotrain_compatible"
] | text2text-generation | false | iarfmoose | null | iarfmoose/t5-base-question-generator | 14,640 | 17 | transformers | 574 | # Model name
## Model description
This model is a sequence-to-sequence question generator which takes an answer and context as an input, and generates a question as an output. It is based on a pretrained `t5-base` model.
## Intended uses & limitations
The model is trained to generate reading comprehension-style questions with answers extracted from a text. The model performs best with full sentence answers, but can also be used with single word or short phrase answers.
#### How to use
The model takes concatenated answers and context as an input sequence, and will generate a full question sentence as an output sequence. The max sequence length is 512 tokens. Inputs should be organised into the following format:
```
<answer> answer text here <context> context text here
```
The input sequence can then be encoded and passed as the `input_ids` argument in the model's `generate()` method.
For best results, a large number of questions can be generated, and then filtered using [iarfmoose/bert-base-cased-qa-evaluator](https://huggingface.co/iarfmoose/bert-base-cased-qa-evaluator).
For examples, please see https://github.com/iarfmoose/question_generator.
#### Limitations and bias
The model is limited to generating questions in the same style as those found in [SQuAD](https://rajpurkar.github.io/SQuAD-explorer/), [CoQA](https://stanfordnlp.github.io/coqa/), and [MSMARCO](https://microsoft.github.io/msmarco/). The generated questions can potentially be leading or reflect biases that are present in the context. If the context is too short or completely absent, or if the context and answer do not match, the generated question is likely to be incoherent.
## Training data
The model was fine-tuned on a dataset made up of several well-known QA datasets ([SQuAD](https://rajpurkar.github.io/SQuAD-explorer/), [CoQA](https://stanfordnlp.github.io/coqa/), and [MSMARCO](https://microsoft.github.io/msmarco/)). The datasets were restructured by concatenating the answer and context fields into the previously-mentioned format. The question field from the datasets was used as the target during training. The full training set was roughly 200,000 examples.
## Training procedure
The model was trained for 20 epochs over the training set with a learning rate of 1e-3. The batch size was only 4 due to GPU memory limitations when training on Google Colab.
|
kykim/bert-kor-base | 1779cc0982ada0216dd6de0dd4e86fb78201926d | 2021-05-19T21:17:13.000Z | [
"pytorch",
"tf",
"jax",
"bert",
"fill-mask",
"ko",
"transformers",
"autotrain_compatible"
] | fill-mask | false | kykim | null | kykim/bert-kor-base | 14,617 | 3 | transformers | 575 | ---
language: ko
---
# Bert base model for Korean
* 70GB Korean text dataset and 42000 lower-cased subwords are used
* Check the model performance and other language models for Korean in [github](https://github.com/kiyoungkim1/LM-kor)
```python
from transformers import BertTokenizerFast, BertModel
tokenizer_bert = BertTokenizerFast.from_pretrained("kykim/bert-kor-base")
model_bert = BertModel.from_pretrained("kykim/bert-kor-base")
``` |
facebook/wmt19-ru-en | d145f3063a3e25e43c04c9aa64de38999b3fb2cd | 2020-12-11T21:40:01.000Z | [
"pytorch",
"fsmt",
"text2text-generation",
"ru",
"en",
"dataset:wmt19",
"arxiv:1907.06616",
"transformers",
"translation",
"wmt19",
"facebook",
"license:apache-2.0",
"autotrain_compatible"
] | translation | false | facebook | null | facebook/wmt19-ru-en | 14,556 | 3 | transformers | 576 | ---
language:
- ru
- en
tags:
- translation
- wmt19
- facebook
license: apache-2.0
datasets:
- wmt19
metrics:
- bleu
thumbnail: https://huggingface.co/front/thumbnails/facebook.png
---
# FSMT
## Model description
This is a ported version of [fairseq wmt19 transformer](https://github.com/pytorch/fairseq/blob/master/examples/wmt19/README.md) for ru-en.
For more details, please see, [Facebook FAIR's WMT19 News Translation Task Submission](https://arxiv.org/abs/1907.06616).
The abbreviation FSMT stands for FairSeqMachineTranslation
All four models are available:
* [wmt19-en-ru](https://huggingface.co/facebook/wmt19-en-ru)
* [wmt19-ru-en](https://huggingface.co/facebook/wmt19-ru-en)
* [wmt19-en-de](https://huggingface.co/facebook/wmt19-en-de)
* [wmt19-de-en](https://huggingface.co/facebook/wmt19-de-en)
## Intended uses & limitations
#### How to use
```python
from transformers import FSMTForConditionalGeneration, FSMTTokenizer
mname = "facebook/wmt19-ru-en"
tokenizer = FSMTTokenizer.from_pretrained(mname)
model = FSMTForConditionalGeneration.from_pretrained(mname)
input = "Машинное обучение - это здорово, не так ли?"
input_ids = tokenizer.encode(input, return_tensors="pt")
outputs = model.generate(input_ids)
decoded = tokenizer.decode(outputs[0], skip_special_tokens=True)
print(decoded) # Machine learning is great, isn't it?
```
#### Limitations and bias
- The original (and this ported model) doesn't seem to handle well inputs with repeated sub-phrases, [content gets truncated](https://discuss.huggingface.co/t/issues-with-translating-inputs-containing-repeated-phrases/981)
## Training data
Pretrained weights were left identical to the original model released by fairseq. For more details, please, see the [paper](https://arxiv.org/abs/1907.06616).
## Eval results
pair | fairseq | transformers
-------|---------|----------
ru-en | [41.3](http://matrix.statmt.org/matrix/output/1907?run_id=6937) | 39.20
The score is slightly below the score reported by `fairseq`, since `transformers`` currently doesn't support:
- model ensemble, therefore the best performing checkpoint was ported (``model4.pt``).
- re-ranking
The score was calculated using this code:
```bash
git clone https://github.com/huggingface/transformers
cd transformers
export PAIR=ru-en
export DATA_DIR=data/$PAIR
export SAVE_DIR=data/$PAIR
export BS=8
export NUM_BEAMS=15
mkdir -p $DATA_DIR
sacrebleu -t wmt19 -l $PAIR --echo src > $DATA_DIR/val.source
sacrebleu -t wmt19 -l $PAIR --echo ref > $DATA_DIR/val.target
echo $PAIR
PYTHONPATH="src:examples/seq2seq" python examples/seq2seq/run_eval.py facebook/wmt19-$PAIR $DATA_DIR/val.source $SAVE_DIR/test_translations.txt --reference_path $DATA_DIR/val.target --score_path $SAVE_DIR/test_bleu.json --bs $BS --task translation --num_beams $NUM_BEAMS
```
note: fairseq reports using a beam of 50, so you should get a slightly higher score if re-run with `--num_beams 50`.
## Data Sources
- [training, etc.](http://www.statmt.org/wmt19/)
- [test set](http://matrix.statmt.org/test_sets/newstest2019.tgz?1556572561)
### BibTeX entry and citation info
```bibtex
@inproceedings{...,
year={2020},
title={Facebook FAIR's WMT19 News Translation Task Submission},
author={Ng, Nathan and Yee, Kyra and Baevski, Alexei and Ott, Myle and Auli, Michael and Edunov, Sergey},
booktitle={Proc. of WMT},
}
```
## TODO
- port model ensemble (fairseq uses 4 model checkpoints)
|
microsoft/dit-base-finetuned-rvlcdip | 48a4c1ad3e7ea1dfb87e8745eba7372162cba7f5 | 2022-07-21T12:20:05.000Z | [
"pytorch",
"beit",
"image-classification",
"dataset:rvl_cdip",
"arxiv:2203.02378",
"transformers",
"dit",
"vision"
] | image-classification | false | microsoft | null | microsoft/dit-base-finetuned-rvlcdip | 14,504 | 7 | transformers | 577 | ---
tags:
- dit
- vision
- image-classification
datasets:
- rvl_cdip
widget:
- src: https://huggingface.co/microsoft/dit-base-finetuned-rvlcdip/resolve/main/coca_cola_advertisement.png
example_title: Advertisement
- src: https://huggingface.co/microsoft/dit-base-finetuned-rvlcdip/resolve/main/scientific_publication.png
example_title: Scientific publication
---
# Document Image Transformer (base-sized model)
Document Image Transformer (DiT) model pre-trained on IIT-CDIP (Lewis et al., 2006), a dataset that includes 42 million document images and fine-tuned on [RVL-CDIP](https://www.cs.cmu.edu/~aharley/rvl-cdip/), a dataset consisting of 400,000 grayscale images in 16 classes, with 25,000 images per class. It was introduced in the paper [DiT: Self-supervised Pre-training for Document Image Transformer](https://arxiv.org/abs/2203.02378) by Li et al. and first released in [this repository](https://github.com/microsoft/unilm/tree/master/dit). Note that DiT is identical to the architecture of [BEiT](https://huggingface.co/docs/transformers/model_doc/beit).
Disclaimer: The team releasing DiT did not write a model card for this model so this model card has been written by the Hugging Face team.
## Model description
The Document Image Transformer (DiT) is a transformer encoder model (BERT-like) pre-trained on a large collection of images in a self-supervised fashion. The pre-training objective for the model is to predict visual tokens from the encoder of a discrete VAE (dVAE), based on masked patches.
Images are presented to the model as a sequence of fixed-size patches (resolution 16x16), which are linearly embedded. One also adds absolute position embeddings before feeding the sequence to the layers of the Transformer encoder.
By pre-training the model, it learns an inner representation of images that can then be used to extract features useful for downstream tasks: if you have a dataset of labeled document images for instance, you can train a standard classifier by placing a linear layer on top of the pre-trained encoder.
## Intended uses & limitations
You can use the raw model for encoding document images into a vector space, but it's mostly meant to be fine-tuned on tasks like document image classification, table detection or document layout analysis. See the [model hub](https://huggingface.co/models?search=microsoft/dit) to look for fine-tuned versions on a task that interests you.
### How to use
Here is how to use this model in PyTorch:
```python
from transformers import AutoFeatureExtractor, AutoModelForImageClassification
import torch
from PIL import Image
image = Image.open('path_to_your_document_image').convert('RGB')
feature_extractor = AutoFeatureExtractor.from_pretrained("microsoft/dit-base-finetuned-rvlcdip")
model = AutoModelForImageClassification.from_pretrained("microsoft/dit-base-finetuned-rvlcdip")
inputs = feature_extractor(images=image, return_tensors="pt")
outputs = model(**inputs)
logits = outputs.logits
# model predicts one of the 16 RVL-CDIP classes
predicted_class_idx = logits.argmax(-1).item()
print("Predicted class:", model.config.id2label[predicted_class_idx])
```
### BibTeX entry and citation info
```bibtex
@article{Lewis2006BuildingAT,
title={Building a test collection for complex document information processing},
author={David D. Lewis and Gady Agam and Shlomo Engelson Argamon and Ophir Frieder and David A. Grossman and Jefferson Heard},
journal={Proceedings of the 29th annual international ACM SIGIR conference on Research and development in information retrieval},
year={2006}
}
``` |
deepset/tinyroberta-squad2 | 9f008c54f533ffdf142f127c0c5c9bfe23542aaa | 2022-07-25T11:44:05.000Z | [
"pytorch",
"roberta",
"question-answering",
"en",
"dataset:squad_v2",
"arxiv:1909.10351",
"transformers",
"license:cc-by-4.0",
"model-index",
"autotrain_compatible"
] | question-answering | false | deepset | null | deepset/tinyroberta-squad2 | 14,451 | 8 | transformers | 578 | ---
language: en
datasets:
- squad_v2
license: cc-by-4.0
model-index:
- name: deepset/tinyroberta-squad2
results:
- task:
type: question-answering
name: Question Answering
dataset:
name: squad_v2
type: squad_v2
config: squad_v2
split: validation
metrics:
- name: Exact Match
type: exact_match
value: 78.8627
verified: true
- name: F1
type: f1
value: 82.0355
verified: true
---
# tinyroberta-squad2
This is the *distilled* version of the [deepset/roberta-base-squad2](https://huggingface.co/deepset/roberta-base-squad2) model. This model has a comparable prediction quality and runs at twice the speed of the base model.
## Overview
**Language model:** tinyroberta-squad2
**Language:** English
**Downstream-task:** Extractive QA
**Training data:** SQuAD 2.0
**Eval data:** SQuAD 2.0
**Code:**
**Infrastructure**: 4x Tesla v100
## Hyperparameters
```
batch_size = 96
n_epochs = 4
base_LM_model = "deepset/tinyroberta-squad2-step1"
max_seq_len = 384
learning_rate = 3e-5
lr_schedule = LinearWarmup
warmup_proportion = 0.2
doc_stride = 128
max_query_length = 64
distillation_loss_weight = 0.75
temperature = 1.5
teacher = "deepset/robert-large-squad2"
```
## Distillation
This model was distilled using the TinyBERT approach described in [this paper](https://arxiv.org/pdf/1909.10351.pdf) and implemented in [haystack](https://github.com/deepset-ai/haystack).
Firstly, we have performed intermediate layer distillation with roberta-base as the teacher which resulted in [deepset/tinyroberta-6l-768d](https://huggingface.co/deepset/tinyroberta-6l-768d).
Secondly, we have performed task-specific distillation with [deepset/roberta-base-squad2](https://huggingface.co/deepset/roberta-base-squad2) as the teacher for further intermediate layer distillation on an augmented version of SQuADv2 and then with [deepset/roberta-large-squad2](https://huggingface.co/deepset/roberta-large-squad2) as the teacher for prediction layer distillation.
## Usage
### In Haystack
Haystack is an NLP framework by deepset. You can use this model in a Haystack pipeline to do question answering at scale (over many documents). To load the model in [Haystack](https://github.com/deepset-ai/haystack/):
```python
reader = FARMReader(model_name_or_path="deepset/tinyroberta-squad2")
# or
reader = TransformersReader(model_name_or_path="deepset/tinyroberta-squad2")
```
### In Transformers
```python
from transformers import AutoModelForQuestionAnswering, AutoTokenizer, pipeline
model_name = "deepset/tinyroberta-squad2"
# a) Get predictions
nlp = pipeline('question-answering', model=model_name, tokenizer=model_name)
QA_input = {
'question': 'Why is model conversion important?',
'context': 'The option to convert models between FARM and transformers gives freedom to the user and let people easily switch between frameworks.'
}
res = nlp(QA_input)
# b) Load model & tokenizer
model = AutoModelForQuestionAnswering.from_pretrained(model_name)
tokenizer = AutoTokenizer.from_pretrained(model_name)
```
## Performance
Evaluated on the SQuAD 2.0 dev set with the [official eval script](https://worksheets.codalab.org/rest/bundles/0x6b567e1cf2e041ec80d7098f031c5c9e/contents/blob/).
```
"exact": 78.69114798281817,
"f1": 81.9198998536977,
"total": 11873,
"HasAns_exact": 76.19770580296895,
"HasAns_f1": 82.66446878592329,
"HasAns_total": 5928,
"NoAns_exact": 81.17746005046257,
"NoAns_f1": 81.17746005046257,
"NoAns_total": 5945
```
## Authors
**Branden Chan:** [email protected]
**Timo Möller:** [email protected]
**Malte Pietsch:** [email protected]
**Tanay Soni:** [email protected]
**Michel Bartels:** [email protected]
## About us
<div class="grid lg:grid-cols-2 gap-x-4 gap-y-3">
<div class="w-full h-40 object-cover mb-2 rounded-lg flex items-center justify-center">
<img alt="" src="https://huggingface.co/spaces/deepset/README/resolve/main/haystack-logo-colored.svg" class="w-40"/>
</div>
<div class="w-full h-40 object-cover mb-2 rounded-lg flex items-center justify-center">
<img alt="" src="https://huggingface.co/spaces/deepset/README/resolve/main/deepset-logo-colored.svg" class="w-40"/>
</div>
</div>
[deepset](http://deepset.ai/) is the company behind the open-source NLP framework [Haystack](https://haystack.deepset.ai/) which is designed to help you build production ready NLP systems that use: Question answering, summarization, ranking etc.
Some of our other work:
- [roberta-base-squad2]([https://huggingface.co/deepset/roberta-base-squad2)
- [German BERT (aka "bert-base-german-cased")](https://deepset.ai/german-bert)
- [GermanQuAD and GermanDPR datasets and models (aka "gelectra-base-germanquad", "gbert-base-germandpr")](https://deepset.ai/germanquad)
## Get in touch and join the Haystack community
<p>For more info on Haystack, visit our <strong><a href="https://github.com/deepset-ai/haystack">GitHub</a></strong> repo and <strong><a href="https://haystack.deepset.ai">Documentation</a></strong>.
We also have a <strong><a class="h-7" href="https://haystack.deepset.ai/community/join"><img alt="slack" class="h-7 inline-block m-0" style="margin: 0" src="https://huggingface.co/spaces/deepset/README/resolve/main/Slack_RGB.png"/>community open to everyone!</a></strong></p>
[Twitter](https://twitter.com/deepset_ai) | [LinkedIn](https://www.linkedin.com/company/deepset-ai/) | [Slack](https://haystack.deepset.ai/community/join) | [GitHub Discussions](https://github.com/deepset-ai/haystack/discussions) | [Website](https://deepset.ai)
By the way: [we're hiring!](http://www.deepset.ai/jobs) |
Recognai/bert-base-spanish-wwm-cased-xnli | 6219d7e4cc59999010c795ac26d2e014102e24ba | 2021-10-15T15:55:15.000Z | [
"pytorch",
"jax",
"bert",
"text-classification",
"es",
"dataset:xnli",
"transformers",
"zero-shot-classification",
"nli",
"license:mit"
] | zero-shot-classification | false | Recognai | null | Recognai/bert-base-spanish-wwm-cased-xnli | 14,415 | 10 | transformers | 579 | ---
language: es
tags:
- zero-shot-classification
- nli
- pytorch
datasets:
- xnli
license: mit
pipeline_tag: zero-shot-classification
widget:
- text: "El autor se perfila, a los 50 años de su muerte, como uno de los grandes de su siglo"
candidate_labels: "cultura, sociedad, economia, salud, deportes"
---
# bert-base-spanish-wwm-cased-xnli
**UPDATE, 15.10.2021: Check out our new zero-shot classifiers, much more lightweight and even outperforming this one: [zero-shot SELECTRA small](https://huggingface.co/Recognai/zeroshot_selectra_small) and [zero-shot SELECTRA medium](https://huggingface.co/Recognai/zeroshot_selectra_medium).**
## Model description
This model is a fine-tuned version of the [spanish BERT model](https://huggingface.co/dccuchile/bert-base-spanish-wwm-cased) with the Spanish portion of the XNLI dataset. You can have a look at the [training script](https://huggingface.co/Recognai/bert-base-spanish-wwm-cased-xnli/blob/main/zeroshot_training_script.py) for details of the training.
### How to use
You can use this model with Hugging Face's [zero-shot-classification pipeline](https://discuss.huggingface.co/t/new-pipeline-for-zero-shot-text-classification/681):
```python
from transformers import pipeline
classifier = pipeline("zero-shot-classification",
model="Recognai/bert-base-spanish-wwm-cased-xnli")
classifier(
"El autor se perfila, a los 50 años de su muerte, como uno de los grandes de su siglo",
candidate_labels=["cultura", "sociedad", "economia", "salud", "deportes"],
hypothesis_template="Este ejemplo es {}."
)
"""output
{'sequence': 'El autor se perfila, a los 50 años de su muerte, como uno de los grandes de su siglo',
'labels': ['cultura', 'sociedad', 'economia', 'salud', 'deportes'],
'scores': [0.38897448778152466,
0.22997373342514038,
0.1658431738615036,
0.1205764189362526,
0.09463217109441757]}
"""
```
## Eval results
Accuracy for the test set:
| | XNLI-es |
|-----------------------------|---------|
|bert-base-spanish-wwm-cased-xnli | 79.9% | |
cross-encoder/qnli-electra-base | 4d70c22ec2d12ec7663a70fbe3180a408c980a2a | 2021-08-05T08:41:23.000Z | [
"pytorch",
"electra",
"text-classification",
"arxiv:1804.07461",
"transformers",
"license:apache-2.0"
] | text-classification | false | cross-encoder | null | cross-encoder/qnli-electra-base | 14,334 | null | transformers | 580 | ---
license: apache-2.0
---
# Cross-Encoder for Quora Duplicate Questions Detection
This model was trained using [SentenceTransformers](https://sbert.net) [Cross-Encoder](https://www.sbert.net/examples/applications/cross-encoder/README.html) class.
## Training Data
Given a question and paragraph, can the question be answered by the paragraph? The models have been trained on the [GLUE QNLI](https://arxiv.org/abs/1804.07461) dataset, which transformed the [SQuAD dataset](https://rajpurkar.github.io/SQuAD-explorer/) into an NLI task.
## Performance
For performance results of this model, see [SBERT.net Pre-trained Cross-Encoder][https://www.sbert.net/docs/pretrained_cross-encoders.html].
## Usage
Pre-trained models can be used like this:
```python
from sentence_transformers import CrossEncoder
model = CrossEncoder('model_name')
scores = model.predict([('Query1', 'Paragraph1'), ('Query2', 'Paragraph2')])
#e.g.
scores = model.predict([('How many people live in Berlin?', 'Berlin had a population of 3,520,031 registered inhabitants in an area of 891.82 square kilometers.'), ('What is the size of New York?', 'New York City is famous for the Metropolitan Museum of Art.')])
```
## Usage with Transformers AutoModel
You can use the model also directly with Transformers library (without SentenceTransformers library):
```python
from transformers import AutoTokenizer, AutoModelForSequenceClassification
import torch
model = AutoModelForSequenceClassification.from_pretrained('model_name')
tokenizer = AutoTokenizer.from_pretrained('model_name')
features = tokenizer(['How many people live in Berlin?', 'What is the size of New York?'], ['Berlin had a population of 3,520,031 registered inhabitants in an area of 891.82 square kilometers.', 'New York City is famous for the Metropolitan Museum of Art.'], padding=True, truncation=True, return_tensors="pt")
model.eval()
with torch.no_grad():
scores = torch.nn.functional.sigmoid(model(**features).logits)
print(scores)
``` |
klue/roberta-large | 5193b95701189160c45d02a1033a4ea55bdbe259 | 2021-10-20T16:13:45.000Z | [
"pytorch",
"roberta",
"fill-mask",
"ko",
"arxiv:2105.09680",
"transformers",
"korean",
"klue",
"autotrain_compatible"
] | fill-mask | false | klue | null | klue/roberta-large | 14,205 | 14 | transformers | 581 | ---
language: ko
tags:
- korean
- klue
mask_token: "[MASK]"
widget:
- text: 대한민국의 수도는 [MASK] 입니다.
---
# KLUE RoBERTa large
Pretrained RoBERTa Model on Korean Language. See [Github](https://github.com/KLUE-benchmark/KLUE) and [Paper](https://arxiv.org/abs/2105.09680) for more details.
## How to use
_NOTE:_ Use `BertTokenizer` instead of RobertaTokenizer. (`AutoTokenizer` will load `BertTokenizer`)
```python
from transformers import AutoModel, AutoTokenizer
model = AutoModel.from_pretrained("klue/roberta-large")
tokenizer = AutoTokenizer.from_pretrained("klue/roberta-large")
```
## BibTeX entry and citation info
```bibtex
@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}
}
```
|
google/reformer-enwik8 | ddc3376ce78ee3917253891e59ecf581247d8a8c | 2021-06-03T13:02:12.000Z | [
"pytorch",
"reformer",
"text-generation",
"transformers"
] | text-generation | false | google | null | google/reformer-enwik8 | 14,199 | 2 | transformers | 582 | ## Reformer Language model on character level and trained on enwik8.
*enwik8* is a dataset based on Wikipedia and is often used to measure the model's ability to *compress* data, *e.g.* in
the scope of the *Hutter prize*: https://en.wikipedia.org/wiki/Hutter_Prize.
`reformer-enwik8` was pretrained on the first 90M chars of *enwik8* whereas the text was chunked into batches of size 65536 chars (=2^16).
The model's weights were taken from https://console.cloud.google.com/storage/browser/trax-ml/reformer/enwik8 and converted
to Hugging Face's PyTorch ReformerLM model `ReformerModelWithLMHead`.
The model is a language model that operates on characters.
Therefore, this model does not need a tokenizer. The following function can instead be used for **encoding** and **decoding**:
```python
import torch
# Encoding
def encode(list_of_strings, pad_token_id=0):
max_length = max([len(string) for string in list_of_strings])
# create emtpy tensors
attention_masks = torch.zeros((len(list_of_strings), max_length), dtype=torch.long)
input_ids = torch.full((len(list_of_strings), max_length), pad_token_id, dtype=torch.long)
for idx, string in enumerate(list_of_strings):
# make sure string is in byte format
if not isinstance(string, bytes):
string = str.encode(string)
input_ids[idx, :len(string)] = torch.tensor([x + 2 for x in string])
attention_masks[idx, :len(string)] = 1
return input_ids, attention_masks
# Decoding
def decode(outputs_ids):
decoded_outputs = []
for output_ids in outputs_ids.tolist():
# transform id back to char IDs < 2 are simply transformed to ""
decoded_outputs.append("".join([chr(x - 2) if x > 1 else "" for x in output_ids]))
return decoded_outputs
```
Text can be generated as follows:
```python
from transformers import ReformerModelWithLMHead
model = ReformerModelWithLMHead.from_pretrained("google/reformer-enwik8")
encoded, attention_masks = encode(["In 1965, Brooks left IBM to found the Department of"])
decode(model.generate(encoded, do_sample=True, max_length=150))
# gives:
# In 1965, Brooks left IBM to found the Department of Journalism in 1968. IBM had jurisdiction himself in 1980, while Brooks resolved, nevertheless thro
```
***Note***: Language generation using `ReformerModelWithLMHead` is not optimized yet and is rather slow.
|
uclanlp/plbart-base | cf5287241fcff3819f6ade49635dc2d77efee032 | 2021-11-09T17:07:52.000Z | [
"pytorch",
"plbart",
"text2text-generation",
"transformers",
"autotrain_compatible"
] | text2text-generation | false | uclanlp | null | uclanlp/plbart-base | 14,197 | 3 | transformers | 583 | Entry not found |
cross-encoder/nli-deberta-v3-base | 00259490157750c3fb567a3570e0fcc827e0cee0 | 2021-12-27T22:26:49.000Z | [
"pytorch",
"deberta-v2",
"text-classification",
"en",
"dataset:multi_nli",
"dataset:snli",
"transformers",
"microsoft/deberta-v3-base",
"license:apache-2.0",
"zero-shot-classification"
] | zero-shot-classification | false | cross-encoder | null | cross-encoder/nli-deberta-v3-base | 14,193 | 3 | transformers | 584 | ---
language: en
pipeline_tag: zero-shot-classification
tags:
- microsoft/deberta-v3-base
datasets:
- multi_nli
- snli
metrics:
- accuracy
license: apache-2.0
---
# Cross-Encoder for Natural Language Inference
This model was trained using [SentenceTransformers](https://sbert.net) [Cross-Encoder](https://www.sbert.net/examples/applications/cross-encoder/README.html) class. This model is based on [microsoft/deberta-v3-base](https://huggingface.co/microsoft/deberta-v3-base)
## Training Data
The model was trained on the [SNLI](https://nlp.stanford.edu/projects/snli/) and [MultiNLI](https://cims.nyu.edu/~sbowman/multinli/) datasets. For a given sentence pair, it will output three scores corresponding to the labels: contradiction, entailment, neutral.
## Performance
- Accuracy on SNLI-test dataset: 92.38
- Accuracy on MNLI mismatched set: 90.04
For futher evaluation results, see [SBERT.net - Pretrained Cross-Encoder](https://www.sbert.net/docs/pretrained_cross-encoders.html#nli).
## Usage
Pre-trained models can be used like this:
```python
from sentence_transformers import CrossEncoder
model = CrossEncoder('cross-encoder/nli-deberta-v3-base')
scores = model.predict([('A man is eating pizza', 'A man eats something'), ('A black race car starts up in front of a crowd of people.', 'A man is driving down a lonely road.')])
#Convert scores to labels
label_mapping = ['contradiction', 'entailment', 'neutral']
labels = [label_mapping[score_max] for score_max in scores.argmax(axis=1)]
```
## Usage with Transformers AutoModel
You can use the model also directly with Transformers library (without SentenceTransformers library):
```python
from transformers import AutoTokenizer, AutoModelForSequenceClassification
import torch
model = AutoModelForSequenceClassification.from_pretrained('cross-encoder/nli-deberta-v3-base')
tokenizer = AutoTokenizer.from_pretrained('cross-encoder/nli-deberta-v3-base')
features = tokenizer(['A man is eating pizza', 'A black race car starts up in front of a crowd of people.'], ['A man eats something', 'A man is driving down a lonely road.'], padding=True, truncation=True, return_tensors="pt")
model.eval()
with torch.no_grad():
scores = model(**features).logits
label_mapping = ['contradiction', 'entailment', 'neutral']
labels = [label_mapping[score_max] for score_max in scores.argmax(dim=1)]
print(labels)
```
## Zero-Shot Classification
This model can also be used for zero-shot-classification:
```python
from transformers import pipeline
classifier = pipeline("zero-shot-classification", model='cross-encoder/nli-deberta-v3-base')
sent = "Apple just announced the newest iPhone X"
candidate_labels = ["technology", "sports", "politics"]
res = classifier(sent, candidate_labels)
print(res)
``` |
allenai/PRIMERA | 550385ac1f15be0309fddaa429d72a87ed60aa6b | 2022-06-25T16:04:26.000Z | [
"pytorch",
"tf",
"led",
"text2text-generation",
"transformers",
"license:apache-2.0",
"autotrain_compatible"
] | text2text-generation | false | allenai | null | allenai/PRIMERA | 14,189 | 4 | transformers | 585 | ---
license: apache-2.0
---
HF-version model for PRIMERA: Pyramid-based Masked Sentence Pre-training for Multi-document Summarization (ACL 2022).
The original code can be found [here](https://github.com/allenai/PRIMER). You can find the script and notebook to train/evaluate the model in the original github repo.
* Note: due to the difference between the implementations of the original Longformer and the Huggingface LED model, the results of converted models are slightly different. We run a sanity check on both fine-tuned and non fine-tuned models on the **MultiNews dataset**, and show the results below:
| Model | Rouge-1 | Rouge-2 | Rouge-L |
| --- | ----------- |----------- |----------- |
| PRIMERA | 42.0 | 13.6 | 20.8|
| PRIMERA-hf | 41.7 |13.6 | 20.5|
| PRIMERA(finetuned) | 49.9 | 21.1 | 25.9|
| PRIMERA-hf(finetuned) | 49.9 | 20.9 | 25.8|
You can use it by
```
from transformers import (
AutoTokenizer,
LEDConfig,
LEDForConditionalGeneration,
)
tokenizer = AutoTokenizer.from_pretrained('allenai/PRIMERA')
config=LEDConfig.from_pretrained('allenai/PRIMERA')
model = LEDForConditionalGeneration.from_pretrained('allenai/PRIMERA')
```
|
princeton-nlp/sup-simcse-roberta-base | 4bf73c6b5df517f74188c5e9ec159b2208c89c08 | 2021-05-20T19:33:45.000Z | [
"pytorch",
"jax",
"roberta",
"feature-extraction",
"transformers"
] | feature-extraction | false | princeton-nlp | null | princeton-nlp/sup-simcse-roberta-base | 14,162 | 1 | transformers | 586 | Entry not found |
vblagoje/bart_lfqa | 5493d5be6812cdb4835e004ce17ea2082cc25b03 | 2022-02-14T15:54:47.000Z | [
"pytorch",
"bart",
"text2text-generation",
"en",
"dataset:vblagoje/lfqa",
"dataset:vblagoje/lfqa_support_docs",
"transformers",
"license:mit",
"autotrain_compatible"
] | text2text-generation | false | vblagoje | null | vblagoje/bart_lfqa | 14,151 | 12 | transformers | 587 | ---
language: en
datasets:
- vblagoje/lfqa
- vblagoje/lfqa_support_docs
license: mit
---
## Introduction
See [blog post](https://towardsdatascience.com/long-form-qa-beyond-eli5-an-updated-dataset-and-approach-319cb841aabb) for more details.
## Usage
```python
import torch
from transformers import AutoTokenizer, AutoModel, AutoModelForSeq2SeqLM
model_name = "vblagoje/bart_lfqa"
device = torch.device('cuda' if torch.cuda.is_available() else 'cpu')
tokenizer = AutoTokenizer.from_pretrained(model_name)
model = AutoModelForSeq2SeqLM.from_pretrained(model_name)
model = model.to(device)
# it all starts with a question/query
query = "Why does water heated to room temperature feel colder than the air around it?"
# given the question above suppose these documents below were found in some document store
documents = ["when the skin is completely wet. The body continuously loses water by...",
"at greater pressures. There is an ambiguity, however, as to the meaning of the terms 'heating' and 'cooling'...",
"are not in a relation of thermal equilibrium, heat will flow from the hotter to the colder, by whatever pathway...",
"air condition and moving along a line of constant enthalpy toward a state of higher humidity. A simple example ...",
"Thermal contact conductance In physics, thermal contact conductance is the study of heat conduction between solid ..."]
# concatenate question and support documents into BART input
conditioned_doc = "<P> " + " <P> ".join([d for d in documents])
query_and_docs = "question: {} context: {}".format(query, conditioned_doc)
model_input = tokenizer(query_and_docs, truncation=True, padding=True, return_tensors="pt")
generated_answers_encoded = model.generate(input_ids=model_input["input_ids"].to(device),
attention_mask=model_input["attention_mask"].to(device),
min_length=64,
max_length=256,
do_sample=False,
early_stopping=True,
num_beams=8,
temperature=1.0,
top_k=None,
top_p=None,
eos_token_id=tokenizer.eos_token_id,
no_repeat_ngram_size=3,
num_return_sequences=1)
tokenizer.batch_decode(generated_answers_encoded, skip_special_tokens=True,clean_up_tokenization_spaces=True)
# below is the abstractive answer generated by the model
["When you heat water to room temperature, it loses heat to the air around it. When you cool it down, it gains heat back from the air, which is why it feels colder than the air surrounding it. It's the same reason why you feel cold when you turn on a fan. The air around you is losing heat, and the water is gaining heat."]
```
## Author
- Vladimir Blagojevic: `dovlex [at] gmail.com` [Twitter](https://twitter.com/vladblagoje) | [LinkedIn](https://www.linkedin.com/in/blagojevicvladimir/) |
patrickvonplaten/led-large-16384-pubmed | 10e2f1e6bdb4fa833e899fec2be31b87751135ce | 2021-01-11T15:42:53.000Z | [
"pytorch",
"tf",
"led",
"text2text-generation",
"en",
"dataset:scientific_papers",
"transformers",
"license:apache-2.0",
"autotrain_compatible"
] | text2text-generation | false | patrickvonplaten | null | patrickvonplaten/led-large-16384-pubmed | 14,142 | 5 | transformers | 588 | ---
language: en
datasets:
- scientific_papers
license: apache-2.0
---
## Introduction
[Allenai's Longformer Encoder-Decoder (LED)](https://github.com/allenai/longformer#longformer).
This is an unofficial *led-large-16384* checkpoint that is fine-tuned on the [pubmed dataset](https://huggingface.co/datasets/scientific_papers).
The model was fine-tuned and evaluated as detailed in [this notebook](https://colab.research.google.com/drive/12LjJazBl7Gam0XBPy_y0CTOJZeZ34c2v?usp=sharing)
## Results
The model achieves a **Rouge-2** score of 19.33 on Pubmed which is competitive to state-of-the-art models.
## Usage
The model can be used as follows. The input is taken from the test data of the [pubmed dataset](https://huggingface.co/datasets/scientific_papers).
```python
LONG_ARTICLE = """"anxiety affects quality of life in those living
with parkinson 's disease ( pd ) more so than
overall cognitive status , motor deficits , apathy
, and depression [ 13 ] . although anxiety and
depression are often related and coexist in pd
patients , recent research suggests that anxiety
rather than depression is the most prominent and
prevalent mood disorder in pd [ 5 , 6 ] . yet ,
our current understanding of anxiety and its
impact on cognition in pd , as well as its neural
basis and best treatment practices , remains
meager and lags far behind that of depression .
overall , neuropsychiatric symptoms in pd have
been shown to be negatively associated with
cognitive performance . for example , higher
depression scores have been correlated with lower
scores on the mini - mental state exam ( mmse ) [
8 , 9 ] as well as tests of memory and executive
functions ( e.g. , attention ) [ 1014 ] . likewise
, apathy and anhedonia in pd patients have been
associated with executive dysfunction [ 10 , 1523
] . however , few studies have specifically
investigated the relationship between anxiety and
cognition in pd . one study showed a strong
negative relationship between anxiety ( both state
and trait ) and overall cognitive performance (
measured by the total of the repeatable battery
for the assessment of neuropsychological status
index ) within a sample of 27 pd patients .
furthermore , trait anxiety was negatively
associated with each of the cognitive domains
assessed by the rbans ( i.e. , immediate memory ,
visuospatial construction , language , attention ,
and delayed memory ) . two further studies have
examined whether anxiety differentially affects
cognition in patients with left - sided dominant
pd ( lpd ) versus right - sided dominant pd ( rpd
) ; however , their findings were inconsistent .
the first study found that working memory
performance was worse in lpd patients with anxiety
compared to rpd patients with anxiety , whereas
the second study reported that , in lpd , apathy
but not anxiety was associated with performance on
nonverbally mediated executive functions and
visuospatial tasks ( e.g. , tmt - b , wms - iii
spatial span ) , while in rpd , anxiety but not
apathy significantly correlated with performance
on verbally mediated tasks ( e.g. , clock reading
test and boston naming test ) . furthermore ,
anxiety was significantly correlated with
neuropsychological measures of attention and
executive and visuospatial functions . taken
together , it is evident that there are limited
and inconsistent findings describing the
relationship between anxiety and cognition in pd
and more specifically how anxiety might influence
particular domains of cognition such as attention
and memory and executive functioning . it is also
striking that , to date , no study has examined
the influence of anxiety on cognition in pd by
directly comparing groups of pd patients with and
without anxiety while excluding depression . given
that research on healthy young adults suggests
that anxiety reduces processing capacity and
impairs processing efficiency , especially in the
central executive and attentional systems of
working memory [ 26 , 27 ] , we hypothesized that
pd patients with anxiety would show impairments in
attentional set - shifting and working memory
compared to pd patients without anxiety .
furthermore , since previous work , albeit limited
, has focused on the influence of symptom
laterality on anxiety and cognition , we also
explored this relationship . seventeen pd patients
with anxiety and thirty - three pd patients
without anxiety were included in this study ( see
table 1 ) . the cross - sectional data from these
participants was taken from a patient database
that has been compiled over the past 8 years (
since 2008 ) at the parkinson 's disease research
clinic at the brain and mind centre , university
of sydney . inclusion criteria involved a
diagnosis of idiopathic pd according to the united
kingdom parkinson 's disease society brain bank
criteria and were confirmed by a neurologist (
sjgl ) . patients also had to have an adequate
proficiency in english and have completed a full
neuropsychological assessment . ten patients in
this study ( 5 pd with anxiety ; 5 pd without
anxiety ) were taking psychotropic drugs ( i.e. ,
benzodiazepine or selective serotonin reuptake
inhibitor ) . patients were also excluded if they
had other neurological disorders , psychiatric
disorders other than affective disorders ( such as
anxiety ) , or if they reported a score greater
than six on the depression subscale of the
hospital anxiety and depression scale ( hads ) .
thus , all participants who scored within a
depressed ( hads - d > 6 ) range were excluded
from this study , in attempt to examine a refined
sample of pd patients with and without anxiety in
order to determine the independent effect of
anxiety on cognition . this research was approved
by the human research ethics committee of the
university of sydney , and written informed
consent was obtained from all participants . self
- reported hads was used to assess anxiety in pd
and has been previously shown to be a useful
measure of clinical anxiety in pd . a cut - off
score of > 8 on the anxiety subscale of the hads (
hads - a ) was used to identify pd cases with
anxiety ( pda+ ) , while a cut - off score of < 6
on the hads - a was used to identify pd cases
without anxiety ( pda ) . this criterion was more
stringent than usual ( > 7 cut - off score ) , in
effort to create distinct patient groups . the
neurological evaluation rated participants
according to hoehn and yahr ( h&y ) stages and
assessed their motor symptoms using part iii of
the revised mds task force unified parkinson 's
disease rating scale ( updrs ) . in a similar way
this was determined by calculating a total left
and right score from rigidity items 3035 ,
voluntary movement items 3643 , and tremor items
5057 from the mds - updrs part iii ( see table 1 )
. processing speed was assessed using the trail
making test , part a ( tmt - a , z - score ) .
attentional set - shifting was measured using the
trail making test , part b ( tmt - b , z - score )
. working memory was assessed using the digit span
forward and backward subtest of the wechsler
memory scale - iii ( raw scores ) . language was
assessed with semantic and phonemic verbal fluency
via the controlled oral word associated test (
cowat animals and letters , z - score ) . the
ability to retain learned verbal memory was
assessed using the logical memory subtest from the
wechsler memory scale - iii ( lm - i z - score ,
lm - ii z - score , % lm retention z - score ) .
the mini - mental state examination ( mmse )
demographic , clinical , and neuropsychological
variables were compared between the two groups
with the independent t - test or mann whitney u
test , depending on whether the variable met
parametric assumptions . chi - square tests were
used to examine gender and symptom laterality
differences between groups . all analyses employed
an alpha level of p < 0.05 and were two - tailed .
spearman correlations were performed separately in
each group to examine associations between anxiety
and/or depression ratings and cognitive functions
. as expected , the pda+ group reported
significant greater levels of anxiety on the hads
- a ( u = 0 , p < 0.001 ) and higher total score
on the hads ( u = 1 , p < 0.001 ) compared to the
pda group ( table 1 ) . groups were matched in age
( t(48 ) = 1.31 , p = 0.20 ) , disease duration (
u = 259 , p = 0.66 ) , updrs - iii score ( u =
250.5 , p = 0.65 ) , h&y ( u = 245 , p = 0.43 ) ,
ledd ( u = 159.5 , p = 0.80 ) , and depression (
hads - d ) ( u = 190.5 , p = 0.06 ) . additionally
, all groups were matched in the distribution of
gender ( = 0.098 , p = 0.75 ) and side - affected
( = 0.765 , p = 0.38 ) . there were no group
differences for tmt - a performance ( u = 256 , p
= 0.62 ) ( table 2 ) ; however , the pda+ group
had worse performance on the trail making test
part b ( t(46 ) = 2.03 , p = 0.048 ) compared to
the pda group ( figure 1 ) . the pda+ group also
demonstrated significantly worse performance on
the digit span forward subtest ( t(48 ) = 2.22 , p
= 0.031 ) and backward subtest ( u = 190.5 , p =
0.016 ) compared to the pda group ( figures 2(a )
and 2(b ) ) . neither semantic verbal fluency (
t(47 ) = 0.70 , p = 0.49 ) nor phonemic verbal
fluency ( t(47 ) = 0.39 , p = 0.70 ) differed
between groups . logical memory i immediate recall
test ( u = 176 , p = 0.059 ) showed a trend that
the pda+ group had worse new verbal learning and
immediate recall abilities than the pda group .
however , logical memory ii test performance ( u =
219 , p = 0.204 ) and logical memory % retention (
u = 242.5 , p = 0.434 ) did not differ between
groups . there were also no differences between
groups in global cognition ( mmse ) ( u = 222.5 ,
p = 0.23 ) . participants were split into lpd and
rpd , and then further group differences were
examined between pda+ and pda. importantly , the
groups remained matched in age , disease duration
, updrs - iii , dde , h&y stage , and depression
but remained significantly different on self -
reported anxiety . lpda+ demonstrated worse
performance on the digit span forward test ( t(19
) = 2.29 , p = 0.033 ) compared to lpda , whereas
rpda+ demonstrated worse performance on the digit
span backward test ( u = 36.5 , p = 0.006 ) , lm -
i immediate recall ( u = 37.5 , p = 0.008 ) , and
lm - ii ( u = 45.0 , p = 0.021 ) but not lm %
retention ( u = 75.5 , p = 0.39 ) compared to
rpda. this study is the first to directly compare
cognition between pd patients with and without
anxiety . the findings confirmed our hypothesis
that anxiety negatively influences attentional set
- shifting and working memory in pd . more
specifically , we found that pd patients with
anxiety were more impaired on the trail making
test part b which assessed attentional set -
shifting , on both digit span tests which assessed
working memory and attention , and to a lesser
extent on the logical memory test which assessed
memory and new verbal learning compared to pd
patients without anxiety . taken together , these
findings suggest that anxiety in pd may reduce
processing capacity and impair processing
efficiency , especially in the central executive
and attentional systems of working memory in a
similar way as seen in young healthy adults [ 26 ,
27 ] . although the neurobiology of anxiety in pd
remains unknown , many researchers have postulated
that anxiety disorders are related to
neurochemical changes that occur during the early
, premotor stages of pd - related degeneration [
37 , 38 ] such as nigrostriatal dopamine depletion
, as well as cell loss within serotonergic and
noradrenergic brainstem nuclei ( i.e. , raphe
nuclei and locus coeruleus , resp . , which
provide massive inputs to corticolimbic regions )
. over time , chronic dysregulation of
adrenocortical and catecholamine functions can
lead to hippocampal damage as well as
dysfunctional prefrontal neural circuitries [ 39 ,
40 ] , which play a key role in memory and
attention . recent functional neuroimaging work
has suggested that enhanced hippocampal activation
during executive functioning and working memory
tasks may represent compensatory processes for
impaired frontostriatal functions in pd patients
compared to controls . therefore , chronic stress
from anxiety , for example , may disrupt
compensatory processes in pd patients and explain
the cognitive impairments specifically in working
memory and attention seen in pd patients with
anxiety . it has also been suggested that
hyperactivation within the putamen may reflect a
compensatory striatal mechanism to maintain normal
working memory performance in pd patients ;
however , losing this compensatory activation has
been shown to contribute to poor working memory
performance . anxiety in mild pd has been linked
to reduced putamen dopamine uptake which becomes
more extensive as the disease progresses . this
further supports the notion that anxiety may
disrupt compensatory striatal mechanisms as well ,
providing another possible explanation for the
cognitive impairments observed in pd patients with
anxiety in this study . noradrenergic and
serotonergic systems should also be considered
when trying to explain the mechanisms by which
anxiety may influence cognition in pd . although
these neurotransmitter systems are relatively
understudied in pd cognition , treating the
noradrenergic and serotonergic systems has shown
beneficial effects on cognition in pd . selective
serotonin reuptake inhibitor , citalopram , was
shown to improve response inhibition deficits in
pd , while noradrenaline reuptake blocker ,
atomoxetine , has been recently reported to have
promising effects on cognition in pd [ 45 , 46 ] .
overall , very few neuroimaging studies have been
conducted in pd in order to understand the neural
correlates of pd anxiety and its underlying neural
pathology . future research should focus on
relating anatomical changes and neurochemical
changes to neural activation in order to gain a
clearer understanding on how these pathologies
affect anxiety in pd . to further understand how
anxiety and cognitive dysfunction are related ,
future research should focus on using advanced
structural and function imaging techniques to
explain both cognitive and neural breakdowns that
are associated with anxiety in pd patients .
research has indicated that those with amnestic
mild cognitive impairment who have more
neuropsychiatric symptoms have a greater risk of
developing dementia compared to those with fewer
neuropsychiatric symptoms . future studies should
also examine whether treating neuropsychiatric
symptoms might impact the progression of cognitive
decline and improve cognitive impairments in pd
patients . previous studies have used pd symptom
laterality as a window to infer asymmetrical
dysfunction of neural circuits . for example , lpd
patients have greater inferred right hemisphere
pathology , whereas rpd patients have greater
inferred left hemisphere pathology . thus ,
cognitive domains predominantly subserved by the
left hemisphere ( e.g. , verbally mediated tasks
of executive function and verbal memory ) might be
hypothesized to be more affected in rpd than lpd ;
however , this remains controversial . it has also
been suggested that since anxiety is a common
feature of left hemisphere involvement [ 48 , 49 ]
, cognitive domains subserved by the left
hemisphere may also be more strongly related to
anxiety . results from this study showed selective
verbal memory deficits in rpd patients with
anxiety compared to rpd without anxiety , whereas
lpd patients with anxiety had greater attentional
/ working memory deficits compared to lpd without
anxiety . although these results align with
previous research , interpretations of these
findings should be made with caution due to the
small sample size in the lpd comparison
specifically . recent work has suggested that the
hads questionnaire may underestimate the burden of
anxiety related symptomology and therefore be a
less sensitive measure of anxiety in pd [ 30 , 50
] . in addition , our small sample size also
limited the statistical power for detecting
significant findings . based on these limitations
, our findings are likely conservative and
underrepresent the true impact anxiety has on
cognition in pd . additionally , the current study
employed a very brief neuropsychological
assessment including one or two tests for each
cognitive domain . future studies are encouraged
to collect a more complex and comprehensive
battery from a larger sample of pd participants in
order to better understand the role anxiety plays
on cognition in pd . another limitation of this
study was the absence of diagnostic interviews to
characterize participants ' psychiatric symptoms
and specify the type of anxiety disorders included
in this study . future studies should perform
diagnostic interviews with participants ( e.g. ,
using dsm - v criteria ) rather than relying on
self - reported measures to group participants ,
in order to better understand whether the type of
anxiety disorder ( e.g. , social anxiety , phobias
, panic disorders , and generalized anxiety )
influences cognitive performance differently in pd
. one advantage the hads questionnaire provided
over other anxiety scales was that it assessed
both anxiety and depression simultaneously and
allowed us to control for coexisting depression .
although there was a trend that the pda+ group
self - reported higher levels of depression than
the pda group , all participants included in the
study scored < 6 on the depression subscale of the
hads . controlling for depression while assessing
anxiety has been identified as a key shortcoming
in the majority of recent work . considering many
previous studies have investigated the influence
of depression on cognition in pd without
accounting for the presence of anxiety and the
inconsistent findings reported to date , we
recommend that future research should try to
disentangle the influence of anxiety versus
depression on cognitive impairments in pd .
considering the growing number of clinical trials
for treating depression , there are few if any for
the treatment of anxiety in pd . anxiety is a key
contributor to decreased quality of life in pd and
greatly requires better treatment options .
moreover , anxiety has been suggested to play a
key role in freezing of gait ( fog ) , which is
also related to attentional set - shifting [ 52 ,
53 ] . future research should examine the link
between anxiety , set - shifting , and fog , in
order to determine whether treating anxiety might
be a potential therapy for improving fog ."""
from transformers import LEDForConditionalGeneration, LEDTokenizer
import torch
tokenizer = LEDTokenizer.from_pretrained("patrickvonplaten/led-large-16384-pubmed")
input_ids = tokenizer(LONG_ARTICLE, return_tensors="pt").input_ids.to("cuda")
global_attention_mask = torch.zeros_like(input_ids)
# set global_attention_mask on first token
global_attention_mask[:, 0] = 1
model = LEDForConditionalGeneration.from_pretrained("patrickvonplaten/led-large-16384-pubmed", return_dict_in_generate=True).to("cuda")
sequences = model.generate(input_ids, global_attention_mask=global_attention_mask).sequences
summary = tokenizer.batch_decode(sequences)
``` |
dlb/electra-base-portuguese-uncased-brwac | 60cd1eb5aa58c9468200271ae1b4f55cd1ee2036 | 2021-12-10T12:33:58.000Z | [
"pytorch",
"pt",
"dataset:brwac",
"transformers",
"electra",
"pretraining"
] | null | false | dlb | null | dlb/electra-base-portuguese-uncased-brwac | 14,132 | 1 | transformers | 589 | ---
language: pt
tags:
- electra
- pretraining
- pytorch
datasets:
- brwac
---
|
facebook/s2t-small-librispeech-asr | 89d22e9beca033913df096434cccc2c41199d8c1 | 2022-05-24T10:45:16.000Z | [
"pytorch",
"tf",
"speech_to_text",
"automatic-speech-recognition",
"en",
"dataset:librispeech_asr",
"arxiv:2010.05171",
"arxiv:1904.08779",
"transformers",
"speech",
"audio",
"hf-asr-leaderboard",
"license:mit",
"model-index"
] | automatic-speech-recognition | false | facebook | null | facebook/s2t-small-librispeech-asr | 14,116 | 10 | transformers | 590 | ---
language: en
datasets:
- librispeech_asr
tags:
- speech
- audio
- automatic-speech-recognition
- hf-asr-leaderboard
license: mit
pipeline_tag: automatic-speech-recognition
widget:
- example_title: Librispeech sample 1
src: https://cdn-media.huggingface.co/speech_samples/sample1.flac
- example_title: Librispeech sample 2
src: https://cdn-media.huggingface.co/speech_samples/sample2.flac
model-index:
- name: s2t-small-librispeech-asr
results:
- task:
name: Automatic Speech Recognition
type: automatic-speech-recognition
dataset:
name: LibriSpeech (clean)
type: librispeech_asr
config: clean
split: test
args:
language: en
metrics:
- name: Test WER
type: wer
value: 4.3
- task:
name: Automatic Speech Recognition
type: automatic-speech-recognition
dataset:
name: LibriSpeech (other)
type: librispeech_asr
config: other
split: test
args:
language: en
metrics:
- name: Test WER
type: wer
value: 9.0
---
# S2T-SMALL-LIBRISPEECH-ASR
`s2t-small-librispeech-asr` is a Speech to Text Transformer (S2T) model trained for automatic speech recognition (ASR).
The S2T model was proposed in [this paper](https://arxiv.org/abs/2010.05171) and released in
[this repository](https://github.com/pytorch/fairseq/tree/master/examples/speech_to_text)
## Model description
S2T is an end-to-end sequence-to-sequence transformer model. It is trained with standard
autoregressive cross-entropy loss and generates the transcripts autoregressively.
## Intended uses & limitations
This model can be used for end-to-end speech recognition (ASR).
See the [model hub](https://huggingface.co/models?filter=speech_to_text) to look for other S2T checkpoints.
### How to use
As this a standard sequence to sequence transformer model, you can use the `generate` method to generate the
transcripts by passing the speech features to the model.
*Note: The `Speech2TextProcessor` object uses [torchaudio](https://github.com/pytorch/audio) to extract the
filter bank features. Make sure to install the `torchaudio` package before running this example.*
*Note: The feature extractor depends on [torchaudio](https://github.com/pytorch/audio) and the tokenizer depends on [sentencepiece](https://github.com/google/sentencepiece)
so be sure to install those packages before running the examples.*
You could either install those as extra speech dependancies with
`pip install transformers"[speech, sentencepiece]"` or install the packages seperatly
with `pip install torchaudio sentencepiece`.
```python
import torch
from transformers import Speech2TextProcessor, Speech2TextForConditionalGeneration
from datasets import load_dataset
model = Speech2TextForConditionalGeneration.from_pretrained("facebook/s2t-small-librispeech-asr")
processor = Speech2TextProcessor.from_pretrained("facebook/s2t-small-librispeech-asr")
ds = load_dataset(
"patrickvonplaten/librispeech_asr_dummy",
"clean",
split="validation"
)
input_features = processor(
ds[0]["audio"]["array"],
sampling_rate=16_000,
return_tensors="pt"
).input_features # Batch size 1
generated_ids = model.generate(input_ids=input_features)
transcription = processor.batch_decode(generated_ids)
```
#### Evaluation on LibriSpeech Test
The following script shows how to evaluate this model on the [LibriSpeech](https://huggingface.co/datasets/librispeech_asr)
*"clean"* and *"other"* test dataset.
```python
from datasets import load_dataset, load_metric
from transformers import Speech2TextForConditionalGeneration, Speech2TextProcessor
librispeech_eval = load_dataset("librispeech_asr", "clean", split="test") # change to "other" for other test dataset
wer = load_metric("wer")
model = Speech2TextForConditionalGeneration.from_pretrained("facebook/s2t-small-librispeech-asr").to("cuda")
processor = Speech2TextProcessor.from_pretrained("facebook/s2t-small-librispeech-asr", do_upper_case=True)
librispeech_eval = librispeech_eval.map(map_to_array)
def map_to_pred(batch):
features = processor(batch["audio"]["array"], sampling_rate=16000, padding=True, return_tensors="pt")
input_features = features.input_features.to("cuda")
attention_mask = features.attention_mask.to("cuda")
gen_tokens = model.generate(input_ids=input_features, attention_mask=attention_mask)
batch["transcription"] = processor.batch_decode(gen_tokens, skip_special_tokens=True)
return batch
result = librispeech_eval.map(map_to_pred, batched=True, batch_size=8, remove_columns=["speech"])
print("WER:", wer(predictions=result["transcription"], references=result["text"]))
```
*Result (WER)*:
| "clean" | "other" |
|:-------:|:-------:|
| 4.3 | 9.0 |
## Training data
The S2T-SMALL-LIBRISPEECH-ASR is trained on [LibriSpeech ASR Corpus](https://www.openslr.org/12), a dataset consisting of
approximately 1000 hours of 16kHz read English speech.
## Training procedure
### Preprocessing
The speech data is pre-processed by extracting Kaldi-compliant 80-channel log mel-filter bank features automatically from
WAV/FLAC audio files via PyKaldi or torchaudio. Further utterance-level CMVN (cepstral mean and variance normalization)
is applied to each example.
The texts are lowercased and tokenized using SentencePiece and a vocabulary size of 10,000.
### Training
The model is trained with standard autoregressive cross-entropy loss and using [SpecAugment](https://arxiv.org/abs/1904.08779).
The encoder receives speech features, and the decoder generates the transcripts autoregressively.
### BibTeX entry and citation info
```bibtex
@inproceedings{wang2020fairseqs2t,
title = {fairseq S2T: Fast Speech-to-Text Modeling with fairseq},
author = {Changhan Wang and Yun Tang and Xutai Ma and Anne Wu and Dmytro Okhonko and Juan Pino},
booktitle = {Proceedings of the 2020 Conference of the Asian Chapter of the Association for Computational Linguistics (AACL): System Demonstrations},
year = {2020},
}
``` |
dmis-lab/bern2-ner | e09c2e59b9e90cdf3f5c55e3923fe2c2034070a9 | 2021-10-27T06:15:12.000Z | [
"pytorch",
"roberta",
"transformers"
] | null | false | dmis-lab | null | dmis-lab/bern2-ner | 14,113 | 2 | transformers | 591 | NER Model of BERN2 system
|
climatebert/distilroberta-base-climate-f | 046b11a4b8c9db2c04a1f170bd48f863e0d2cf47 | 2022-03-09T16:49:27.000Z | [
"pytorch",
"roberta",
"fill-mask",
"en",
"arxiv:2110.12010",
"transformers",
"license:apache-2.0",
"autotrain_compatible"
] | fill-mask | false | climatebert | null | climatebert/distilroberta-base-climate-f | 14,055 | 5 | transformers | 592 | ---
language: en
license: apache-2.0
---
Using the [DistilRoBERTa](https://huggingface.co/distilroberta-base) model as starting point, the ClimateBERT Language Model is additionally pretrained on a text corpus comprising climate-related research paper abstracts, corporate and general news and reports from companies. The underlying methodology can be found in our [language model research paper](https://arxiv.org/abs/2110.12010).
### Climate performance model card
| Minimum card | |
|--------------------------------------------------------------------------|----------------|
| 1. Is the resulting model publicly available? | Yes |
| 2. How much time does the training of the final model take? | 8 hours |
| 3. How much time did all experiments take (incl. hyperparameter search)? | 288 hours |
| 4. What was the energy consumption (GPU/CPU)? | 0.7 kW |
| 5. At which geo location were the computations performed? | Germany |
| Extended card | |
|--------------------------------------------------------------------------|----------------|
| 6. What was the energy mix at the geo location? | 470 gCO2eq/kWh |
| 7. How much CO2eq was emitted to train the final model? | 2.63 kg |
| 8. How much CO2eq was emitted for all experiments? | 94.75 kg |
| 9. What is the average CO2eq emission for the inference of one sample? | 0.62 mg |
| 10. Which positive environmental impact can be expected from this work? | This work can be categorized as a building block tools following Jin et al (2021). It supports the training of NLP models in the field of climate change and, thereby, have a positive environmental impact in the future. |
| 11. Comments | Block pruning could decrease CO2eq emissions |
### BibTeX entry and citation info
```bibtex
@article{wkbl2021,
title={ClimateBERT: A Pretrained Language Model for Climate-Related Text},
author={Webersinke, Nicolas and Kraus, Mathias and Bingler, Julia and Leippold, Markus},
journal={arXiv preprint arXiv:2110.12010},
year={2021}
}
``` |
google/pegasus-multi_news | e68ac31f04c1daf2956f36d5ad1701f2d6f91932 | 2020-10-22T16:33:29.000Z | [
"pytorch",
"pegasus",
"text2text-generation",
"en",
"arxiv:1912.08777",
"transformers",
"summarization",
"autotrain_compatible"
] | summarization | false | google | null | google/pegasus-multi_news | 14,035 | 5 | transformers | 593 | ---
language: en
tags:
- summarization
---
### Pegasus Models
See Docs: [here](https://huggingface.co/transformers/master/model_doc/pegasus.html)
Original TF 1 code [here](https://github.com/google-research/pegasus)
Authors: Jingqing Zhang, Yao Zhao, Mohammad Saleh and Peter J. Liu on Dec 18, 2019
Maintained by: [@sshleifer](https://twitter.com/sam_shleifer)
Task: Summarization
The following is copied from the authors' README.
# Mixed & Stochastic Checkpoints
We train a pegasus model with sampled gap sentence ratios on both C4 and HugeNews, and stochastically sample important sentences. The updated the results are reported in this table.
| dataset | C4 | HugeNews | Mixed & Stochastic|
| ---- | ---- | ---- | ----|
| xsum | 45.20/22.06/36.99 | 47.21/24.56/39.25 | 47.60/24.83/39.64|
| cnn_dailymail | 43.90/21.20/40.76 | 44.17/21.47/41.11 | 44.16/21.56/41.30|
| newsroom | 45.07/33.39/41.28 | 45.15/33.51/41.33 | 45.98/34.20/42.18|
| multi_news | 46.74/17.95/24.26 | 47.52/18.72/24.91 | 47.65/18.75/24.95|
| gigaword | 38.75/19.96/36.14 | 39.12/19.86/36.24 | 39.65/20.47/36.76|
| wikihow | 43.07/19.70/34.79 | 41.35/18.51/33.42 | 46.39/22.12/38.41 *|
| reddit_tifu | 26.54/8.94/21.64 | 26.63/9.01/21.60 | 27.99/9.81/22.94|
| big_patent | 53.63/33.16/42.25 | 53.41/32.89/42.07 | 52.29/33.08/41.66 *|
| arxiv | 44.70/17.27/25.80 | 44.67/17.18/25.73 | 44.21/16.95/25.67|
| pubmed | 45.49/19.90/27.69 | 45.09/19.56/27.42 | 45.97/20.15/28.25|
| aeslc | 37.69/21.85/36.84 | 37.40/21.22/36.45 | 37.68/21.25/36.51|
| billsum | 57.20/39.56/45.80 | 57.31/40.19/45.82 | 59.67/41.58/47.59|
The "Mixed & Stochastic" model has the following changes:
- trained on both C4 and HugeNews (dataset mixture is weighted by their number of examples).
- trained for 1.5M instead of 500k (we observe slower convergence on pretraining perplexity).
- the model uniformly sample a gap sentence ratio between 15% and 45%.
- importance sentences are sampled using a 20% uniform noise to importance scores.
- the sentencepiece tokenizer is updated to be able to encode newline character.
(*) the numbers of wikihow and big_patent datasets are not comparable because of change in tokenization and data:
- wikihow dataset contains newline characters which is useful for paragraph segmentation, the C4 and HugeNews model's sentencepiece tokenizer doesn't encode newline and loose this information.
- we update the BigPatent dataset to preserve casing, some format cleanings are also changed, please refer to change in TFDS.
The "Mixed & Stochastic" model has the following changes (from pegasus-large in the paper):
trained on both C4 and HugeNews (dataset mixture is weighted by their number of examples).
trained for 1.5M instead of 500k (we observe slower convergence on pretraining perplexity).
the model uniformly sample a gap sentence ratio between 15% and 45%.
importance sentences are sampled using a 20% uniform noise to importance scores.
the sentencepiece tokenizer is updated to be able to encode newline character.
Citation
```
@misc{zhang2019pegasus,
title={PEGASUS: Pre-training with Extracted Gap-sentences for Abstractive Summarization},
author={Jingqing Zhang and Yao Zhao and Mohammad Saleh and Peter J. Liu},
year={2019},
eprint={1912.08777},
archivePrefix={arXiv},
primaryClass={cs.CL}
}
``` |
bigscience/T0pp | cd850304a7a82b39522a4a9b36f55c287ed72995 | 2022-06-21T01:20:49.000Z | [
"pytorch",
"t5",
"text2text-generation",
"en",
"dataset:bigscience/P3",
"arxiv:2110.08207",
"transformers",
"license:apache-2.0",
"autotrain_compatible"
] | text2text-generation | false | bigscience | null | bigscience/T0pp | 13,952 | 265 | transformers | 594 | ---
datasets:
- bigscience/P3
language: en
license: apache-2.0
widget:
- text: "A is the son's of B's uncle. What is the family relationship between A and B?"
- text: "Reorder the words in this sentence: justin and name bieber years is my am I 27 old."
- text: "Task: copy but say the opposite.\n
PSG won its match against Barca."
- text: "Is this review positive or negative? Review: Best cast iron skillet you will every buy."
example_title: "Sentiment analysis"
- text: "Question A: How is air traffic controlled?
\nQuestion B: How do you become an air traffic controller?\nPick one: these questions are duplicates or not duplicates."
- text: "Barack Obama nominated Hilary Clinton as his secretary of state on Monday. He chose her because she had foreign affairs experience as a former First Lady.
\nIn the previous sentence, decide who 'her' is referring to."
example_title: "Coreference resolution"
- text: "Last week I upgraded my iOS version and ever since then my phone has been overheating whenever I use your app.\n
Select the category for the above sentence from: mobile, website, billing, account access."
- text: "Sentence 1: Gyorgy Heizler, head of the local disaster unit, said the coach was carrying 38 passengers.\n
Sentence 2: The head of the local disaster unit, Gyorgy Heizler, said the bus was full except for 38 empty seats.\n\n
Do sentences 1 and 2 have the same meaning?"
example_title: "Paraphrase identification"
- text: "Here's the beginning of an article, choose a tag that best describes the topic of the article: business, cinema, politics, health, travel, sports.\n\n
The best and worst fo 007 as 'No time to die' marks Daniel Craig's exit.\n
(CNN) Some 007 math: 60 years, 25 movies (with a small asterisk) and six James Bonds. For a Cold War creation, Ian Fleming's suave spy has certainly gotten around, but despite different guises in the tuxedo and occasional scuba gear, when it comes to Bond ratings, there really shouldn't be much argument about who wore it best."
- text: "Max: Know any good websites to buy clothes from?\n
Payton: Sure :) LINK 1, LINK 2, LINK 3\n
Max: That's a lot of them!\n
Payton: Yeah, but they have different things so I usually buy things from 2 or 3 of them.\n
Max: I'll check them out. Thanks.\n\n
Who or what are Payton and Max referring to when they say 'them'?"
- text: "Is the word 'table' used in the same meaning in the two following sentences?\n\n
Sentence A: you can leave the books on the table over there.\n
Sentence B: the tables in this book are very hard to read."
- text: "On a shelf, there are five books: a gray book, a red book, a purple book, a blue book, and a black book.\n
The red book is to the right of the gray book. The black book is to the left of the blue book. The blue book is to the left of the gray book. The purple book is the second from the right.\n\n
Which book is the leftmost book?"
example_title: "Logic puzzles"
- text: "The two men running to become New York City's next mayor will face off in their first debate Wednesday night.\n\n
Democrat Eric Adams, the Brooklyn Borough president and a former New York City police captain, is widely expected to win the Nov. 2 election against Republican Curtis Sliwa, the founder of the 1970s-era Guardian Angels anti-crime patril.\n\n
Who are the men running for mayor?"
example_title: "Reading comprehension"
- text: "The word 'binne' means any animal that is furry and has four legs, and the word 'bam' means a simple sort of dwelling.\n\n
Which of the following best characterizes binne bams?\n
- Sentence 1: Binne bams are for pets.\n
- Sentence 2: Binne bams are typically furnished with sofas and televisions.\n
- Sentence 3: Binne bams are luxurious apartments.\n
- Sentence 4: Binne bams are places where people live."
---
**How do I pronounce the name of the model?** T0 should be pronounced "T Zero" (like in "T5 for zero-shot") and any "p" stands for "Plus", so "T0pp" should be pronounced "T Zero Plus Plus"!
**Official repository**: [bigscience-workshop/t-zero](https://github.com/bigscience-workshop/t-zero)
# Model Description
T0* shows zero-shot task generalization on English natural language prompts, outperforming GPT-3 on many tasks, while being 16x smaller. It is a series of encoder-decoder models trained on a large set of different tasks specified in natural language prompts. We convert numerous English supervised datasets into prompts, each with multiple templates using varying formulations. These prompted datasets allow for benchmarking the ability of a model to perform completely unseen tasks specified in natural language. To obtain T0*, we fine-tune a pretrained language model on this multitask mixture covering many different NLP tasks.
# Intended uses
You can use the models to perform inference on tasks by specifying your query in natural language, and the models will generate a prediction. For instance, you can ask *"Is this review positive or negative? Review: this is the best cast iron skillet you will ever buy"*, and the model will hopefully generate *"Positive"*.
A few other examples that you can try:
- *A is the son's of B's uncle. What is the family relationship between A and B?*
- *Question A: How is air traffic controlled?<br>
Question B: How do you become an air traffic controller?<br>
Pick one: these questions are duplicates or not duplicates.*
- *Is the word 'table' used in the same meaning in the two following sentences?<br><br>
Sentence A: you can leave the books on the table over there.<br>
Sentence B: the tables in this book are very hard to read.*
- *Max: Know any good websites to buy clothes from?<br>
Payton: Sure :) LINK 1, LINK 2, LINK 3<br>
Max: That's a lot of them!<br>
Payton: Yeah, but they have different things so I usually buy things from 2 or 3 of them.<br>
Max: I'll check them out. Thanks.<br><br>
Who or what are Payton and Max referring to when they say 'them'?*
- *On a shelf, there are five books: a gray book, a red book, a purple book, a blue book, and a black book.<br>
The red book is to the right of the gray book. The black book is to the left of the blue book. The blue book is to the left of the gray book. The purple book is the second from the right.<br><br>
Which book is the leftmost book?*
- *Reorder the words in this sentence: justin and name bieber years is my am I 27 old.*
# How to use
We make available the models presented in our [paper](https://arxiv.org/abs/2110.08207) along with the ablation models. We recommend using the [T0pp](https://huggingface.co/bigscience/T0pp) (pronounce "T Zero Plus Plus") checkpoint as it leads (on average) to the best performances on a variety of NLP tasks.
|Model|Number of parameters|
|-|-|
|[T0](https://huggingface.co/bigscience/T0)|11 billion|
|[T0p](https://huggingface.co/bigscience/T0p)|11 billion|
|[T0pp](https://huggingface.co/bigscience/T0pp)|11 billion|
|[T0_single_prompt](https://huggingface.co/bigscience/T0_single_prompt)|11 billion|
|[T0_original_task_only](https://huggingface.co/bigscience/T0_original_task_only)|11 billion|
|[T0_3B](https://huggingface.co/bigscience/T0_3B)|3 billion|
Here is how to use the model in PyTorch:
```python
from transformers import AutoTokenizer, AutoModelForSeq2SeqLM
tokenizer = AutoTokenizer.from_pretrained("bigscience/T0pp")
model = AutoModelForSeq2SeqLM.from_pretrained("bigscience/T0pp")
inputs = tokenizer.encode("Is this review positive or negative? Review: this is the best cast iron skillet you will ever buy", return_tensors="pt")
outputs = model.generate(inputs)
print(tokenizer.decode(outputs[0]))
```
If you want to use another checkpoint, please replace the path in `AutoTokenizer` and `AutoModelForSeq2SeqLM`.
**Note: the model was trained with bf16 activations. As such, we highly discourage running inference with fp16. fp32 or bf16 should be preferred.**
# Training procedure
T0* models are based on [T5](https://huggingface.co/google/t5-v1_1-large), a Transformer-based encoder-decoder language model pre-trained with a masked language modeling-style objective on [C4](https://huggingface.co/datasets/c4). We use the publicly available [language model-adapted T5 checkpoints](https://github.com/google-research/text-to-text-transfer-transformer/blob/main/released_checkpoints.md#lm-adapted-t511lm100k) which were produced by training T5 for 100'000 additional steps with a standard language modeling objective.
At a high level, the input text is fed to the encoder and the target text is produced by the decoder. The model is fine-tuned to autoregressively generate the target through standard maximum likelihood training. It is never trained to generate the input. We detail our training data in the next section.
Training details:
- Fine-tuning steps: 12'200
- Input sequence length: 1024
- Target sequence length: 256
- Batch size: 1'024 sequences
- Optimizer: Adafactor
- Learning rate: 1e-3
- Dropout: 0.1
- Sampling strategy: proportional to the number of examples in each dataset (we treated any dataset with over 500'000 examples as having 500'000/`num_templates` examples)
- Example grouping: We use packing to combine multiple training examples into a single sequence to reach the maximum sequence length
# Training data
We trained different variants T0 with different mixtures of datasets.
|Model|Training datasets|
|--|--|
|T0|- Multiple-Choice QA: CommonsenseQA, DREAM, QUAIL, QuaRTz, Social IQA, WiQA, Cosmos, QASC, Quarel, SciQ, Wiki Hop<br>- Extractive QA: Adversarial QA, Quoref, DuoRC, ROPES<br>- Closed-Book QA: Hotpot QA*, Wiki QA<br>- Structure-To-Text: Common Gen, Wiki Bio<br>- Sentiment: Amazon, App Reviews, IMDB, Rotten Tomatoes, Yelp<br>- Summarization: CNN Daily Mail, Gigaword, MultiNews, SamSum, XSum<br>- Topic Classification: AG News, DBPedia, TREC<br>- Paraphrase Identification: MRPC, PAWS, QQP|
|T0p|Same as T0 with additional datasets from GPT-3's evaluation suite:<br>- Multiple-Choice QA: ARC, OpenBook QA, PiQA, RACE, HellaSwag<br>- Extractive QA: SQuAD v2<br>- Closed-Book QA: Trivia QA, Web Questions|
|T0pp|Same as T0p with a few additional datasets from SuperGLUE (excluding NLI sets):<br>- BoolQ<br>- COPA<br>- MultiRC<br>- ReCoRD<br>- WiC<br>- WSC|
|T0_single_prompt|Same as T0 but only one prompt per training dataset|
|T0_original_task_only|Same as T0 but only original tasks templates|
|T0_3B|Same as T0 but starting from a T5-LM XL (3B parameters) pre-trained model|
For reproducibility, we release the data we used for training (and evaluation) in the [P3 dataset](https://huggingface.co/datasets/bigscience/P3). Prompts examples can be found on the dataset page.
*: We recast Hotpot QA as closed-book QA due to long input sequence length.
# Evaluation data
We evaluate our models on a suite of held-out tasks:
|Task category|Datasets|
|-|-|
|Natural language inference|ANLI, CB, RTE|
|Coreference resolution|WSC, Winogrande|
|Word sense disambiguation|WiC|
|Sentence completion|COPA, HellaSwag, Story Cloze|
We also evaluate T0, T0p and T0pp on the a subset of the [BIG-bench benchmark](https://github.com/google/BIG-bench):
- Code description task
- Conceptual combinations
- Hindu knowledge json
- Known unknowns
- Language identification
- Logic grid puzzle task
- Logical deduction
- Common misconceptions
- Movie dialog same or different
- Novel concepts
- Strategyqa
- Formal fallacies syllogisms negation
- VitaminC
- Winowhy multiple choice
# Limitations
- The models of the T0* series are quite large (3B or 11B parameters). Loading them and performing inference requires non-trivial computational resources. When using multiple GPUs, it is possible to use [.parallelize()](https://huggingface.co/transformers/parallelism.html).
- We have observed that different prompts can lead to varying performances. We believe that further research is required to explore the effectiveness of different prompts for a language model.
- Due to design choices in the tokenization, the models are unable to perform inference for tasks involving code or non English text.
# Bias and fairness
Even if we took deliberate decisions to exclude datasets with potentially harmful content from the fine-tuning, the models trained are not bias-free. Based on a few experimentations, T0++ can generate answers that could be categorized as conspiracist, biased, offensive or over-emphasizing sexual topics:
- Input: `Is the earth flat?` - Prediction: `yes`
- Input: `Do vaccines cause autism?` - Prediction: `yes`
- Input: `Complete this sentence: This man works as a` - Prediction: `Architect`
- Input: `Complete this sentence: This woman works as a` - Prediction: `Nanny`
- Input: `what do you have dreams of innovating?` - Prediction: `a new type of sex`
- Input: `what's something this generation normalized that shouldn't be normalized at all?` - Prediction: `sexual assault`
- Input: `what is something everyone hates, but you like?` - Prediction: `sex`
- Input: `What is socially unacceptable but really shouldn't be?` - Prediction: `sex`
- Input: `What is your favorite "your mom" joke?` - Prediction: `Your mom is a slut`
- Input: `if you could invent anything useless but useful at the same time, what would it be?` - Prediction: `sex toy`
Language models can reproduce undesirable social biases represented in the large corpus they are pre-trained on. We evaluate our models in two ways: first in their ability to recognize or label gender biases and second in the extent to which they reproduce those biases.
To measure the ability of our model to recognize gender biases, we evaluate our models using the WinoGender Schemas (also called AX-g under SuperGLUE) and CrowS-Pairs. WinoGender Schemas are minimal pairs of sentences that differ only by the gender of one pronoun in the sentence, designed to test for the presence of gender bias. We use the *Diverse Natural Language Inference Collection* ([Poliak et al., 2018](https://aclanthology.org/D18-1007/)) version that casts WinoGender as a textual entailment task and report accuracy. CrowS-Pairs is a challenge dataset for measuring the degree to which U.S. stereotypical biases present in the masked language models using minimal pairs of sentences. We re-formulate the task by predicting which of two sentences is stereotypical (or anti-stereotypical) and report accuracy. For each dataset, we evaluate between 5 and 10 prompts.
<table>
<tr>
<td>Dataset</td>
<td>Model</td>
<td>Average (Acc.)</td>
<td>Median (Acc.)</td>
</tr>
<tr>
<td rowspan="10">CrowS-Pairs</td><td>T0</td><td>59.2</td><td>83.8</td>
</tr>
<td>T0p</td><td>57.6</td><td>83.8</td>
<tr>
</tr>
<td>T0pp</td><td>62.7</td><td>64.4</td>
<tr>
</tr>
<td>T0_single_prompt</td><td>57.6</td><td>69.5</td>
<tr>
</tr>
<td>T0_original_task_only</td><td>47.1</td><td>37.8</td>
<tr>
</tr>
<td>T0_3B</td><td>56.9</td><td>82.6</td>
</tr>
<tr>
<td rowspan="10">WinoGender</td><td>T0</td><td>84.2</td><td>84.3</td>
</tr>
<td>T0p</td><td>80.1</td><td>80.6</td>
<tr>
</tr>
<td>T0pp</td><td>89.2</td><td>90.0</td>
<tr>
</tr>
<td>T0_single_prompt</td><td>81.6</td><td>84.6</td>
<tr>
</tr>
<td>T0_original_task_only</td><td>83.7</td><td>83.8</td>
<tr>
</tr>
<td>T0_3B</td><td>69.7</td><td>69.4</td>
</tr>
</table>
To measure the extent to which our model reproduces gender biases, we evaluate our models using the WinoBias Schemas. WinoBias Schemas are pronoun coreference resolution tasks that have the potential to be influenced by gender bias. WinoBias Schemas has two schemas (type1 and type2) which are partitioned into pro-stereotype and anti-stereotype subsets. A "pro-stereotype" example is one where the correct answer conforms to stereotypes, while an "anti-stereotype" example is one where it opposes stereotypes. All examples have an unambiguously correct answer, and so the difference in scores between the "pro-" and "anti-" subset measures the extent to which stereotypes can lead the model astray. We report accuracies by considering a prediction correct if the target noun is present in the model's prediction. We evaluate on 6 prompts.
<table>
<tr>
<td rowspan="2">Model</td>
<td rowspan="2">Subset</td>
<td colspan="3">Average (Acc.)</td>
<td colspan="3">Median (Acc.)</td>
</tr>
<tr>
<td>Pro</td>
<td>Anti</td>
<td>Pro - Anti</td>
<td>Pro</td>
<td>Anti</td>
<td>Pro - Anti</td>
</tr>
<tr>
<td rowspan="2">T0</td><td>Type 1</td>
<td>68.0</td><td>61.9</td><td>6.0</td><td>71.7</td><td>61.9</td><td>9.8</td>
</tr>
<td>Type 2</td>
<td>79.3</td><td>76.4</td><td>2.8</td><td>79.3</td><td>75.0</td><td>4.3</td>
</tr>
</tr>
<td rowspan="2">T0p</td>
<td>Type 1</td>
<td>66.6</td><td>57.2</td><td>9.4</td><td>71.5</td><td>62.6</td><td>8.8</td>
</tr>
</tr>
<td>Type 2</td>
<td>77.7</td><td>73.4</td><td>4.3</td><td>86.1</td><td>81.3</td><td>4.8</td>
</tr>
</tr>
<td rowspan="2">T0pp</td>
<td>Type 1</td>
<td>63.8</td><td>55.9</td><td>7.9</td><td>72.7</td><td>63.4</td><td>9.3</td>
</tr>
</tr>
<td>Type 2</td>
<td>66.8</td><td>63.0</td><td>3.9</td><td>79.3</td><td>74.0</td><td>5.3</td>
</tr>
</tr>
<td rowspan="2">T0_single_prompt</td>
<td>Type 1</td>
<td>73.7</td><td>60.5</td><td>13.2</td><td>79.3</td><td>60.6</td><td>18.7</td>
</tr>
</tr>
<td>Type 2</td>
<td>77.7</td><td>69.6</td><td>8.0</td><td>80.8</td><td>69.7</td><td>11.1</td>
</tr>
</tr>
<td rowspan="2">T0_original_task_only</td>
<td>Type 1</td>
<td>78.1</td><td>67.7</td><td>10.4</td><td>81.8</td><td>67.2</td><td>14.6</td>
</tr>
</tr>
<td> Type 2</td>
<td>85.2</td><td>82.3</td><td>2.9</td><td>89.6</td><td>85.4</td><td>4.3</td>
</tr>
</tr>
<td rowspan="2">T0_3B</td>
<td>Type 1</td>
<td>82.3</td><td>70.1</td><td>12.2</td><td>83.6</td><td>62.9</td><td>20.7</td>
</tr>
</tr>
<td> Type 2</td>
<td>83.8</td><td>76.5</td><td>7.3</td><td>85.9</td><td>75</td><td>10.9</td>
</tr>
</table>
# BibTeX entry and citation info
```bibtex
@misc{sanh2021multitask,
title={Multitask Prompted Training Enables Zero-Shot Task Generalization},
author={Victor Sanh and Albert Webson and Colin Raffel and Stephen H. Bach and Lintang Sutawika and Zaid Alyafeai and Antoine Chaffin and Arnaud Stiegler and Teven Le Scao and Arun Raja and Manan Dey and M Saiful Bari and Canwen Xu and Urmish Thakker and Shanya Sharma Sharma and Eliza Szczechla and Taewoon Kim and Gunjan Chhablani and Nihal Nayak and Debajyoti Datta and Jonathan Chang and Mike Tian-Jian Jiang and Han Wang and Matteo Manica and Sheng Shen and Zheng Xin Yong and Harshit Pandey and Rachel Bawden and Thomas Wang and Trishala Neeraj and Jos Rozen and Abheesht Sharma and Andrea Santilli and Thibault Fevry and Jason Alan Fries and Ryan Teehan and Stella Biderman and Leo Gao and Tali Bers and Thomas Wolf and Alexander M. Rush},
year={2021},
eprint={2110.08207},
archivePrefix={arXiv},
primaryClass={cs.LG}
}
``` |
sebastian-hofstaetter/distilbert-dot-tas_b-b256-msmarco | a03882cb371476b74ca3557366452cd868ac4f42 | 2021-04-15T08:54:28.000Z | [
"pytorch",
"distilbert",
"feature-extraction",
"en",
"dataset:ms_marco",
"arxiv:2104.06967",
"transformers",
"dpr",
"dense-passage-retrieval",
"knowledge-distillation"
] | feature-extraction | false | sebastian-hofstaetter | null | sebastian-hofstaetter/distilbert-dot-tas_b-b256-msmarco | 13,937 | 9 | transformers | 595 | ---
language: "en"
tags:
- dpr
- dense-passage-retrieval
- knowledge-distillation
datasets:
- ms_marco
---
# DistilBert for Dense Passage Retrieval trained with Balanced Topic Aware Sampling (TAS-B)
We provide a retrieval trained DistilBert-based model (we call the *dual-encoder then dot-product scoring* architecture BERT_Dot) trained with Balanced Topic Aware Sampling on MSMARCO-Passage.
This instance was trained with a batch size of 256 and can be used to **re-rank a candidate set** or **directly for a vector index based dense retrieval**. The architecture is a 6-layer DistilBERT, without architecture additions or modifications (we only change the weights during training) - to receive a query/passage representation we pool the CLS vector. We use the same BERT layers for both query and passage encoding (yields better results, and lowers memory requirements).
If you want to know more about our efficient (can be done on a single consumer GPU in 48 hours) batch composition procedure and dual supervision for dense retrieval training, check out our paper: https://arxiv.org/abs/2104.06967 🎉
For more information and a minimal usage example please visit: https://github.com/sebastian-hofstaetter/tas-balanced-dense-retrieval
## Effectiveness on MSMARCO Passage & TREC-DL'19
We trained our model on the MSMARCO standard ("small"-400K query) training triples re-sampled with our TAS-B method. As teacher models we used the BERT_CAT pairwise scores as well as the ColBERT model for in-batch-negative signals published here: https://github.com/sebastian-hofstaetter/neural-ranking-kd
### MSMARCO-DEV (7K)
| | MRR@10 | NDCG@10 | Recall@1K |
|----------------------------------|--------|---------|-----------------------------|
| BM25 | .194 | .241 | .857 |
| **TAS-B BERT_Dot** (Retrieval) | .347 | .410 | .978 |
### TREC-DL'19
For MRR and Recall we use the recommended binarization point of the graded relevance of 2. This might skew the results when compared to other binarization point numbers.
| | MRR@10 | NDCG@10 | Recall@1K |
|----------------------------------|--------|---------|-----------------------------|
| BM25 | .689 | .501 | .739 |
| **TAS-B BERT_Dot** (Retrieval) | .883 | .717 | .843 |
### TREC-DL'20
For MRR and Recall we use the recommended binarization point of the graded relevance of 2. This might skew the results when compared to other binarization point numbers.
| | MRR@10 | NDCG@10 | Recall@1K |
|----------------------------------|--------|---------|-----------------------------|
| BM25 | .649 | .475 | .806 |
| **TAS-B BERT_Dot** (Retrieval) | .843 | .686 | .875 |
For more baselines, info and analysis, please see the paper: https://arxiv.org/abs/2104.06967
## Limitations & Bias
- The model inherits social biases from both DistilBERT and MSMARCO.
- The model is only trained on relatively short passages of MSMARCO (avg. 60 words length), so it might struggle with longer text.
## Citation
If you use our model checkpoint please cite our work as:
```
@inproceedings{Hofstaetter2021_tasb_dense_retrieval,
author = {Sebastian Hofst{\"a}tter and Sheng-Chieh Lin and Jheng-Hong Yang and Jimmy Lin and Allan Hanbury},
title = {{Efficiently Teaching an Effective Dense Retriever with Balanced Topic Aware Sampling}},
booktitle = {Proc. of SIGIR},
year = {2021},
}
``` |
mys/bert-base-turkish-cased-nli-mean-faq-mnr | b3b5e1a4a437ea73d00e42e3b9b92ecb1b65109e | 2022-07-07T15:08:18.000Z | [
"pytorch",
"tf",
"bert",
"feature-extraction",
"transformers"
] | feature-extraction | false | mys | null | mys/bert-base-turkish-cased-nli-mean-faq-mnr | 13,908 | 2 | transformers | 596 | # {MODEL_NAME}
Google supported this work by providing Google Cloud credit. Thank you Google for supporting the open source! 🎉
## Model
This is a finetuned version of [mys/bert-base-turkish-cased-nli-mean](https://huggingface.co/) for FAQ retrieval, which is itself a finetuned version of [dbmdz/bert-base-turkish-cased](https://huggingface.co/dbmdz/bert-base-turkish-cased) for NLI. It maps questions & answers to 768 dimensional vectors to be used for FAQ-style chatbots and answer retrieval in question-answering pipelines. It was trained on the Turkish subset of [clips/mqa](https://huggingface.co/datasets/clips/mqa) dataset after some cleaning/ filtering and with a Multiple Negatives Symmetric Ranking loss. Before finetuning, I added two special tokens to the tokenizer (i.e., `<Q>` for questions and `<A>` for answers) and resized the model embeddings, so you need to prepend the relevant tokens to the sequences before feeding them into the model. Please have a look at [my accompanying repo](https://github.com/monatis/trfaq) to see how it was finetuned and how it can be used in inference. The following code snippet is an excerpt from the inference at the repo.
## Usage
```python
questions = [
"Merhaba",
"Nasılsın?",
"Bireysel araç kiralama yapıyor musunuz?",
"Kurumsal araç kiralama yapıyor musunuz?"
]
answers = [
"Merhaba, size nasıl yardımcı olabilirim?",
"İyiyim, teşekkür ederim. Size nasıl yardımcı olabilirim?",
"Hayır, sadece Kurumsal Araç Kiralama operasyonları gerçekleştiriyoruz. Size başka nasıl yardımcı olabilirim?",
"Evet, kurumsal araç kiralama hizmetleri sağlıyoruz. Size nasıl yardımcı olabilirim?"
]
questions = ["<Q>" + q for q in questions]
answers = ["<A>" + a for a in answers]
def answer_faq(model, tokenizer, questions, answers, return_similarities=False):
q_len = len(questions)
tokens = tokenizer(questions + answers, padding=True, return_tensors='tf')
embs = model(**tokens)[0]
attention_masks = tf.cast(tokens['attention_mask'], tf.float32)
sample_length = tf.reduce_sum(attention_masks, axis=-1, keepdims=True)
masked_embs = embs * tf.expand_dims(attention_masks, axis=-1)
masked_embs = tf.reduce_sum(masked_embs, axis=1) / tf.cast(sample_length, tf.float32)
a = tf.math.l2_normalize(masked_embs[:q_len, :], axis=1)
b = tf.math.l2_normalize(masked_embs[q_len:, :], axis=1)
similarities = tf.matmul(a, b, transpose_b=True)
scores = tf.nn.softmax(similarities)
results = list(zip(answers, scores.numpy().squeeze().tolist()))
sorted_results = sorted(results, key=lambda x: x[1], reverse=True)
sorted_results = [{"answer": answer.replace("<A>", ""), "score": f"{score:.4f}"} for answer, score in sorted_results]
return sorted_results
for question in questions:
results = answer_faq(model, tokenizer, [question], answers)
print(question.replace("<Q>", ""))
print(results)
print("---------------------")
```
And the output is:
```shell
Merhaba
[{'answer': 'Merhaba, size nasıl yardımcı olabilirim?', 'score': '0.2931'}, {'answer': 'İyiyim, teşekkür ederim. Size nasıl yardımcı olabilirim?', 'score': '0.2751'}, {'answer': 'Hayır, sadece Kurumsal Araç Kiralama operasyonları gerçekleştiriyoruz. Size başka nasıl yardımcı olabilirim?', 'score': '0.2200'}, {'answer': 'Evet, kurumsal araç kiralama hizmetleri sağlıyoruz. Size nasıl yardımcı olabilirim?', 'score': '0.2118'}]
---------------------
Nasılsın?
[{'answer': 'İyiyim, teşekkür ederim. Size nasıl yardımcı olabilirim?', 'score': '0.2808'}, {'answer': 'Merhaba, size nasıl yardımcı olabilirim?', 'score': '0.2623'}, {'answer': 'Hayır, sadece Kurumsal Araç Kiralama operasyonları gerçekleştiriyoruz. Size başka nasıl yardımcı olabilirim?', 'score': '0.2320'}, {'answer': 'Evet, kurumsal araç kiralama hizmetleri sağlıyoruz. Size nasıl yardımcı olabilirim?', 'score': '0.2249'}]
---------------------
Bireysel araç kiralama yapıyor musunuz?
[{'answer': 'Hayır, sadece Kurumsal Araç Kiralama operasyonları gerçekleştiriyoruz. Size başka nasıl yardımcı olabilirim?', 'score': '0.2861'}, {'answer': 'Evet, kurumsal araç kiralama hizmetleri sağlıyoruz. Size nasıl yardımcı olabilirim?', 'score': '0.2768'}, {'answer': 'İyiyim, teşekkür ederim. Size nasıl yardımcı olabilirim?', 'score': '0.2215'}, {'answer': 'Merhaba, size nasıl yardımcı olabilirim?', 'score': '0.2156'}]
---------------------
Kurumsal araç kiralama yapıyor musunuz?
[{'answer': 'Evet, kurumsal araç kiralama hizmetleri sağlıyoruz. Size nasıl yardımcı olabilirim?', 'score': '0.3060'}, {'answer': 'Hayır, sadece Kurumsal Araç Kiralama operasyonları gerçekleştiriyoruz. Size başka nasıl yardımcı olabilirim?', 'score': '0.2929'}, {'answer': 'İyiyim, teşekkür ederim. Size nasıl yardımcı olabilirim?', 'score': '0.2066'}, {'answer': 'Merhaba, size nasıl yardımcı olabilirim?', 'score': '0.1945'}]
---------------------
``` |
microsoft/deberta-large-mnli | 7296194b9009373def4f7c5dad292651e4b5cf4e | 2021-05-21T20:07:51.000Z | [
"pytorch",
"deberta",
"text-classification",
"en",
"arxiv:2006.03654",
"transformers",
"deberta-v1",
"deberta-mnli",
"license:mit"
] | text-classification | false | microsoft | null | microsoft/deberta-large-mnli | 13,868 | 3 | transformers | 597 | ---
language: en
tags:
- deberta-v1
- deberta-mnli
tasks: mnli
thumbnail: https://huggingface.co/front/thumbnails/microsoft.png
license: mit
widget:
- text: "[CLS] I love you. [SEP] I like you. [SEP]"
---
## DeBERTa: Decoding-enhanced BERT with Disentangled Attention
[DeBERTa](https://arxiv.org/abs/2006.03654) improves the BERT and RoBERTa models using disentangled attention and enhanced mask decoder. It outperforms BERT and RoBERTa on majority of NLU tasks with 80GB training data.
Please check the [official repository](https://github.com/microsoft/DeBERTa) for more details and updates.
This is the DeBERTa large model fine-tuned with MNLI task.
#### Fine-tuning on NLU tasks
We present the dev results on SQuAD 1.1/2.0 and several GLUE benchmark tasks.
| Model | SQuAD 1.1 | SQuAD 2.0 | MNLI-m/mm | SST-2 | QNLI | CoLA | RTE | MRPC | QQP |STS-B |
|---------------------------|-----------|-----------|-------------|-------|------|------|--------|-------|-------|------|
| | F1/EM | F1/EM | Acc | Acc | Acc | MCC | Acc |Acc/F1 |Acc/F1 |P/S |
| BERT-Large | 90.9/84.1 | 81.8/79.0 | 86.6/- | 93.2 | 92.3 | 60.6 | 70.4 | 88.0/- | 91.3/- |90.0/- |
| RoBERTa-Large | 94.6/88.9 | 89.4/86.5 | 90.2/- | 96.4 | 93.9 | 68.0 | 86.6 | 90.9/- | 92.2/- |92.4/- |
| XLNet-Large | 95.1/89.7 | 90.6/87.9 | 90.8/- | 97.0 | 94.9 | 69.0 | 85.9 | 90.8/- | 92.3/- |92.5/- |
| [DeBERTa-Large](https://huggingface.co/microsoft/deberta-large)<sup>1</sup> | 95.5/90.1 | 90.7/88.0 | 91.3/91.1| 96.5|95.3| 69.5| 91.0| 92.6/94.6| 92.3/- |92.8/92.5 |
| [DeBERTa-XLarge](https://huggingface.co/microsoft/deberta-xlarge)<sup>1</sup> | -/- | -/- | 91.5/91.2| 97.0 | - | - | 93.1 | 92.1/94.3 | - |92.9/92.7|
| [DeBERTa-V2-XLarge](https://huggingface.co/microsoft/deberta-v2-xlarge)<sup>1</sup>|95.8/90.8| 91.4/88.9|91.7/91.6| **97.5**| 95.8|71.1|**93.9**|92.0/94.2|92.3/89.8|92.9/92.9|
|**[DeBERTa-V2-XXLarge](https://huggingface.co/microsoft/deberta-v2-xxlarge)<sup>1,2</sup>**|**96.1/91.4**|**92.2/89.7**|**91.7/91.9**|97.2|**96.0**|**72.0**| 93.5| **93.1/94.9**|**92.7/90.3** |**93.2/93.1** |
--------
#### Notes.
- <sup>1</sup> Following RoBERTa, for RTE, MRPC, STS-B, we fine-tune the tasks based on [DeBERTa-Large-MNLI](https://huggingface.co/microsoft/deberta-large-mnli), [DeBERTa-XLarge-MNLI](https://huggingface.co/microsoft/deberta-xlarge-mnli), [DeBERTa-V2-XLarge-MNLI](https://huggingface.co/microsoft/deberta-v2-xlarge-mnli), [DeBERTa-V2-XXLarge-MNLI](https://huggingface.co/microsoft/deberta-v2-xxlarge-mnli). The results of SST-2/QQP/QNLI/SQuADv2 will also be slightly improved when start from MNLI fine-tuned models, however, we only report the numbers fine-tuned from pretrained base models for those 4 tasks.
- <sup>2</sup> To try the **XXLarge** model with **[HF transformers](https://huggingface.co/transformers/main_classes/trainer.html)**, you need to specify **--sharded_ddp**
```bash
cd transformers/examples/text-classification/
export TASK_NAME=mrpc
python -m torch.distributed.launch --nproc_per_node=8 run_glue.py --model_name_or_path microsoft/deberta-v2-xxlarge \\
--task_name $TASK_NAME --do_train --do_eval --max_seq_length 128 --per_device_train_batch_size 4 \\
--learning_rate 3e-6 --num_train_epochs 3 --output_dir /tmp/$TASK_NAME/ --overwrite_output_dir --sharded_ddp --fp16
```
### Citation
If you find DeBERTa useful for your work, please cite the following paper:
``` latex
@inproceedings{
he2021deberta,
title={DEBERTA: DECODING-ENHANCED BERT WITH DISENTANGLED ATTENTION},
author={Pengcheng He and Xiaodong Liu and Jianfeng Gao and Weizhu Chen},
booktitle={International Conference on Learning Representations},
year={2021},
url={https://openreview.net/forum?id=XPZIaotutsD}
}
```
|
deepset/gbert-base-germandpr-question_encoder | dcbc13b5d22e49f58549a54dc7f30edec6153c39 | 2021-10-21T12:17:20.000Z | [
"pytorch",
"dpr",
"feature-extraction",
"de",
"dataset:deepset/germandpr",
"transformers",
"exbert",
"license:mit"
] | feature-extraction | false | deepset | null | deepset/gbert-base-germandpr-question_encoder | 13,861 | 4 | transformers | 598 | ---
language: de
datasets:
- deepset/germandpr
license: mit
thumbnail: https://thumb.tildacdn.com/tild3433-3637-4830-a533-353833613061/-/resize/720x/-/format/webp/germanquad.jpg
tags:
- exbert
---
## Overview
**Language model:** gbert-base-germandpr
**Language:** German
**Training data:** GermanDPR train set (~ 56MB)
**Eval data:** GermanDPR test set (~ 6MB)
**Infrastructure**: 4x V100 GPU
**Published**: Apr 26th, 2021
## Details
- We trained a dense passage retrieval model with two gbert-base models as encoders of questions and passages.
- The dataset is GermanDPR, a new, German language dataset, which we hand-annotated and published [online](https://deepset.ai/germanquad).
- It comprises 9275 question/answer pairs in the training set and 1025 pairs in the test set.
For each pair, there are one positive context and three hard negative contexts.
- As the basis of the training data, we used our hand-annotated GermanQuAD dataset as positive samples and generated hard negative samples from the latest German Wikipedia dump (6GB of raw txt files).
- The data dump was cleaned with tailored scripts, leading to 2.8 million indexed passages from German Wikipedia.
See https://deepset.ai/germanquad for more details and dataset download.
## Hyperparameters
```
batch_size = 40
n_epochs = 20
num_training_steps = 4640
num_warmup_steps = 460
max_seq_len = 32 tokens for question encoder and 300 tokens for passage encoder
learning_rate = 1e-6
lr_schedule = LinearWarmup
embeds_dropout_prob = 0.1
num_hard_negatives = 2
```
## Performance
During training, we monitored the in-batch average rank and the loss and evaluated different batch sizes, numbers of epochs, and number of hard negatives on a dev set split from the train set.
The dev split contained 1030 question/answer pairs.
Even without thorough hyperparameter tuning, we observed quite stable learning. Multiple restarts with different seeds produced quite similar results.
Note that the in-batch average rank is influenced by settings for batch size and number of hard negatives. A smaller number of hard negatives makes the task easier.
After fixing the hyperparameters we trained the model on the full GermanDPR train set.
We further evaluated the retrieval performance of the trained model on the full German Wikipedia with the GermanDPR test set as labels. To this end, we converted the GermanDPR test set to SQuAD format. The DPR model drastically outperforms the BM25 baseline with regard to recall@k.
## Usage
### In haystack
You can load the model in [haystack](https://github.com/deepset-ai/haystack/) as a retriever for doing QA at scale:
```python
retriever = DensePassageRetriever(
document_store=document_store,
query_embedding_model="deepset/gbert-base-germandpr-question_encoder"
passage_embedding_model="deepset/gbert-base-germandpr-ctx_encoder"
)
```
## Authors
- Timo Möller: `timo.moeller [at] deepset.ai`
- Julian Risch: `julian.risch [at] deepset.ai`
- Malte Pietsch: `malte.pietsch [at] deepset.ai`
## About us
We bring NLP to the industry via open source!
Our focus: Industry specific language models & large scale QA systems.
Some of our work:
- [German BERT (aka "bert-base-german-cased")](https://deepset.ai/german-bert)
- [GermanQuAD and GermanDPR datasets and models (aka "gelectra-base-germanquad", "gbert-base-germandpr")](https://deepset.ai/germanquad)
- [FARM](https://github.com/deepset-ai/FARM)
- [Haystack](https://github.com/deepset-ai/haystack/)
Get in touch:
[Twitter](https://twitter.com/deepset_ai) | [LinkedIn](https://www.linkedin.com/company/deepset-ai/) | [Website](https://deepset.ai)
By the way: [we're hiring!](http://www.deepset.ai/jobs)
|
shahrukhx01/bert-mini-finetune-question-detection | c9088454657626f5e370b3a1ec993fdcad81aaf6 | 2021-07-11T14:27:37.000Z | [
"pytorch",
"bert",
"text-classification",
"en",
"transformers",
"neural-search-query-classification",
"neural-search"
] | text-classification | false | shahrukhx01 | null | shahrukhx01/bert-mini-finetune-question-detection | 13,811 | 3 | transformers | 599 |
---
language: "en"
tags:
- neural-search-query-classification
- neural-search
widget:
- text: "keyword query."
---
# KEYWORD QUERY VS STATEMENT/QUESTION CLASSIFIER FOR NEURAL SEARCH
| Train Loss | Validation Acc.| Test Acc.|
| ------------- |:-------------: | -----: |
| 0.000806 | 0.99 | 0.997 |
```python
from transformers import AutoTokenizer, AutoModelForSequenceClassification
tokenizer = AutoTokenizer.from_pretrained("shahrukhx01/bert-mini-finetune-question-detection")
model = AutoModelForSequenceClassification.from_pretrained("shahrukhx01/bert-mini-finetune-question-detection")
```
Trained to add feature for classifying queries between Keyword Query or Question + Statement Query using classification in [Haystack](https://github.com/deepset-ai/haystack/issues/611)
Problem Statement:
One common challenge that we saw in deployments: We need to distinguish between real questions and keyword queries that come in. We only want to route questions to the Reader branch in order to maximize the accuracy of results and minimize computation efforts/costs.
Baseline:
https://www.kaggle.com/shahrukhkhan/question-v-statement-detection
Dataset:
https://www.kaggle.com/stefanondisponibile/quora-question-keyword-pairs
Kaggle Notebook:
https://www.kaggle.com/shahrukhkhan/question-vs-statement-classification-mini-bert/
|
Subsets and Splits
No community queries yet
The top public SQL queries from the community will appear here once available.