modelId
stringlengths 5
139
| author
stringlengths 2
42
| last_modified
timestamp[us, tz=UTC]date 2020-02-15 11:33:14
2025-07-14 06:27:53
| downloads
int64 0
223M
| likes
int64 0
11.7k
| library_name
stringclasses 519
values | tags
listlengths 1
4.05k
| pipeline_tag
stringclasses 55
values | createdAt
timestamp[us, tz=UTC]date 2022-03-02 23:29:04
2025-07-14 06:27:45
| card
stringlengths 11
1.01M
|
|---|---|---|---|---|---|---|---|---|---|
sd-concepts-library/towerplace | sd-concepts-library | 2022-09-20T03:35:13Z | 0 | 0 | null | [
"license:mit",
"region:us"
]
| null | 2022-09-20T01:37:57Z | ---
license: mit
---
### TowerPlace on Stable Diffusion
This is the `<TowerPlace>` concept taught to Stable Diffusion via Textual Inversion. You can load this concept into the [Stable Conceptualizer](https://colab.research.google.com/github/huggingface/notebooks/blob/main/diffusers/stable_conceptualizer_inference.ipynb) notebook. You can also train your own concepts and load them into the concept libraries using [this notebook](https://colab.research.google.com/github/huggingface/notebooks/blob/main/diffusers/sd_textual_inversion_training.ipynb).
Here is the new concept you will be able to use as a `style`:
|
Ricardmc99/q-Taxi-v3 | Ricardmc99 | 2022-09-20T02:54:49Z | 0 | 0 | null | [
"Taxi-v3",
"q-learning",
"reinforcement-learning",
"custom-implementation",
"model-index",
"region:us"
]
| reinforcement-learning | 2022-09-20T02:54:41Z | ---
tags:
- Taxi-v3
- q-learning
- reinforcement-learning
- custom-implementation
model-index:
- name: q-Taxi-v3
results:
- task:
type: reinforcement-learning
name: reinforcement-learning
dataset:
name: Taxi-v3
type: Taxi-v3
metrics:
- type: mean_reward
value: 7.50 +/- 2.72
name: mean_reward
verified: false
---
# **Q-Learning** Agent playing **Taxi-v3**
This is a trained model of a **Q-Learning** agent playing **Taxi-v3** .
## Usage
```python
model = load_from_hub(repo_id="Ricardmc99/q-Taxi-v3", filename="q-learning.pkl")
# Don't forget to check if you need to add additional attributes (is_slippery=False etc)
env = gym.make(model["env_id"])
evaluate_agent(env, model["max_steps"], model["n_eval_episodes"], model["qtable"], model["eval_seed"])
```
|
Ricardmc99/q-FrozenLake-v1-4x4-noSlippery | Ricardmc99 | 2022-09-20T02:45:34Z | 0 | 0 | null | [
"FrozenLake-v1-4x4-no_slippery",
"q-learning",
"reinforcement-learning",
"custom-implementation",
"model-index",
"region:us"
]
| reinforcement-learning | 2022-09-20T02:45:25Z | ---
tags:
- FrozenLake-v1-4x4-no_slippery
- q-learning
- reinforcement-learning
- custom-implementation
model-index:
- name: q-FrozenLake-v1-4x4-noSlippery
results:
- task:
type: reinforcement-learning
name: reinforcement-learning
dataset:
name: FrozenLake-v1-4x4-no_slippery
type: FrozenLake-v1-4x4-no_slippery
metrics:
- type: mean_reward
value: 0.00 +/- 0.00
name: mean_reward
verified: false
---
# **Q-Learning** Agent playing **FrozenLake-v1**
This is a trained model of a **Q-Learning** agent playing **FrozenLake-v1** .
## Usage
```python
model = load_from_hub(repo_id="Ricardmc99/q-FrozenLake-v1-4x4-noSlippery", filename="q-learning.pkl")
# Don't forget to check if you need to add additional attributes (is_slippery=False etc)
env = gym.make(model["env_id"])
evaluate_agent(env, model["max_steps"], model["n_eval_episodes"], model["qtable"], model["eval_seed"])
```
|
rram12/q-FrozenLake-v1-4x4-noSlippery | rram12 | 2022-09-20T02:26:17Z | 0 | 0 | null | [
"FrozenLake-v1-4x4-no_slippery",
"q-learning",
"reinforcement-learning",
"custom-implementation",
"model-index",
"region:us"
]
| reinforcement-learning | 2022-09-20T02:26:11Z | ---
tags:
- FrozenLake-v1-4x4-no_slippery
- q-learning
- reinforcement-learning
- custom-implementation
model-index:
- name: q-FrozenLake-v1-4x4-noSlippery
results:
- task:
type: reinforcement-learning
name: reinforcement-learning
dataset:
name: FrozenLake-v1-4x4-no_slippery
type: FrozenLake-v1-4x4-no_slippery
metrics:
- type: mean_reward
value: 1.00 +/- 0.00
name: mean_reward
verified: false
---
# **Q-Learning** Agent playing **FrozenLake-v1**
This is a trained model of a **Q-Learning** agent playing **FrozenLake-v1** .
## Usage
```python
model = load_from_hub(repo_id="rram12/q-FrozenLake-v1-4x4-noSlippery", filename="q-learning.pkl")
# Don't forget to check if you need to add additional attributes (is_slippery=False etc)
env = gym.make(model["env_id"])
evaluate_agent(env, model["max_steps"], model["n_eval_episodes"], model["qtable"], model["eval_seed"])
```
|
weirdguitarist/wav2vec2-base-stac-local | weirdguitarist | 2022-09-20T01:58:36Z | 19 | 0 | transformers | [
"transformers",
"pytorch",
"wav2vec2",
"automatic-speech-recognition",
"generated_from_trainer",
"license:apache-2.0",
"endpoints_compatible",
"region:us"
]
| automatic-speech-recognition | 2022-09-13T10:27:39Z | ---
license: apache-2.0
tags:
- generated_from_trainer
model-index:
- name: wav2vec2-base-stac-local
results: []
---
<!-- This model card has been generated automatically according to the information the Trainer had access to. You
should probably proofread and complete it, then remove this comment. -->
# wav2vec2-base-stac-local
This model is a fine-tuned version of [facebook/wav2vec2-xls-r-300m](https://huggingface.co/facebook/wav2vec2-xls-r-300m) on the None dataset.
It achieves the following results on the evaluation set:
- Loss: 1.9746
- Wer: 0.7828
- Cer: 0.3202
## Model description
More information needed
## Intended uses & limitations
More information needed
## Training and evaluation data
More information needed
## Training procedure
### Training hyperparameters
The following hyperparameters were used during training:
- learning_rate: 0.0001
- train_batch_size: 1
- eval_batch_size: 1
- seed: 42
- gradient_accumulation_steps: 2
- total_train_batch_size: 2
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
- lr_scheduler_type: linear
- lr_scheduler_warmup_steps: 500
- num_epochs: 30
### Training results
| Training Loss | Epoch | Step | Validation Loss | Wer | Cer |
|:-------------:|:-----:|:-----:|:---------------:|:------:|:------:|
| 2.0603 | 1.0 | 2369 | 2.1282 | 0.9517 | 0.5485 |
| 1.6155 | 2.0 | 4738 | 1.6196 | 0.9060 | 0.4565 |
| 1.3462 | 3.0 | 7107 | 1.4331 | 0.8379 | 0.3983 |
| 1.1819 | 4.0 | 9476 | 1.3872 | 0.8233 | 0.3717 |
| 1.0189 | 5.0 | 11845 | 1.4066 | 0.8328 | 0.3660 |
| 0.9026 | 6.0 | 14214 | 1.3502 | 0.8198 | 0.3508 |
| 0.777 | 7.0 | 16583 | 1.3016 | 0.7922 | 0.3433 |
| 0.7109 | 8.0 | 18952 | 1.2662 | 0.8302 | 0.3510 |
| 0.6766 | 9.0 | 21321 | 1.4321 | 0.8103 | 0.3368 |
| 0.6078 | 10.0 | 23690 | 1.3592 | 0.7871 | 0.3360 |
| 0.5958 | 11.0 | 26059 | 1.4389 | 0.7819 | 0.3397 |
| 0.5094 | 12.0 | 28428 | 1.3391 | 0.8017 | 0.3239 |
| 0.4567 | 13.0 | 30797 | 1.4718 | 0.8026 | 0.3347 |
| 0.4448 | 14.0 | 33166 | 1.7450 | 0.8043 | 0.3424 |
| 0.3976 | 15.0 | 35535 | 1.4581 | 0.7888 | 0.3283 |
| 0.3449 | 16.0 | 37904 | 1.5688 | 0.8078 | 0.3397 |
| 0.3046 | 17.0 | 40273 | 1.8630 | 0.8060 | 0.3448 |
| 0.2983 | 18.0 | 42642 | 1.8400 | 0.8190 | 0.3425 |
| 0.2728 | 19.0 | 45011 | 1.6726 | 0.8034 | 0.3280 |
| 0.2579 | 20.0 | 47380 | 1.6661 | 0.8138 | 0.3249 |
| 0.2169 | 21.0 | 49749 | 1.7389 | 0.8138 | 0.3277 |
| 0.2498 | 22.0 | 52118 | 1.7205 | 0.7948 | 0.3207 |
| 0.1831 | 23.0 | 54487 | 1.8641 | 0.8103 | 0.3229 |
| 0.1927 | 24.0 | 56856 | 1.8724 | 0.7784 | 0.3251 |
| 0.1649 | 25.0 | 59225 | 1.9187 | 0.7974 | 0.3277 |
| 0.1594 | 26.0 | 61594 | 1.9022 | 0.7828 | 0.3220 |
| 0.1338 | 27.0 | 63963 | 1.9303 | 0.7862 | 0.3212 |
| 0.1441 | 28.0 | 66332 | 1.9528 | 0.7845 | 0.3207 |
| 0.129 | 29.0 | 68701 | 1.9676 | 0.7819 | 0.3212 |
| 0.1169 | 30.0 | 71070 | 1.9746 | 0.7828 | 0.3202 |
### Framework versions
- Transformers 4.17.0
- Pytorch 1.8.1+cu102
- Datasets 1.18.3
- Tokenizers 0.12.1
|
rram12/ppo-lunarlander | rram12 | 2022-09-20T00:53:42Z | 0 | 0 | stable-baselines3 | [
"stable-baselines3",
"LunarLander-v2",
"deep-reinforcement-learning",
"reinforcement-learning",
"model-index",
"region:us"
]
| reinforcement-learning | 2022-09-20T00:52:50Z | ---
library_name: stable-baselines3
tags:
- LunarLander-v2
- deep-reinforcement-learning
- reinforcement-learning
- stable-baselines3
model-index:
- name: PPO
results:
- metrics:
- type: mean_reward
value: 283.68 +/- 20.00
name: mean_reward
task:
type: reinforcement-learning
name: reinforcement-learning
dataset:
name: LunarLander-v2
type: LunarLander-v2
---
# **PPO** Agent playing **LunarLander-v2**
This is a trained model of a **PPO** agent playing **LunarLander-v2**
using the [stable-baselines3 library](https://github.com/DLR-RM/stable-baselines3).
## Usage (with Stable-baselines3)
TODO: Add your code
```python
from stable_baselines3 import ...
from huggingface_sb3 import load_from_hub
...
```
|
sd-concepts-library/wojaks-now | sd-concepts-library | 2022-09-20T00:19:17Z | 0 | 4 | null | [
"license:mit",
"region:us"
]
| null | 2022-09-20T00:19:10Z | ---
license: mit
---
### wojaks-now on Stable Diffusion
This is the `<red-wojak>` concept taught to Stable Diffusion via Textual Inversion. You can load this concept into the [Stable Conceptualizer](https://colab.research.google.com/github/huggingface/notebooks/blob/main/diffusers/stable_conceptualizer_inference.ipynb) notebook. You can also train your own concepts and load them into the concept libraries using [this notebook](https://colab.research.google.com/github/huggingface/notebooks/blob/main/diffusers/sd_textual_inversion_training.ipynb).
Here is the new concept you will be able to use as an `object`:
|
sd-concepts-library/shiny-polyman | sd-concepts-library | 2022-09-19T23:57:20Z | 0 | 0 | null | [
"license:mit",
"region:us"
]
| null | 2022-09-19T23:57:09Z | ---
license: mit
---
### Shiny polyman on Stable Diffusion
This is the `<shiny-polyman>` concept taught to Stable Diffusion via Textual Inversion. You can load this concept into the [Stable Conceptualizer](https://colab.research.google.com/github/huggingface/notebooks/blob/main/diffusers/stable_conceptualizer_inference.ipynb) notebook. You can also train your own concepts and load them into the concept libraries using [this notebook](https://colab.research.google.com/github/huggingface/notebooks/blob/main/diffusers/sd_textual_inversion_training.ipynb).
Here is the new concept you will be able to use as an `object`:
|
SandraB/mt5-small-mlsum_training_sample | SandraB | 2022-09-19T23:36:24Z | 111 | 0 | transformers | [
"transformers",
"pytorch",
"tensorboard",
"mt5",
"text2text-generation",
"summarization",
"generated_from_trainer",
"dataset:mlsum",
"license:apache-2.0",
"model-index",
"autotrain_compatible",
"endpoints_compatible",
"region:us"
]
| summarization | 2022-09-19T13:17:26Z | ---
license: apache-2.0
tags:
- summarization
- generated_from_trainer
datasets:
- mlsum
metrics:
- rouge
model-index:
- name: mt5-small-mlsum_training_sample
results:
- task:
name: Sequence-to-sequence Language Modeling
type: text2text-generation
dataset:
name: mlsum
type: mlsum
config: de
split: train
args: de
metrics:
- name: Rouge1
type: rouge
value: 28.2078
---
<!-- This model card has been generated automatically according to the information the Trainer had access to. You
should probably proofread and complete it, then remove this comment. -->
# mt5-small-mlsum_training_sample
This model is a fine-tuned version of [google/mt5-small](https://huggingface.co/google/mt5-small) on the mlsum dataset.
It achieves the following results on the evaluation set:
- Loss: 1.9727
- Rouge1: 28.2078
- Rouge2: 19.0712
- Rougel: 26.2267
- Rougelsum: 26.9462
## Model description
More information needed
## Intended uses & limitations
More information needed
## Training and evaluation data
More information needed
## Training procedure
### Training hyperparameters
The following hyperparameters were used during training:
- learning_rate: 0.001
- train_batch_size: 8
- eval_batch_size: 8
- seed: 42
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
- lr_scheduler_type: linear
- num_epochs: 8
### Training results
| Training Loss | Epoch | Step | Validation Loss | Rouge1 | Rouge2 | Rougel | Rougelsum |
|:-------------:|:-----:|:-----:|:---------------:|:-------:|:-------:|:-------:|:---------:|
| 1.3193 | 1.0 | 6875 | 2.1352 | 25.8941 | 17.4672 | 24.2858 | 24.924 |
| 1.2413 | 2.0 | 13750 | 2.0528 | 26.6221 | 18.1166 | 24.8233 | 25.5111 |
| 1.1844 | 3.0 | 20625 | 1.9783 | 27.0518 | 18.3457 | 25.2288 | 25.8919 |
| 1.0403 | 4.0 | 27500 | 1.9487 | 27.8154 | 18.9701 | 25.9435 | 26.6578 |
| 0.9582 | 5.0 | 34375 | 1.9374 | 27.6863 | 18.7723 | 25.7667 | 26.4694 |
| 0.8992 | 6.0 | 41250 | 1.9353 | 27.8959 | 18.919 | 26.0434 | 26.7262 |
| 0.8109 | 7.0 | 48125 | 1.9492 | 28.0644 | 18.8873 | 26.0628 | 26.757 |
| 0.7705 | 8.0 | 55000 | 1.9727 | 28.2078 | 19.0712 | 26.2267 | 26.9462 |
### Framework versions
- Transformers 4.22.1
- Pytorch 1.12.1+cu113
- Datasets 2.4.0
- Tokenizers 0.12.1
|
Isaacp/xlm-roberta-base-finetuned-panx-de-fr | Isaacp | 2022-09-19T23:18:16Z | 103 | 0 | transformers | [
"transformers",
"pytorch",
"xlm-roberta",
"token-classification",
"generated_from_trainer",
"license:mit",
"autotrain_compatible",
"endpoints_compatible",
"region:us"
]
| token-classification | 2022-09-15T21:55:15Z | ---
license: mit
tags:
- generated_from_trainer
metrics:
- f1
model-index:
- name: xlm-roberta-base-finetuned-panx-de-fr
results: []
---
<!-- This model card has been generated automatically according to the information the Trainer had access to. You
should probably proofread and complete it, then remove this comment. -->
# xlm-roberta-base-finetuned-panx-de-fr
This model is a fine-tuned version of [xlm-roberta-base](https://huggingface.co/xlm-roberta-base) on the None dataset.
It achieves the following results on the evaluation set:
- Loss: 0.1637
- F1: 0.8599
## Model description
More information needed
## Intended uses & limitations
More information needed
## Training and evaluation data
More information needed
## Training procedure
### Training hyperparameters
The following hyperparameters were used during training:
- learning_rate: 5e-05
- train_batch_size: 24
- eval_batch_size: 24
- seed: 42
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
- lr_scheduler_type: linear
- num_epochs: 3
### Training results
| Training Loss | Epoch | Step | Validation Loss | F1 |
|:-------------:|:-----:|:----:|:---------------:|:------:|
| 0.2897 | 1.0 | 715 | 0.1759 | 0.8369 |
| 0.1462 | 2.0 | 1430 | 0.1587 | 0.8506 |
| 0.0931 | 3.0 | 2145 | 0.1637 | 0.8599 |
### Framework versions
- Transformers 4.11.3
- Pytorch 1.12.1+cu113
- Datasets 1.16.1
- Tokenizers 0.10.3
|
sd-concepts-library/kawaii-colors | sd-concepts-library | 2022-09-19T23:08:01Z | 0 | 26 | null | [
"license:mit",
"region:us"
]
| null | 2022-09-15T20:07:40Z | ---
license: mit
---
### Kawaii Colors on Stable Diffusion
This is the `<kawaii-colors-style>` concept taught to Stable Diffusion via Textual Inversion. You can load this concept into the [Stable Conceptualizer](https://colab.research.google.com/github/huggingface/notebooks/blob/main/diffusers/stable_conceptualizer_inference.ipynb) notebook. You can also train your own concepts and load them into the concept libraries using [this notebook](https://colab.research.google.com/github/huggingface/notebooks/blob/main/diffusers/sd_textual_inversion_training.ipynb).
Here is the new concept you will be able to use as a `style`:
|
sd-concepts-library/overprettified | sd-concepts-library | 2022-09-19T22:12:07Z | 0 | 0 | null | [
"license:mit",
"region:us"
]
| null | 2022-09-19T22:12:02Z | ---
license: mit
---
### overprettified on Stable Diffusion
This is the `<overprettified>` concept taught to Stable Diffusion via Textual Inversion. You can load this concept into the [Stable Conceptualizer](https://colab.research.google.com/github/huggingface/notebooks/blob/main/diffusers/stable_conceptualizer_inference.ipynb) notebook. You can also train your own concepts and load them into the concept libraries using [this notebook](https://colab.research.google.com/github/huggingface/notebooks/blob/main/diffusers/sd_textual_inversion_training.ipynb).
Here is the new concept you will be able to use as an `object`:
|
Isaacp/xlm-roberta-base-finetuned-panx-de | Isaacp | 2022-09-19T21:44:18Z | 103 | 0 | transformers | [
"transformers",
"pytorch",
"tensorboard",
"xlm-roberta",
"token-classification",
"generated_from_trainer",
"dataset:xtreme",
"license:mit",
"model-index",
"autotrain_compatible",
"endpoints_compatible",
"region:us"
]
| token-classification | 2022-09-15T21:21:23Z | ---
license: mit
tags:
- generated_from_trainer
datasets:
- xtreme
metrics:
- f1
model-index:
- name: xlm-roberta-base-finetuned-panx-de
results:
- task:
name: Token Classification
type: token-classification
dataset:
name: xtreme
type: xtreme
args: PAN-X.de
metrics:
- name: F1
type: f1
value: 0.861006932605978
---
<!-- This model card has been generated automatically according to the information the Trainer had access to. You
should probably proofread and complete it, then remove this comment. -->
# xlm-roberta-base-finetuned-panx-de
This model is a fine-tuned version of [xlm-roberta-base](https://huggingface.co/xlm-roberta-base) on the xtreme dataset.
It achieves the following results on the evaluation set:
- Loss: 0.1397
- F1: 0.8610
## Model description
More information needed
## Intended uses & limitations
More information needed
## Training and evaluation data
More information needed
## Training procedure
### Training hyperparameters
The following hyperparameters were used during training:
- learning_rate: 5e-05
- train_batch_size: 24
- eval_batch_size: 24
- seed: 42
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
- lr_scheduler_type: linear
- num_epochs: 3
### Training results
| Training Loss | Epoch | Step | Validation Loss | F1 |
|:-------------:|:-----:|:----:|:---------------:|:------:|
| 0.2663 | 1.0 | 525 | 0.1628 | 0.8273 |
| 0.1338 | 2.0 | 1050 | 0.1457 | 0.8396 |
| 0.0844 | 3.0 | 1575 | 0.1397 | 0.8610 |
### Framework versions
- Transformers 4.11.3
- Pytorch 1.12.1+cu113
- Datasets 1.16.1
- Tokenizers 0.10.3
|
research-backup/roberta-large-semeval2012-average-no-mask-prompt-d-nce-conceptnet-validated | research-backup | 2022-09-19T21:43:02Z | 105 | 0 | transformers | [
"transformers",
"pytorch",
"roberta",
"feature-extraction",
"dataset:relbert/semeval2012_relational_similarity",
"model-index",
"text-embeddings-inference",
"endpoints_compatible",
"region:us"
]
| feature-extraction | 2022-09-17T04:16:44Z | ---
datasets:
- relbert/semeval2012_relational_similarity
model-index:
- name: relbert/roberta-large-semeval2012-average-no-mask-prompt-d-nce-conceptnet-validated
results:
- task:
name: Relation Mapping
type: sorting-task
dataset:
name: Relation Mapping
args: relbert/relation_mapping
type: relation-mapping
metrics:
- name: Accuracy
type: accuracy
value: 0.8758333333333334
- task:
name: Analogy Questions (SAT full)
type: multiple-choice-qa
dataset:
name: SAT full
args: relbert/analogy_questions
type: analogy-questions
metrics:
- name: Accuracy
type: accuracy
value: 0.6978609625668449
- task:
name: Analogy Questions (SAT)
type: multiple-choice-qa
dataset:
name: SAT
args: relbert/analogy_questions
type: analogy-questions
metrics:
- name: Accuracy
type: accuracy
value: 0.7002967359050445
- task:
name: Analogy Questions (BATS)
type: multiple-choice-qa
dataset:
name: BATS
args: relbert/analogy_questions
type: analogy-questions
metrics:
- name: Accuracy
type: accuracy
value: 0.8171206225680934
- task:
name: Analogy Questions (Google)
type: multiple-choice-qa
dataset:
name: Google
args: relbert/analogy_questions
type: analogy-questions
metrics:
- name: Accuracy
type: accuracy
value: 0.954
- task:
name: Analogy Questions (U2)
type: multiple-choice-qa
dataset:
name: U2
args: relbert/analogy_questions
type: analogy-questions
metrics:
- name: Accuracy
type: accuracy
value: 0.7017543859649122
- task:
name: Analogy Questions (U4)
type: multiple-choice-qa
dataset:
name: U4
args: relbert/analogy_questions
type: analogy-questions
metrics:
- name: Accuracy
type: accuracy
value: 0.6782407407407407
- task:
name: Lexical Relation Classification (BLESS)
type: classification
dataset:
name: BLESS
args: relbert/lexical_relation_classification
type: relation-classification
metrics:
- name: F1
type: f1
value: 0.9228567123700467
- name: F1 (macro)
type: f1_macro
value: 0.9181948193231166
- task:
name: Lexical Relation Classification (CogALexV)
type: classification
dataset:
name: CogALexV
args: relbert/lexical_relation_classification
type: relation-classification
metrics:
- name: F1
type: f1
value: 0.862206572769953
- name: F1 (macro)
type: f1_macro
value: 0.709511061556657
- task:
name: Lexical Relation Classification (EVALution)
type: classification
dataset:
name: BLESS
args: relbert/lexical_relation_classification
type: relation-classification
metrics:
- name: F1
type: f1
value: 0.6782231852654388
- name: F1 (macro)
type: f1_macro
value: 0.6681046248112636
- task:
name: Lexical Relation Classification (K&H+N)
type: classification
dataset:
name: K&H+N
args: relbert/lexical_relation_classification
type: relation-classification
metrics:
- name: F1
type: f1
value: 0.9577797871600473
- name: F1 (macro)
type: f1_macro
value: 0.8797850563531711
- task:
name: Lexical Relation Classification (ROOT09)
type: classification
dataset:
name: ROOT09
args: relbert/lexical_relation_classification
type: relation-classification
metrics:
- name: F1
type: f1
value: 0.9034785333751175
- name: F1 (macro)
type: f1_macro
value: 0.9016285634954788
---
# relbert/roberta-large-semeval2012-average-no-mask-prompt-d-nce-conceptnet-validated
RelBERT fine-tuned from [roberta-large](https://huggingface.co/roberta-large) on
[relbert/semeval2012_relational_similarity](https://huggingface.co/datasets/relbert/semeval2012_relational_similarity).
Fine-tuning is done via [RelBERT](https://github.com/asahi417/relbert) library (see the repository for more detail).
It achieves the following results on the relation understanding tasks:
- Analogy Question ([dataset](https://huggingface.co/datasets/relbert/analogy_questions), [full result](https://huggingface.co/relbert/roberta-large-semeval2012-average-no-mask-prompt-d-nce-conceptnet-validated/raw/main/analogy.json)):
- Accuracy on SAT (full): 0.6978609625668449
- Accuracy on SAT: 0.7002967359050445
- Accuracy on BATS: 0.8171206225680934
- Accuracy on U2: 0.7017543859649122
- Accuracy on U4: 0.6782407407407407
- Accuracy on Google: 0.954
- Lexical Relation Classification ([dataset](https://huggingface.co/datasets/relbert/lexical_relation_classification), [full result](https://huggingface.co/relbert/roberta-large-semeval2012-average-no-mask-prompt-d-nce-conceptnet-validated/raw/main/classification.json)):
- Micro F1 score on BLESS: 0.9228567123700467
- Micro F1 score on CogALexV: 0.862206572769953
- Micro F1 score on EVALution: 0.6782231852654388
- Micro F1 score on K&H+N: 0.9577797871600473
- Micro F1 score on ROOT09: 0.9034785333751175
- Relation Mapping ([dataset](https://huggingface.co/datasets/relbert/relation_mapping), [full result](https://huggingface.co/relbert/roberta-large-semeval2012-average-no-mask-prompt-d-nce-conceptnet-validated/raw/main/relation_mapping.json)):
- Accuracy on Relation Mapping: 0.8758333333333334
### Usage
This model can be used through the [relbert library](https://github.com/asahi417/relbert). Install the library via pip
```shell
pip install relbert
```
and activate model as below.
```python
from relbert import RelBERT
model = RelBERT("relbert/roberta-large-semeval2012-average-no-mask-prompt-d-nce-conceptnet-validated")
vector = model.get_embedding(['Tokyo', 'Japan']) # shape of (1024, )
```
### Training hyperparameters
The following hyperparameters were used during training:
- model: roberta-large
- max_length: 64
- mode: average_no_mask
- data: relbert/semeval2012_relational_similarity
- split: train
- data_eval: relbert/conceptnet_high_confidence
- split_eval: full
- template_mode: manual
- template: I wasn’t aware of this relationship, but I just read in the encyclopedia that <subj> is the <mask> of <obj>
- loss_function: nce_logout
- classification_loss: False
- temperature_nce_constant: 0.05
- temperature_nce_rank: {'min': 0.01, 'max': 0.05, 'type': 'linear'}
- epoch: 30
- batch: 128
- lr: 5e-06
- lr_decay: False
- lr_warmup: 1
- weight_decay: 0
- random_seed: 0
- exclude_relation: None
- exclude_relation_eval: None
- n_sample: 640
- gradient_accumulation: 8
The full configuration can be found at [fine-tuning parameter file](https://huggingface.co/relbert/roberta-large-semeval2012-average-no-mask-prompt-d-nce-conceptnet-validated/raw/main/trainer_config.json).
### Reference
If you use any resource from RelBERT, please consider to cite our [paper](https://aclanthology.org/2021.eacl-demos.7/).
```
@inproceedings{ushio-etal-2021-distilling-relation-embeddings,
title = "{D}istilling {R}elation {E}mbeddings from {P}re-trained {L}anguage {M}odels",
author = "Ushio, Asahi and
Schockaert, Steven and
Camacho-Collados, Jose",
booktitle = "EMNLP 2021",
year = "2021",
address = "Online",
publisher = "Association for Computational Linguistics",
}
```
|
research-backup/roberta-large-semeval2012-average-prompt-e-nce-conceptnet-validated | research-backup | 2022-09-19T21:31:09Z | 106 | 0 | transformers | [
"transformers",
"pytorch",
"roberta",
"feature-extraction",
"dataset:relbert/semeval2012_relational_similarity",
"model-index",
"text-embeddings-inference",
"endpoints_compatible",
"region:us"
]
| feature-extraction | 2022-09-15T21:58:57Z | ---
datasets:
- relbert/semeval2012_relational_similarity
model-index:
- name: relbert/roberta-large-semeval2012-average-prompt-e-nce-conceptnet-validated
results:
- task:
name: Relation Mapping
type: sorting-task
dataset:
name: Relation Mapping
args: relbert/relation_mapping
type: relation-mapping
metrics:
- name: Accuracy
type: accuracy
value: 0.8490079365079365
- task:
name: Analogy Questions (SAT full)
type: multiple-choice-qa
dataset:
name: SAT full
args: relbert/analogy_questions
type: analogy-questions
metrics:
- name: Accuracy
type: accuracy
value: 0.6256684491978609
- task:
name: Analogy Questions (SAT)
type: multiple-choice-qa
dataset:
name: SAT
args: relbert/analogy_questions
type: analogy-questions
metrics:
- name: Accuracy
type: accuracy
value: 0.6379821958456974
- task:
name: Analogy Questions (BATS)
type: multiple-choice-qa
dataset:
name: BATS
args: relbert/analogy_questions
type: analogy-questions
metrics:
- name: Accuracy
type: accuracy
value: 0.7354085603112841
- task:
name: Analogy Questions (Google)
type: multiple-choice-qa
dataset:
name: Google
args: relbert/analogy_questions
type: analogy-questions
metrics:
- name: Accuracy
type: accuracy
value: 0.882
- task:
name: Analogy Questions (U2)
type: multiple-choice-qa
dataset:
name: U2
args: relbert/analogy_questions
type: analogy-questions
metrics:
- name: Accuracy
type: accuracy
value: 0.6403508771929824
- task:
name: Analogy Questions (U4)
type: multiple-choice-qa
dataset:
name: U4
args: relbert/analogy_questions
type: analogy-questions
metrics:
- name: Accuracy
type: accuracy
value: 0.6273148148148148
- task:
name: Lexical Relation Classification (BLESS)
type: classification
dataset:
name: BLESS
args: relbert/lexical_relation_classification
type: relation-classification
metrics:
- name: F1
type: f1
value: 0.926020792526744
- name: F1 (macro)
type: f1_macro
value: 0.9203622322435366
- task:
name: Lexical Relation Classification (CogALexV)
type: classification
dataset:
name: CogALexV
args: relbert/lexical_relation_classification
type: relation-classification
metrics:
- name: F1
type: f1
value: 0.8514084507042253
- name: F1 (macro)
type: f1_macro
value: 0.6818973595530592
- task:
name: Lexical Relation Classification (EVALution)
type: classification
dataset:
name: BLESS
args: relbert/lexical_relation_classification
type: relation-classification
metrics:
- name: F1
type: f1
value: 0.6868905742145178
- name: F1 (macro)
type: f1_macro
value: 0.6727836285482613
- task:
name: Lexical Relation Classification (K&H+N)
type: classification
dataset:
name: K&H+N
args: relbert/lexical_relation_classification
type: relation-classification
metrics:
- name: F1
type: f1
value: 0.955762676497183
- name: F1 (macro)
type: f1_macro
value: 0.8743237384020923
- task:
name: Lexical Relation Classification (ROOT09)
type: classification
dataset:
name: ROOT09
args: relbert/lexical_relation_classification
type: relation-classification
metrics:
- name: F1
type: f1
value: 0.898464431212786
- name: F1 (macro)
type: f1_macro
value: 0.8965223263877973
---
# relbert/roberta-large-semeval2012-average-prompt-e-nce-conceptnet-validated
RelBERT fine-tuned from [roberta-large](https://huggingface.co/roberta-large) on
[relbert/semeval2012_relational_similarity](https://huggingface.co/datasets/relbert/semeval2012_relational_similarity).
Fine-tuning is done via [RelBERT](https://github.com/asahi417/relbert) library (see the repository for more detail).
It achieves the following results on the relation understanding tasks:
- Analogy Question ([dataset](https://huggingface.co/datasets/relbert/analogy_questions), [full result](https://huggingface.co/relbert/roberta-large-semeval2012-average-prompt-e-nce-conceptnet-validated/raw/main/analogy.json)):
- Accuracy on SAT (full): 0.6256684491978609
- Accuracy on SAT: 0.6379821958456974
- Accuracy on BATS: 0.7354085603112841
- Accuracy on U2: 0.6403508771929824
- Accuracy on U4: 0.6273148148148148
- Accuracy on Google: 0.882
- Lexical Relation Classification ([dataset](https://huggingface.co/datasets/relbert/lexical_relation_classification), [full result](https://huggingface.co/relbert/roberta-large-semeval2012-average-prompt-e-nce-conceptnet-validated/raw/main/classification.json)):
- Micro F1 score on BLESS: 0.926020792526744
- Micro F1 score on CogALexV: 0.8514084507042253
- Micro F1 score on EVALution: 0.6868905742145178
- Micro F1 score on K&H+N: 0.955762676497183
- Micro F1 score on ROOT09: 0.898464431212786
- Relation Mapping ([dataset](https://huggingface.co/datasets/relbert/relation_mapping), [full result](https://huggingface.co/relbert/roberta-large-semeval2012-average-prompt-e-nce-conceptnet-validated/raw/main/relation_mapping.json)):
- Accuracy on Relation Mapping: 0.8490079365079365
### Usage
This model can be used through the [relbert library](https://github.com/asahi417/relbert). Install the library via pip
```shell
pip install relbert
```
and activate model as below.
```python
from relbert import RelBERT
model = RelBERT("relbert/roberta-large-semeval2012-average-prompt-e-nce-conceptnet-validated")
vector = model.get_embedding(['Tokyo', 'Japan']) # shape of (1024, )
```
### Training hyperparameters
The following hyperparameters were used during training:
- model: roberta-large
- max_length: 64
- mode: average
- data: relbert/semeval2012_relational_similarity
- split: train
- data_eval: relbert/conceptnet_high_confidence
- split_eval: full
- template_mode: manual
- template: I wasn’t aware of this relationship, but I just read in the encyclopedia that <obj> is <subj>’s <mask>
- loss_function: nce_logout
- classification_loss: False
- temperature_nce_constant: 0.05
- temperature_nce_rank: {'min': 0.01, 'max': 0.05, 'type': 'linear'}
- epoch: 29
- batch: 128
- lr: 5e-06
- lr_decay: False
- lr_warmup: 1
- weight_decay: 0
- random_seed: 0
- exclude_relation: None
- exclude_relation_eval: None
- n_sample: 640
- gradient_accumulation: 8
The full configuration can be found at [fine-tuning parameter file](https://huggingface.co/relbert/roberta-large-semeval2012-average-prompt-e-nce-conceptnet-validated/raw/main/trainer_config.json).
### Reference
If you use any resource from RelBERT, please consider to cite our [paper](https://aclanthology.org/2021.eacl-demos.7/).
```
@inproceedings{ushio-etal-2021-distilling-relation-embeddings,
title = "{D}istilling {R}elation {E}mbeddings from {P}re-trained {L}anguage {M}odels",
author = "Ushio, Asahi and
Schockaert, Steven and
Camacho-Collados, Jose",
booktitle = "EMNLP 2021",
year = "2021",
address = "Online",
publisher = "Association for Computational Linguistics",
}
```
|
research-backup/roberta-large-semeval2012-average-prompt-b-nce-conceptnet-validated | research-backup | 2022-09-19T21:20:17Z | 104 | 0 | transformers | [
"transformers",
"pytorch",
"roberta",
"feature-extraction",
"dataset:relbert/semeval2012_relational_similarity",
"model-index",
"text-embeddings-inference",
"endpoints_compatible",
"region:us"
]
| feature-extraction | 2022-09-14T23:37:25Z | ---
datasets:
- relbert/semeval2012_relational_similarity
model-index:
- name: relbert/roberta-large-semeval2012-average-prompt-b-nce-conceptnet-validated
results:
- task:
name: Relation Mapping
type: sorting-task
dataset:
name: Relation Mapping
args: relbert/relation_mapping
type: relation-mapping
metrics:
- name: Accuracy
type: accuracy
value: 0.888095238095238
- task:
name: Analogy Questions (SAT full)
type: multiple-choice-qa
dataset:
name: SAT full
args: relbert/analogy_questions
type: analogy-questions
metrics:
- name: Accuracy
type: accuracy
value: 0.6283422459893048
- task:
name: Analogy Questions (SAT)
type: multiple-choice-qa
dataset:
name: SAT
args: relbert/analogy_questions
type: analogy-questions
metrics:
- name: Accuracy
type: accuracy
value: 0.629080118694362
- task:
name: Analogy Questions (BATS)
type: multiple-choice-qa
dataset:
name: BATS
args: relbert/analogy_questions
type: analogy-questions
metrics:
- name: Accuracy
type: accuracy
value: 0.7959977765425236
- task:
name: Analogy Questions (Google)
type: multiple-choice-qa
dataset:
name: Google
args: relbert/analogy_questions
type: analogy-questions
metrics:
- name: Accuracy
type: accuracy
value: 0.92
- task:
name: Analogy Questions (U2)
type: multiple-choice-qa
dataset:
name: U2
args: relbert/analogy_questions
type: analogy-questions
metrics:
- name: Accuracy
type: accuracy
value: 0.5701754385964912
- task:
name: Analogy Questions (U4)
type: multiple-choice-qa
dataset:
name: U4
args: relbert/analogy_questions
type: analogy-questions
metrics:
- name: Accuracy
type: accuracy
value: 0.6134259259259259
- task:
name: Lexical Relation Classification (BLESS)
type: classification
dataset:
name: BLESS
args: relbert/lexical_relation_classification
type: relation-classification
metrics:
- name: F1
type: f1
value: 0.9172819044749133
- name: F1 (macro)
type: f1_macro
value: 0.9134777544987239
- task:
name: Lexical Relation Classification (CogALexV)
type: classification
dataset:
name: CogALexV
args: relbert/lexical_relation_classification
type: relation-classification
metrics:
- name: F1
type: f1
value: 0.8516431924882629
- name: F1 (macro)
type: f1_macro
value: 0.6909836328773065
- task:
name: Lexical Relation Classification (EVALution)
type: classification
dataset:
name: BLESS
args: relbert/lexical_relation_classification
type: relation-classification
metrics:
- name: F1
type: f1
value: 0.6738894907908992
- name: F1 (macro)
type: f1_macro
value: 0.6623942225782876
- task:
name: Lexical Relation Classification (K&H+N)
type: classification
dataset:
name: K&H+N
args: relbert/lexical_relation_classification
type: relation-classification
metrics:
- name: F1
type: f1
value: 0.9517284551714544
- name: F1 (macro)
type: f1_macro
value: 0.8593035416288995
- task:
name: Lexical Relation Classification (ROOT09)
type: classification
dataset:
name: ROOT09
args: relbert/lexical_relation_classification
type: relation-classification
metrics:
- name: F1
type: f1
value: 0.9000313381385145
- name: F1 (macro)
type: f1_macro
value: 0.8976663712913519
---
# relbert/roberta-large-semeval2012-average-prompt-b-nce-conceptnet-validated
RelBERT fine-tuned from [roberta-large](https://huggingface.co/roberta-large) on
[relbert/semeval2012_relational_similarity](https://huggingface.co/datasets/relbert/semeval2012_relational_similarity).
Fine-tuning is done via [RelBERT](https://github.com/asahi417/relbert) library (see the repository for more detail).
It achieves the following results on the relation understanding tasks:
- Analogy Question ([dataset](https://huggingface.co/datasets/relbert/analogy_questions), [full result](https://huggingface.co/relbert/roberta-large-semeval2012-average-prompt-b-nce-conceptnet-validated/raw/main/analogy.json)):
- Accuracy on SAT (full): 0.6283422459893048
- Accuracy on SAT: 0.629080118694362
- Accuracy on BATS: 0.7959977765425236
- Accuracy on U2: 0.5701754385964912
- Accuracy on U4: 0.6134259259259259
- Accuracy on Google: 0.92
- Lexical Relation Classification ([dataset](https://huggingface.co/datasets/relbert/lexical_relation_classification), [full result](https://huggingface.co/relbert/roberta-large-semeval2012-average-prompt-b-nce-conceptnet-validated/raw/main/classification.json)):
- Micro F1 score on BLESS: 0.9172819044749133
- Micro F1 score on CogALexV: 0.8516431924882629
- Micro F1 score on EVALution: 0.6738894907908992
- Micro F1 score on K&H+N: 0.9517284551714544
- Micro F1 score on ROOT09: 0.9000313381385145
- Relation Mapping ([dataset](https://huggingface.co/datasets/relbert/relation_mapping), [full result](https://huggingface.co/relbert/roberta-large-semeval2012-average-prompt-b-nce-conceptnet-validated/raw/main/relation_mapping.json)):
- Accuracy on Relation Mapping: 0.888095238095238
### Usage
This model can be used through the [relbert library](https://github.com/asahi417/relbert). Install the library via pip
```shell
pip install relbert
```
and activate model as below.
```python
from relbert import RelBERT
model = RelBERT("relbert/roberta-large-semeval2012-average-prompt-b-nce-conceptnet-validated")
vector = model.get_embedding(['Tokyo', 'Japan']) # shape of (1024, )
```
### Training hyperparameters
The following hyperparameters were used during training:
- model: roberta-large
- max_length: 64
- mode: average
- data: relbert/semeval2012_relational_similarity
- split: train
- data_eval: relbert/conceptnet_high_confidence
- split_eval: full
- template_mode: manual
- template: Today, I finally discovered the relation between <subj> and <obj> : <obj> is <subj>'s <mask>
- loss_function: nce_logout
- classification_loss: False
- temperature_nce_constant: 0.05
- temperature_nce_rank: {'min': 0.01, 'max': 0.05, 'type': 'linear'}
- epoch: 30
- batch: 128
- lr: 5e-06
- lr_decay: False
- lr_warmup: 1
- weight_decay: 0
- random_seed: 0
- exclude_relation: None
- exclude_relation_eval: None
- n_sample: 640
- gradient_accumulation: 8
The full configuration can be found at [fine-tuning parameter file](https://huggingface.co/relbert/roberta-large-semeval2012-average-prompt-b-nce-conceptnet-validated/raw/main/trainer_config.json).
### Reference
If you use any resource from RelBERT, please consider to cite our [paper](https://aclanthology.org/2021.eacl-demos.7/).
```
@inproceedings{ushio-etal-2021-distilling-relation-embeddings,
title = "{D}istilling {R}elation {E}mbeddings from {P}re-trained {L}anguage {M}odels",
author = "Ushio, Asahi and
Schockaert, Steven and
Camacho-Collados, Jose",
booktitle = "EMNLP 2021",
year = "2021",
address = "Online",
publisher = "Association for Computational Linguistics",
}
```
|
research-backup/roberta-large-semeval2012-mask-prompt-d-nce-conceptnet-validated | research-backup | 2022-09-19T21:08:47Z | 105 | 0 | transformers | [
"transformers",
"pytorch",
"roberta",
"feature-extraction",
"dataset:relbert/semeval2012_relational_similarity",
"model-index",
"text-embeddings-inference",
"endpoints_compatible",
"region:us"
]
| feature-extraction | 2022-09-14T00:41:39Z | ---
datasets:
- relbert/semeval2012_relational_similarity
model-index:
- name: relbert/roberta-large-semeval2012-mask-prompt-d-nce-conceptnet-validated
results:
- task:
name: Relation Mapping
type: sorting-task
dataset:
name: Relation Mapping
args: relbert/relation_mapping
type: relation-mapping
metrics:
- name: Accuracy
type: accuracy
value: 0.840515873015873
- task:
name: Analogy Questions (SAT full)
type: multiple-choice-qa
dataset:
name: SAT full
args: relbert/analogy_questions
type: analogy-questions
metrics:
- name: Accuracy
type: accuracy
value: 0.6818181818181818
- task:
name: Analogy Questions (SAT)
type: multiple-choice-qa
dataset:
name: SAT
args: relbert/analogy_questions
type: analogy-questions
metrics:
- name: Accuracy
type: accuracy
value: 0.685459940652819
- task:
name: Analogy Questions (BATS)
type: multiple-choice-qa
dataset:
name: BATS
args: relbert/analogy_questions
type: analogy-questions
metrics:
- name: Accuracy
type: accuracy
value: 0.8076709282934964
- task:
name: Analogy Questions (Google)
type: multiple-choice-qa
dataset:
name: Google
args: relbert/analogy_questions
type: analogy-questions
metrics:
- name: Accuracy
type: accuracy
value: 0.94
- task:
name: Analogy Questions (U2)
type: multiple-choice-qa
dataset:
name: U2
args: relbert/analogy_questions
type: analogy-questions
metrics:
- name: Accuracy
type: accuracy
value: 0.6535087719298246
- task:
name: Analogy Questions (U4)
type: multiple-choice-qa
dataset:
name: U4
args: relbert/analogy_questions
type: analogy-questions
metrics:
- name: Accuracy
type: accuracy
value: 0.6388888888888888
- task:
name: Lexical Relation Classification (BLESS)
type: classification
dataset:
name: BLESS
args: relbert/lexical_relation_classification
type: relation-classification
metrics:
- name: F1
type: f1
value: 0.9281301792978756
- name: F1 (macro)
type: f1_macro
value: 0.9254620165261186
- task:
name: Lexical Relation Classification (CogALexV)
type: classification
dataset:
name: CogALexV
args: relbert/lexical_relation_classification
type: relation-classification
metrics:
- name: F1
type: f1
value: 0.8856807511737088
- name: F1 (macro)
type: f1_macro
value: 0.7505936116426153
- task:
name: Lexical Relation Classification (EVALution)
type: classification
dataset:
name: BLESS
args: relbert/lexical_relation_classification
type: relation-classification
metrics:
- name: F1
type: f1
value: 0.7210184182015169
- name: F1 (macro)
type: f1_macro
value: 0.707381518416115
- task:
name: Lexical Relation Classification (K&H+N)
type: classification
dataset:
name: K&H+N
args: relbert/lexical_relation_classification
type: relation-classification
metrics:
- name: F1
type: f1
value: 0.9625791194268624
- name: F1 (macro)
type: f1_macro
value: 0.8830231594217628
- task:
name: Lexical Relation Classification (ROOT09)
type: classification
dataset:
name: ROOT09
args: relbert/lexical_relation_classification
type: relation-classification
metrics:
- name: F1
type: f1
value: 0.9207145095581323
- name: F1 (macro)
type: f1_macro
value: 0.9189981669115016
---
# relbert/roberta-large-semeval2012-mask-prompt-d-nce-conceptnet-validated
RelBERT fine-tuned from [roberta-large](https://huggingface.co/roberta-large) on
[relbert/semeval2012_relational_similarity](https://huggingface.co/datasets/relbert/semeval2012_relational_similarity).
Fine-tuning is done via [RelBERT](https://github.com/asahi417/relbert) library (see the repository for more detail).
It achieves the following results on the relation understanding tasks:
- Analogy Question ([dataset](https://huggingface.co/datasets/relbert/analogy_questions), [full result](https://huggingface.co/relbert/roberta-large-semeval2012-mask-prompt-d-nce-conceptnet-validated/raw/main/analogy.json)):
- Accuracy on SAT (full): 0.6818181818181818
- Accuracy on SAT: 0.685459940652819
- Accuracy on BATS: 0.8076709282934964
- Accuracy on U2: 0.6535087719298246
- Accuracy on U4: 0.6388888888888888
- Accuracy on Google: 0.94
- Lexical Relation Classification ([dataset](https://huggingface.co/datasets/relbert/lexical_relation_classification), [full result](https://huggingface.co/relbert/roberta-large-semeval2012-mask-prompt-d-nce-conceptnet-validated/raw/main/classification.json)):
- Micro F1 score on BLESS: 0.9281301792978756
- Micro F1 score on CogALexV: 0.8856807511737088
- Micro F1 score on EVALution: 0.7210184182015169
- Micro F1 score on K&H+N: 0.9625791194268624
- Micro F1 score on ROOT09: 0.9207145095581323
- Relation Mapping ([dataset](https://huggingface.co/datasets/relbert/relation_mapping), [full result](https://huggingface.co/relbert/roberta-large-semeval2012-mask-prompt-d-nce-conceptnet-validated/raw/main/relation_mapping.json)):
- Accuracy on Relation Mapping: 0.840515873015873
### Usage
This model can be used through the [relbert library](https://github.com/asahi417/relbert). Install the library via pip
```shell
pip install relbert
```
and activate model as below.
```python
from relbert import RelBERT
model = RelBERT("relbert/roberta-large-semeval2012-mask-prompt-d-nce-conceptnet-validated")
vector = model.get_embedding(['Tokyo', 'Japan']) # shape of (1024, )
```
### Training hyperparameters
The following hyperparameters were used during training:
- model: roberta-large
- max_length: 64
- mode: mask
- data: relbert/semeval2012_relational_similarity
- split: train
- data_eval: relbert/conceptnet_high_confidence
- split_eval: full
- template_mode: manual
- template: I wasn’t aware of this relationship, but I just read in the encyclopedia that <subj> is the <mask> of <obj>
- loss_function: nce_logout
- classification_loss: False
- temperature_nce_constant: 0.05
- temperature_nce_rank: {'min': 0.01, 'max': 0.05, 'type': 'linear'}
- epoch: 25
- batch: 128
- lr: 5e-06
- lr_decay: False
- lr_warmup: 1
- weight_decay: 0
- random_seed: 0
- exclude_relation: None
- exclude_relation_eval: None
- n_sample: 640
- gradient_accumulation: 8
The full configuration can be found at [fine-tuning parameter file](https://huggingface.co/relbert/roberta-large-semeval2012-mask-prompt-d-nce-conceptnet-validated/raw/main/trainer_config.json).
### Reference
If you use any resource from RelBERT, please consider to cite our [paper](https://aclanthology.org/2021.eacl-demos.7/).
```
@inproceedings{ushio-etal-2021-distilling-relation-embeddings,
title = "{D}istilling {R}elation {E}mbeddings from {P}re-trained {L}anguage {M}odels",
author = "Ushio, Asahi and
Schockaert, Steven and
Camacho-Collados, Jose",
booktitle = "EMNLP 2021",
year = "2021",
address = "Online",
publisher = "Association for Computational Linguistics",
}
```
|
research-backup/roberta-large-semeval2012-mask-prompt-c-nce-conceptnet-validated | research-backup | 2022-09-19T21:04:57Z | 95 | 0 | transformers | [
"transformers",
"pytorch",
"roberta",
"feature-extraction",
"dataset:relbert/semeval2012_relational_similarity",
"model-index",
"text-embeddings-inference",
"endpoints_compatible",
"region:us"
]
| feature-extraction | 2022-09-13T17:13:38Z | ---
datasets:
- relbert/semeval2012_relational_similarity
model-index:
- name: relbert/roberta-large-semeval2012-mask-prompt-c-nce-conceptnet-validated
results:
- task:
name: Relation Mapping
type: sorting-task
dataset:
name: Relation Mapping
args: relbert/relation_mapping
type: relation-mapping
metrics:
- name: Accuracy
type: accuracy
value: 0.7792857142857142
- task:
name: Analogy Questions (SAT full)
type: multiple-choice-qa
dataset:
name: SAT full
args: relbert/analogy_questions
type: analogy-questions
metrics:
- name: Accuracy
type: accuracy
value: 0.553475935828877
- task:
name: Analogy Questions (SAT)
type: multiple-choice-qa
dataset:
name: SAT
args: relbert/analogy_questions
type: analogy-questions
metrics:
- name: Accuracy
type: accuracy
value: 0.5578635014836796
- task:
name: Analogy Questions (BATS)
type: multiple-choice-qa
dataset:
name: BATS
args: relbert/analogy_questions
type: analogy-questions
metrics:
- name: Accuracy
type: accuracy
value: 0.6909394107837687
- task:
name: Analogy Questions (Google)
type: multiple-choice-qa
dataset:
name: Google
args: relbert/analogy_questions
type: analogy-questions
metrics:
- name: Accuracy
type: accuracy
value: 0.836
- task:
name: Analogy Questions (U2)
type: multiple-choice-qa
dataset:
name: U2
args: relbert/analogy_questions
type: analogy-questions
metrics:
- name: Accuracy
type: accuracy
value: 0.5043859649122807
- task:
name: Analogy Questions (U4)
type: multiple-choice-qa
dataset:
name: U4
args: relbert/analogy_questions
type: analogy-questions
metrics:
- name: Accuracy
type: accuracy
value: 0.5254629629629629
- task:
name: Lexical Relation Classification (BLESS)
type: classification
dataset:
name: BLESS
args: relbert/lexical_relation_classification
type: relation-classification
metrics:
- name: F1
type: f1
value: 0.9249660991411782
- name: F1 (macro)
type: f1_macro
value: 0.9197953117418775
- task:
name: Lexical Relation Classification (CogALexV)
type: classification
dataset:
name: CogALexV
args: relbert/lexical_relation_classification
type: relation-classification
metrics:
- name: F1
type: f1
value: 0.8671361502347418
- name: F1 (macro)
type: f1_macro
value: 0.7069093285623891
- task:
name: Lexical Relation Classification (EVALution)
type: classification
dataset:
name: BLESS
args: relbert/lexical_relation_classification
type: relation-classification
metrics:
- name: F1
type: f1
value: 0.6917659804983749
- name: F1 (macro)
type: f1_macro
value: 0.6801085174277005
- task:
name: Lexical Relation Classification (K&H+N)
type: classification
dataset:
name: K&H+N
args: relbert/lexical_relation_classification
type: relation-classification
metrics:
- name: F1
type: f1
value: 0.9645962300897266
- name: F1 (macro)
type: f1_macro
value: 0.9019275591441719
- task:
name: Lexical Relation Classification (ROOT09)
type: classification
dataset:
name: ROOT09
args: relbert/lexical_relation_classification
type: relation-classification
metrics:
- name: F1
type: f1
value: 0.8968975242870574
- name: F1 (macro)
type: f1_macro
value: 0.8958380600988481
---
# relbert/roberta-large-semeval2012-mask-prompt-c-nce-conceptnet-validated
RelBERT fine-tuned from [roberta-large](https://huggingface.co/roberta-large) on
[relbert/semeval2012_relational_similarity](https://huggingface.co/datasets/relbert/semeval2012_relational_similarity).
Fine-tuning is done via [RelBERT](https://github.com/asahi417/relbert) library (see the repository for more detail).
It achieves the following results on the relation understanding tasks:
- Analogy Question ([dataset](https://huggingface.co/datasets/relbert/analogy_questions), [full result](https://huggingface.co/relbert/roberta-large-semeval2012-mask-prompt-c-nce-conceptnet-validated/raw/main/analogy.json)):
- Accuracy on SAT (full): 0.553475935828877
- Accuracy on SAT: 0.5578635014836796
- Accuracy on BATS: 0.6909394107837687
- Accuracy on U2: 0.5043859649122807
- Accuracy on U4: 0.5254629629629629
- Accuracy on Google: 0.836
- Lexical Relation Classification ([dataset](https://huggingface.co/datasets/relbert/lexical_relation_classification), [full result](https://huggingface.co/relbert/roberta-large-semeval2012-mask-prompt-c-nce-conceptnet-validated/raw/main/classification.json)):
- Micro F1 score on BLESS: 0.9249660991411782
- Micro F1 score on CogALexV: 0.8671361502347418
- Micro F1 score on EVALution: 0.6917659804983749
- Micro F1 score on K&H+N: 0.9645962300897266
- Micro F1 score on ROOT09: 0.8968975242870574
- Relation Mapping ([dataset](https://huggingface.co/datasets/relbert/relation_mapping), [full result](https://huggingface.co/relbert/roberta-large-semeval2012-mask-prompt-c-nce-conceptnet-validated/raw/main/relation_mapping.json)):
- Accuracy on Relation Mapping: 0.7792857142857142
### Usage
This model can be used through the [relbert library](https://github.com/asahi417/relbert). Install the library via pip
```shell
pip install relbert
```
and activate model as below.
```python
from relbert import RelBERT
model = RelBERT("relbert/roberta-large-semeval2012-mask-prompt-c-nce-conceptnet-validated")
vector = model.get_embedding(['Tokyo', 'Japan']) # shape of (1024, )
```
### Training hyperparameters
The following hyperparameters were used during training:
- model: roberta-large
- max_length: 64
- mode: mask
- data: relbert/semeval2012_relational_similarity
- split: train
- data_eval: relbert/conceptnet_high_confidence
- split_eval: full
- template_mode: manual
- template: Today, I finally discovered the relation between <subj> and <obj> : <mask>
- loss_function: nce_logout
- classification_loss: False
- temperature_nce_constant: 0.05
- temperature_nce_rank: {'min': 0.01, 'max': 0.05, 'type': 'linear'}
- epoch: 30
- batch: 128
- lr: 5e-06
- lr_decay: False
- lr_warmup: 1
- weight_decay: 0
- random_seed: 0
- exclude_relation: None
- exclude_relation_eval: None
- n_sample: 640
- gradient_accumulation: 8
The full configuration can be found at [fine-tuning parameter file](https://huggingface.co/relbert/roberta-large-semeval2012-mask-prompt-c-nce-conceptnet-validated/raw/main/trainer_config.json).
### Reference
If you use any resource from RelBERT, please consider to cite our [paper](https://aclanthology.org/2021.eacl-demos.7/).
```
@inproceedings{ushio-etal-2021-distilling-relation-embeddings,
title = "{D}istilling {R}elation {E}mbeddings from {P}re-trained {L}anguage {M}odels",
author = "Ushio, Asahi and
Schockaert, Steven and
Camacho-Collados, Jose",
booktitle = "EMNLP 2021",
year = "2021",
address = "Online",
publisher = "Association for Computational Linguistics",
}
```
|
sd-concepts-library/joe-mad | sd-concepts-library | 2022-09-19T20:51:56Z | 0 | 2 | null | [
"license:mit",
"region:us"
]
| null | 2022-09-19T20:51:52Z | ---
license: mit
---
### Joe Mad on Stable Diffusion
This is the `<joe-mad>` concept taught to Stable Diffusion via Textual Inversion. You can load this concept into the [Stable Conceptualizer](https://colab.research.google.com/github/huggingface/notebooks/blob/main/diffusers/stable_conceptualizer_inference.ipynb) notebook. You can also train your own concepts and load them into the concept libraries using [this notebook](https://colab.research.google.com/github/huggingface/notebooks/blob/main/diffusers/sd_textual_inversion_training.ipynb).
Here is the new concept you will be able to use as a `style`:
|
research-backup/roberta-large-semeval2012-average-no-mask-prompt-e-nce-classification-conceptnet-validated | research-backup | 2022-09-19T20:49:24Z | 103 | 0 | transformers | [
"transformers",
"pytorch",
"roberta",
"feature-extraction",
"dataset:relbert/semeval2012_relational_similarity",
"model-index",
"text-embeddings-inference",
"endpoints_compatible",
"region:us"
]
| feature-extraction | 2022-09-11T19:39:08Z | ---
datasets:
- relbert/semeval2012_relational_similarity
model-index:
- name: relbert/roberta-large-semeval2012-average-no-mask-prompt-e-nce-classification-conceptnet-validated
results:
- task:
name: Relation Mapping
type: sorting-task
dataset:
name: Relation Mapping
args: relbert/relation_mapping
type: relation-mapping
metrics:
- name: Accuracy
type: accuracy
value: 0.7637698412698413
- task:
name: Analogy Questions (SAT full)
type: multiple-choice-qa
dataset:
name: SAT full
args: relbert/analogy_questions
type: analogy-questions
metrics:
- name: Accuracy
type: accuracy
value: 0.5133689839572193
- task:
name: Analogy Questions (SAT)
type: multiple-choice-qa
dataset:
name: SAT
args: relbert/analogy_questions
type: analogy-questions
metrics:
- name: Accuracy
type: accuracy
value: 0.516320474777448
- task:
name: Analogy Questions (BATS)
type: multiple-choice-qa
dataset:
name: BATS
args: relbert/analogy_questions
type: analogy-questions
metrics:
- name: Accuracy
type: accuracy
value: 0.5958866036687048
- task:
name: Analogy Questions (Google)
type: multiple-choice-qa
dataset:
name: Google
args: relbert/analogy_questions
type: analogy-questions
metrics:
- name: Accuracy
type: accuracy
value: 0.748
- task:
name: Analogy Questions (U2)
type: multiple-choice-qa
dataset:
name: U2
args: relbert/analogy_questions
type: analogy-questions
metrics:
- name: Accuracy
type: accuracy
value: 0.4605263157894737
- task:
name: Analogy Questions (U4)
type: multiple-choice-qa
dataset:
name: U4
args: relbert/analogy_questions
type: analogy-questions
metrics:
- name: Accuracy
type: accuracy
value: 0.5231481481481481
- task:
name: Lexical Relation Classification (BLESS)
type: classification
dataset:
name: BLESS
args: relbert/lexical_relation_classification
type: relation-classification
metrics:
- name: F1
type: f1
value: 0.9025161970769926
- name: F1 (macro)
type: f1_macro
value: 0.8979165451427438
- task:
name: Lexical Relation Classification (CogALexV)
type: classification
dataset:
name: CogALexV
args: relbert/lexical_relation_classification
type: relation-classification
metrics:
- name: F1
type: f1
value: 0.8328638497652581
- name: F1 (macro)
type: f1_macro
value: 0.6469572777603673
- task:
name: Lexical Relation Classification (EVALution)
type: classification
dataset:
name: BLESS
args: relbert/lexical_relation_classification
type: relation-classification
metrics:
- name: F1
type: f1
value: 0.6630552546045504
- name: F1 (macro)
type: f1_macro
value: 0.6493250582245075
- task:
name: Lexical Relation Classification (K&H+N)
type: classification
dataset:
name: K&H+N
args: relbert/lexical_relation_classification
type: relation-classification
metrics:
- name: F1
type: f1
value: 0.9562495652778744
- name: F1 (macro)
type: f1_macro
value: 0.8695137253747418
- task:
name: Lexical Relation Classification (ROOT09)
type: classification
dataset:
name: ROOT09
args: relbert/lexical_relation_classification
type: relation-classification
metrics:
- name: F1
type: f1
value: 0.8906298965841429
- name: F1 (macro)
type: f1_macro
value: 0.8885946595123109
---
# relbert/roberta-large-semeval2012-average-no-mask-prompt-e-nce-classification-conceptnet-validated
RelBERT fine-tuned from [roberta-large](https://huggingface.co/roberta-large) on
[relbert/semeval2012_relational_similarity](https://huggingface.co/datasets/relbert/semeval2012_relational_similarity).
Fine-tuning is done via [RelBERT](https://github.com/asahi417/relbert) library (see the repository for more detail).
It achieves the following results on the relation understanding tasks:
- Analogy Question ([dataset](https://huggingface.co/datasets/relbert/analogy_questions), [full result](https://huggingface.co/relbert/roberta-large-semeval2012-average-no-mask-prompt-e-nce-classification-conceptnet-validated/raw/main/analogy.json)):
- Accuracy on SAT (full): 0.5133689839572193
- Accuracy on SAT: 0.516320474777448
- Accuracy on BATS: 0.5958866036687048
- Accuracy on U2: 0.4605263157894737
- Accuracy on U4: 0.5231481481481481
- Accuracy on Google: 0.748
- Lexical Relation Classification ([dataset](https://huggingface.co/datasets/relbert/lexical_relation_classification), [full result](https://huggingface.co/relbert/roberta-large-semeval2012-average-no-mask-prompt-e-nce-classification-conceptnet-validated/raw/main/classification.json)):
- Micro F1 score on BLESS: 0.9025161970769926
- Micro F1 score on CogALexV: 0.8328638497652581
- Micro F1 score on EVALution: 0.6630552546045504
- Micro F1 score on K&H+N: 0.9562495652778744
- Micro F1 score on ROOT09: 0.8906298965841429
- Relation Mapping ([dataset](https://huggingface.co/datasets/relbert/relation_mapping), [full result](https://huggingface.co/relbert/roberta-large-semeval2012-average-no-mask-prompt-e-nce-classification-conceptnet-validated/raw/main/relation_mapping.json)):
- Accuracy on Relation Mapping: 0.7637698412698413
### Usage
This model can be used through the [relbert library](https://github.com/asahi417/relbert). Install the library via pip
```shell
pip install relbert
```
and activate model as below.
```python
from relbert import RelBERT
model = RelBERT("relbert/roberta-large-semeval2012-average-no-mask-prompt-e-nce-classification-conceptnet-validated")
vector = model.get_embedding(['Tokyo', 'Japan']) # shape of (1024, )
```
### Training hyperparameters
The following hyperparameters were used during training:
- model: roberta-large
- max_length: 64
- mode: average_no_mask
- data: relbert/semeval2012_relational_similarity
- split: train
- data_eval: relbert/conceptnet_high_confidence
- split_eval: full
- template_mode: manual
- template: I wasn’t aware of this relationship, but I just read in the encyclopedia that <obj> is <subj>’s <mask>
- loss_function: nce_logout
- classification_loss: True
- temperature_nce_constant: 0.05
- temperature_nce_rank: {'min': 0.01, 'max': 0.05, 'type': 'linear'}
- epoch: 30
- batch: 128
- lr: 5e-06
- lr_decay: False
- lr_warmup: 1
- weight_decay: 0
- random_seed: 0
- exclude_relation: None
- exclude_relation_eval: None
- n_sample: 640
- gradient_accumulation: 8
The full configuration can be found at [fine-tuning parameter file](https://huggingface.co/relbert/roberta-large-semeval2012-average-no-mask-prompt-e-nce-classification-conceptnet-validated/raw/main/trainer_config.json).
### Reference
If you use any resource from RelBERT, please consider to cite our [paper](https://aclanthology.org/2021.eacl-demos.7/).
```
@inproceedings{ushio-etal-2021-distilling-relation-embeddings,
title = "{D}istilling {R}elation {E}mbeddings from {P}re-trained {L}anguage {M}odels",
author = "Ushio, Asahi and
Schockaert, Steven and
Camacho-Collados, Jose",
booktitle = "EMNLP 2021",
year = "2021",
address = "Online",
publisher = "Association for Computational Linguistics",
}
```
|
research-backup/roberta-large-semeval2012-average-no-mask-prompt-b-nce-classification-conceptnet-validated | research-backup | 2022-09-19T20:37:42Z | 105 | 0 | transformers | [
"transformers",
"pytorch",
"roberta",
"feature-extraction",
"dataset:relbert/semeval2012_relational_similarity",
"model-index",
"text-embeddings-inference",
"endpoints_compatible",
"region:us"
]
| feature-extraction | 2022-09-10T17:17:52Z | ---
datasets:
- relbert/semeval2012_relational_similarity
model-index:
- name: relbert/roberta-large-semeval2012-average-no-mask-prompt-b-nce-classification-conceptnet-validated
results:
- task:
name: Relation Mapping
type: sorting-task
dataset:
name: Relation Mapping
args: relbert/relation_mapping
type: relation-mapping
metrics:
- name: Accuracy
type: accuracy
value: 0.8167460317460318
- task:
name: Analogy Questions (SAT full)
type: multiple-choice-qa
dataset:
name: SAT full
args: relbert/analogy_questions
type: analogy-questions
metrics:
- name: Accuracy
type: accuracy
value: 0.516042780748663
- task:
name: Analogy Questions (SAT)
type: multiple-choice-qa
dataset:
name: SAT
args: relbert/analogy_questions
type: analogy-questions
metrics:
- name: Accuracy
type: accuracy
value: 0.5281899109792285
- task:
name: Analogy Questions (BATS)
type: multiple-choice-qa
dataset:
name: BATS
args: relbert/analogy_questions
type: analogy-questions
metrics:
- name: Accuracy
type: accuracy
value: 0.632017787659811
- task:
name: Analogy Questions (Google)
type: multiple-choice-qa
dataset:
name: Google
args: relbert/analogy_questions
type: analogy-questions
metrics:
- name: Accuracy
type: accuracy
value: 0.724
- task:
name: Analogy Questions (U2)
type: multiple-choice-qa
dataset:
name: U2
args: relbert/analogy_questions
type: analogy-questions
metrics:
- name: Accuracy
type: accuracy
value: 0.4342105263157895
- task:
name: Analogy Questions (U4)
type: multiple-choice-qa
dataset:
name: U4
args: relbert/analogy_questions
type: analogy-questions
metrics:
- name: Accuracy
type: accuracy
value: 0.5069444444444444
- task:
name: Lexical Relation Classification (BLESS)
type: classification
dataset:
name: BLESS
args: relbert/lexical_relation_classification
type: relation-classification
metrics:
- name: F1
type: f1
value: 0.9034202199789061
- name: F1 (macro)
type: f1_macro
value: 0.893273397921436
- task:
name: Lexical Relation Classification (CogALexV)
type: classification
dataset:
name: CogALexV
args: relbert/lexical_relation_classification
type: relation-classification
metrics:
- name: F1
type: f1
value: 0.8342723004694835
- name: F1 (macro)
type: f1_macro
value: 0.6453699846432566
- task:
name: Lexical Relation Classification (EVALution)
type: classification
dataset:
name: BLESS
args: relbert/lexical_relation_classification
type: relation-classification
metrics:
- name: F1
type: f1
value: 0.6581798483206934
- name: F1 (macro)
type: f1_macro
value: 0.640639393261134
- task:
name: Lexical Relation Classification (K&H+N)
type: classification
dataset:
name: K&H+N
args: relbert/lexical_relation_classification
type: relation-classification
metrics:
- name: F1
type: f1
value: 0.9604228976838005
- name: F1 (macro)
type: f1_macro
value: 0.8814339609725079
- task:
name: Lexical Relation Classification (ROOT09)
type: classification
dataset:
name: ROOT09
args: relbert/lexical_relation_classification
type: relation-classification
metrics:
- name: F1
type: f1
value: 0.8909432779692886
- name: F1 (macro)
type: f1_macro
value: 0.8914692333897629
---
# relbert/roberta-large-semeval2012-average-no-mask-prompt-b-nce-classification-conceptnet-validated
RelBERT fine-tuned from [roberta-large](https://huggingface.co/roberta-large) on
[relbert/semeval2012_relational_similarity](https://huggingface.co/datasets/relbert/semeval2012_relational_similarity).
Fine-tuning is done via [RelBERT](https://github.com/asahi417/relbert) library (see the repository for more detail).
It achieves the following results on the relation understanding tasks:
- Analogy Question ([dataset](https://huggingface.co/datasets/relbert/analogy_questions), [full result](https://huggingface.co/relbert/roberta-large-semeval2012-average-no-mask-prompt-b-nce-classification-conceptnet-validated/raw/main/analogy.json)):
- Accuracy on SAT (full): 0.516042780748663
- Accuracy on SAT: 0.5281899109792285
- Accuracy on BATS: 0.632017787659811
- Accuracy on U2: 0.4342105263157895
- Accuracy on U4: 0.5069444444444444
- Accuracy on Google: 0.724
- Lexical Relation Classification ([dataset](https://huggingface.co/datasets/relbert/lexical_relation_classification), [full result](https://huggingface.co/relbert/roberta-large-semeval2012-average-no-mask-prompt-b-nce-classification-conceptnet-validated/raw/main/classification.json)):
- Micro F1 score on BLESS: 0.9034202199789061
- Micro F1 score on CogALexV: 0.8342723004694835
- Micro F1 score on EVALution: 0.6581798483206934
- Micro F1 score on K&H+N: 0.9604228976838005
- Micro F1 score on ROOT09: 0.8909432779692886
- Relation Mapping ([dataset](https://huggingface.co/datasets/relbert/relation_mapping), [full result](https://huggingface.co/relbert/roberta-large-semeval2012-average-no-mask-prompt-b-nce-classification-conceptnet-validated/raw/main/relation_mapping.json)):
- Accuracy on Relation Mapping: 0.8167460317460318
### Usage
This model can be used through the [relbert library](https://github.com/asahi417/relbert). Install the library via pip
```shell
pip install relbert
```
and activate model as below.
```python
from relbert import RelBERT
model = RelBERT("relbert/roberta-large-semeval2012-average-no-mask-prompt-b-nce-classification-conceptnet-validated")
vector = model.get_embedding(['Tokyo', 'Japan']) # shape of (1024, )
```
### Training hyperparameters
The following hyperparameters were used during training:
- model: roberta-large
- max_length: 64
- mode: average_no_mask
- data: relbert/semeval2012_relational_similarity
- split: train
- data_eval: relbert/conceptnet_high_confidence
- split_eval: full
- template_mode: manual
- template: Today, I finally discovered the relation between <subj> and <obj> : <obj> is <subj>'s <mask>
- loss_function: nce_logout
- classification_loss: True
- temperature_nce_constant: 0.05
- temperature_nce_rank: {'min': 0.01, 'max': 0.05, 'type': 'linear'}
- epoch: 30
- batch: 128
- lr: 5e-06
- lr_decay: False
- lr_warmup: 1
- weight_decay: 0
- random_seed: 0
- exclude_relation: None
- exclude_relation_eval: None
- n_sample: 640
- gradient_accumulation: 8
The full configuration can be found at [fine-tuning parameter file](https://huggingface.co/relbert/roberta-large-semeval2012-average-no-mask-prompt-b-nce-classification-conceptnet-validated/raw/main/trainer_config.json).
### Reference
If you use any resource from RelBERT, please consider to cite our [paper](https://aclanthology.org/2021.eacl-demos.7/).
```
@inproceedings{ushio-etal-2021-distilling-relation-embeddings,
title = "{D}istilling {R}elation {E}mbeddings from {P}re-trained {L}anguage {M}odels",
author = "Ushio, Asahi and
Schockaert, Steven and
Camacho-Collados, Jose",
booktitle = "EMNLP 2021",
year = "2021",
address = "Online",
publisher = "Association for Computational Linguistics",
}
```
|
research-backup/roberta-large-semeval2012-average-no-mask-prompt-a-nce-classification-conceptnet-validated | research-backup | 2022-09-19T20:34:07Z | 103 | 0 | transformers | [
"transformers",
"pytorch",
"roberta",
"feature-extraction",
"dataset:relbert/semeval2012_relational_similarity",
"model-index",
"text-embeddings-inference",
"endpoints_compatible",
"region:us"
]
| feature-extraction | 2022-09-10T08:32:21Z | ---
datasets:
- relbert/semeval2012_relational_similarity
model-index:
- name: relbert/roberta-large-semeval2012-average-no-mask-prompt-a-nce-classification-conceptnet-validated
results:
- task:
name: Relation Mapping
type: sorting-task
dataset:
name: Relation Mapping
args: relbert/relation_mapping
type: relation-mapping
metrics:
- name: Accuracy
type: accuracy
value: 0.7367857142857143
- task:
name: Analogy Questions (SAT full)
type: multiple-choice-qa
dataset:
name: SAT full
args: relbert/analogy_questions
type: analogy-questions
metrics:
- name: Accuracy
type: accuracy
value: 0.3342245989304813
- task:
name: Analogy Questions (SAT)
type: multiple-choice-qa
dataset:
name: SAT
args: relbert/analogy_questions
type: analogy-questions
metrics:
- name: Accuracy
type: accuracy
value: 0.33827893175074186
- task:
name: Analogy Questions (BATS)
type: multiple-choice-qa
dataset:
name: BATS
args: relbert/analogy_questions
type: analogy-questions
metrics:
- name: Accuracy
type: accuracy
value: 0.3968871595330739
- task:
name: Analogy Questions (Google)
type: multiple-choice-qa
dataset:
name: Google
args: relbert/analogy_questions
type: analogy-questions
metrics:
- name: Accuracy
type: accuracy
value: 0.592
- task:
name: Analogy Questions (U2)
type: multiple-choice-qa
dataset:
name: U2
args: relbert/analogy_questions
type: analogy-questions
metrics:
- name: Accuracy
type: accuracy
value: 0.3201754385964912
- task:
name: Analogy Questions (U4)
type: multiple-choice-qa
dataset:
name: U4
args: relbert/analogy_questions
type: analogy-questions
metrics:
- name: Accuracy
type: accuracy
value: 0.3125
- task:
name: Lexical Relation Classification (BLESS)
type: classification
dataset:
name: BLESS
args: relbert/lexical_relation_classification
type: relation-classification
metrics:
- name: F1
type: f1
value: 0.9022148561096881
- name: F1 (macro)
type: f1_macro
value: 0.8962429050248129
- task:
name: Lexical Relation Classification (CogALexV)
type: classification
dataset:
name: CogALexV
args: relbert/lexical_relation_classification
type: relation-classification
metrics:
- name: F1
type: f1
value: 0.8049295774647888
- name: F1 (macro)
type: f1_macro
value: 0.6122481358269966
- task:
name: Lexical Relation Classification (EVALution)
type: classification
dataset:
name: BLESS
args: relbert/lexical_relation_classification
type: relation-classification
metrics:
- name: F1
type: f1
value: 0.652762730227519
- name: F1 (macro)
type: f1_macro
value: 0.6101323743101166
- task:
name: Lexical Relation Classification (K&H+N)
type: classification
dataset:
name: K&H+N
args: relbert/lexical_relation_classification
type: relation-classification
metrics:
- name: F1
type: f1
value: 0.9603533421437018
- name: F1 (macro)
type: f1_macro
value: 0.8709644325592566
- task:
name: Lexical Relation Classification (ROOT09)
type: classification
dataset:
name: ROOT09
args: relbert/lexical_relation_classification
type: relation-classification
metrics:
- name: F1
type: f1
value: 0.8874960827326857
- name: F1 (macro)
type: f1_macro
value: 0.8864394662565577
---
# relbert/roberta-large-semeval2012-average-no-mask-prompt-a-nce-classification-conceptnet-validated
RelBERT fine-tuned from [roberta-large](https://huggingface.co/roberta-large) on
[relbert/semeval2012_relational_similarity](https://huggingface.co/datasets/relbert/semeval2012_relational_similarity).
Fine-tuning is done via [RelBERT](https://github.com/asahi417/relbert) library (see the repository for more detail).
It achieves the following results on the relation understanding tasks:
- Analogy Question ([dataset](https://huggingface.co/datasets/relbert/analogy_questions), [full result](https://huggingface.co/relbert/roberta-large-semeval2012-average-no-mask-prompt-a-nce-classification-conceptnet-validated/raw/main/analogy.json)):
- Accuracy on SAT (full): 0.3342245989304813
- Accuracy on SAT: 0.33827893175074186
- Accuracy on BATS: 0.3968871595330739
- Accuracy on U2: 0.3201754385964912
- Accuracy on U4: 0.3125
- Accuracy on Google: 0.592
- Lexical Relation Classification ([dataset](https://huggingface.co/datasets/relbert/lexical_relation_classification), [full result](https://huggingface.co/relbert/roberta-large-semeval2012-average-no-mask-prompt-a-nce-classification-conceptnet-validated/raw/main/classification.json)):
- Micro F1 score on BLESS: 0.9022148561096881
- Micro F1 score on CogALexV: 0.8049295774647888
- Micro F1 score on EVALution: 0.652762730227519
- Micro F1 score on K&H+N: 0.9603533421437018
- Micro F1 score on ROOT09: 0.8874960827326857
- Relation Mapping ([dataset](https://huggingface.co/datasets/relbert/relation_mapping), [full result](https://huggingface.co/relbert/roberta-large-semeval2012-average-no-mask-prompt-a-nce-classification-conceptnet-validated/raw/main/relation_mapping.json)):
- Accuracy on Relation Mapping: 0.7367857142857143
### Usage
This model can be used through the [relbert library](https://github.com/asahi417/relbert). Install the library via pip
```shell
pip install relbert
```
and activate model as below.
```python
from relbert import RelBERT
model = RelBERT("relbert/roberta-large-semeval2012-average-no-mask-prompt-a-nce-classification-conceptnet-validated")
vector = model.get_embedding(['Tokyo', 'Japan']) # shape of (1024, )
```
### Training hyperparameters
The following hyperparameters were used during training:
- model: roberta-large
- max_length: 64
- mode: average_no_mask
- data: relbert/semeval2012_relational_similarity
- split: train
- data_eval: relbert/conceptnet_high_confidence
- split_eval: full
- template_mode: manual
- template: Today, I finally discovered the relation between <subj> and <obj> : <subj> is the <mask> of <obj>
- loss_function: nce_logout
- classification_loss: True
- temperature_nce_constant: 0.05
- temperature_nce_rank: {'min': 0.01, 'max': 0.05, 'type': 'linear'}
- epoch: 1
- batch: 128
- lr: 5e-06
- lr_decay: False
- lr_warmup: 1
- weight_decay: 0
- random_seed: 0
- exclude_relation: None
- exclude_relation_eval: None
- n_sample: 640
- gradient_accumulation: 8
The full configuration can be found at [fine-tuning parameter file](https://huggingface.co/relbert/roberta-large-semeval2012-average-no-mask-prompt-a-nce-classification-conceptnet-validated/raw/main/trainer_config.json).
### Reference
If you use any resource from RelBERT, please consider to cite our [paper](https://aclanthology.org/2021.eacl-demos.7/).
```
@inproceedings{ushio-etal-2021-distilling-relation-embeddings,
title = "{D}istilling {R}elation {E}mbeddings from {P}re-trained {L}anguage {M}odels",
author = "Ushio, Asahi and
Schockaert, Steven and
Camacho-Collados, Jose",
booktitle = "EMNLP 2021",
year = "2021",
address = "Online",
publisher = "Association for Computational Linguistics",
}
```
|
research-backup/roberta-large-semeval2012-mask-prompt-d-nce-classification-conceptnet-validated | research-backup | 2022-09-19T20:14:34Z | 104 | 0 | transformers | [
"transformers",
"pytorch",
"roberta",
"feature-extraction",
"dataset:relbert/semeval2012_relational_similarity",
"model-index",
"text-embeddings-inference",
"endpoints_compatible",
"region:us"
]
| feature-extraction | 2022-09-07T21:38:26Z | ---
datasets:
- relbert/semeval2012_relational_similarity
model-index:
- name: relbert/roberta-large-semeval2012-mask-prompt-d-nce-classification-conceptnet-validated
results:
- task:
name: Relation Mapping
type: sorting-task
dataset:
name: Relation Mapping
args: relbert/relation_mapping
type: relation-mapping
metrics:
- name: Accuracy
type: accuracy
value: 0.8544444444444445
- task:
name: Analogy Questions (SAT full)
type: multiple-choice-qa
dataset:
name: SAT full
args: relbert/analogy_questions
type: analogy-questions
metrics:
- name: Accuracy
type: accuracy
value: 0.6524064171122995
- task:
name: Analogy Questions (SAT)
type: multiple-choice-qa
dataset:
name: SAT
args: relbert/analogy_questions
type: analogy-questions
metrics:
- name: Accuracy
type: accuracy
value: 0.6498516320474778
- task:
name: Analogy Questions (BATS)
type: multiple-choice-qa
dataset:
name: BATS
args: relbert/analogy_questions
type: analogy-questions
metrics:
- name: Accuracy
type: accuracy
value: 0.7509727626459144
- task:
name: Analogy Questions (Google)
type: multiple-choice-qa
dataset:
name: Google
args: relbert/analogy_questions
type: analogy-questions
metrics:
- name: Accuracy
type: accuracy
value: 0.902
- task:
name: Analogy Questions (U2)
type: multiple-choice-qa
dataset:
name: U2
args: relbert/analogy_questions
type: analogy-questions
metrics:
- name: Accuracy
type: accuracy
value: 0.6271929824561403
- task:
name: Analogy Questions (U4)
type: multiple-choice-qa
dataset:
name: U4
args: relbert/analogy_questions
type: analogy-questions
metrics:
- name: Accuracy
type: accuracy
value: 0.625
- task:
name: Lexical Relation Classification (BLESS)
type: classification
dataset:
name: BLESS
args: relbert/lexical_relation_classification
type: relation-classification
metrics:
- name: F1
type: f1
value: 0.9246647581738737
- name: F1 (macro)
type: f1_macro
value: 0.9201116139693363
- task:
name: Lexical Relation Classification (CogALexV)
type: classification
dataset:
name: CogALexV
args: relbert/lexical_relation_classification
type: relation-classification
metrics:
- name: F1
type: f1
value: 0.8826291079812206
- name: F1 (macro)
type: f1_macro
value: 0.74506786895136
- task:
name: Lexical Relation Classification (EVALution)
type: classification
dataset:
name: BLESS
args: relbert/lexical_relation_classification
type: relation-classification
metrics:
- name: F1
type: f1
value: 0.7172264355362946
- name: F1 (macro)
type: f1_macro
value: 0.703292242462215
- task:
name: Lexical Relation Classification (K&H+N)
type: classification
dataset:
name: K&H+N
args: relbert/lexical_relation_classification
type: relation-classification
metrics:
- name: F1
type: f1
value: 0.9616748974055783
- name: F1 (macro)
type: f1_macro
value: 0.8934154139843127
- task:
name: Lexical Relation Classification (ROOT09)
type: classification
dataset:
name: ROOT09
args: relbert/lexical_relation_classification
type: relation-classification
metrics:
- name: F1
type: f1
value: 0.9094327796928863
- name: F1 (macro)
type: f1_macro
value: 0.906471425124189
---
# relbert/roberta-large-semeval2012-mask-prompt-d-nce-classification-conceptnet-validated
RelBERT fine-tuned from [roberta-large](https://huggingface.co/roberta-large) on
[relbert/semeval2012_relational_similarity](https://huggingface.co/datasets/relbert/semeval2012_relational_similarity).
Fine-tuning is done via [RelBERT](https://github.com/asahi417/relbert) library (see the repository for more detail).
It achieves the following results on the relation understanding tasks:
- Analogy Question ([dataset](https://huggingface.co/datasets/relbert/analogy_questions), [full result](https://huggingface.co/relbert/roberta-large-semeval2012-mask-prompt-d-nce-classification-conceptnet-validated/raw/main/analogy.json)):
- Accuracy on SAT (full): 0.6524064171122995
- Accuracy on SAT: 0.6498516320474778
- Accuracy on BATS: 0.7509727626459144
- Accuracy on U2: 0.6271929824561403
- Accuracy on U4: 0.625
- Accuracy on Google: 0.902
- Lexical Relation Classification ([dataset](https://huggingface.co/datasets/relbert/lexical_relation_classification), [full result](https://huggingface.co/relbert/roberta-large-semeval2012-mask-prompt-d-nce-classification-conceptnet-validated/raw/main/classification.json)):
- Micro F1 score on BLESS: 0.9246647581738737
- Micro F1 score on CogALexV: 0.8826291079812206
- Micro F1 score on EVALution: 0.7172264355362946
- Micro F1 score on K&H+N: 0.9616748974055783
- Micro F1 score on ROOT09: 0.9094327796928863
- Relation Mapping ([dataset](https://huggingface.co/datasets/relbert/relation_mapping), [full result](https://huggingface.co/relbert/roberta-large-semeval2012-mask-prompt-d-nce-classification-conceptnet-validated/raw/main/relation_mapping.json)):
- Accuracy on Relation Mapping: 0.8544444444444445
### Usage
This model can be used through the [relbert library](https://github.com/asahi417/relbert). Install the library via pip
```shell
pip install relbert
```
and activate model as below.
```python
from relbert import RelBERT
model = RelBERT("relbert/roberta-large-semeval2012-mask-prompt-d-nce-classification-conceptnet-validated")
vector = model.get_embedding(['Tokyo', 'Japan']) # shape of (1024, )
```
### Training hyperparameters
The following hyperparameters were used during training:
- model: roberta-large
- max_length: 64
- mode: mask
- data: relbert/semeval2012_relational_similarity
- split: train
- data_eval: relbert/conceptnet_high_confidence
- split_eval: full
- template_mode: manual
- template: I wasn’t aware of this relationship, but I just read in the encyclopedia that <subj> is the <mask> of <obj>
- loss_function: nce_logout
- classification_loss: True
- temperature_nce_constant: 0.05
- temperature_nce_rank: {'min': 0.01, 'max': 0.05, 'type': 'linear'}
- epoch: 30
- batch: 128
- lr: 5e-06
- lr_decay: False
- lr_warmup: 1
- weight_decay: 0
- random_seed: 0
- exclude_relation: None
- exclude_relation_eval: None
- n_sample: 640
- gradient_accumulation: 8
The full configuration can be found at [fine-tuning parameter file](https://huggingface.co/relbert/roberta-large-semeval2012-mask-prompt-d-nce-classification-conceptnet-validated/raw/main/trainer_config.json).
### Reference
If you use any resource from RelBERT, please consider to cite our [paper](https://aclanthology.org/2021.eacl-demos.7/).
```
@inproceedings{ushio-etal-2021-distilling-relation-embeddings,
title = "{D}istilling {R}elation {E}mbeddings from {P}re-trained {L}anguage {M}odels",
author = "Ushio, Asahi and
Schockaert, Steven and
Camacho-Collados, Jose",
booktitle = "EMNLP 2021",
year = "2021",
address = "Online",
publisher = "Association for Computational Linguistics",
}
```
|
research-backup/roberta-large-semeval2012-mask-prompt-c-nce-classification-conceptnet-validated | research-backup | 2022-09-19T20:10:10Z | 102 | 0 | transformers | [
"transformers",
"pytorch",
"roberta",
"feature-extraction",
"dataset:relbert/semeval2012_relational_similarity",
"model-index",
"text-embeddings-inference",
"endpoints_compatible",
"region:us"
]
| feature-extraction | 2022-09-07T10:51:00Z | ---
datasets:
- relbert/semeval2012_relational_similarity
model-index:
- name: relbert/roberta-large-semeval2012-mask-prompt-c-nce-classification-conceptnet-validated
results:
- task:
name: Relation Mapping
type: sorting-task
dataset:
name: Relation Mapping
args: relbert/relation_mapping
type: relation-mapping
metrics:
- name: Accuracy
type: accuracy
value: 0.5911706349206349
- task:
name: Analogy Questions (SAT full)
type: multiple-choice-qa
dataset:
name: SAT full
args: relbert/analogy_questions
type: analogy-questions
metrics:
- name: Accuracy
type: accuracy
value: 0.3235294117647059
- task:
name: Analogy Questions (SAT)
type: multiple-choice-qa
dataset:
name: SAT
args: relbert/analogy_questions
type: analogy-questions
metrics:
- name: Accuracy
type: accuracy
value: 0.314540059347181
- task:
name: Analogy Questions (BATS)
type: multiple-choice-qa
dataset:
name: BATS
args: relbert/analogy_questions
type: analogy-questions
metrics:
- name: Accuracy
type: accuracy
value: 0.4118954974986103
- task:
name: Analogy Questions (Google)
type: multiple-choice-qa
dataset:
name: Google
args: relbert/analogy_questions
type: analogy-questions
metrics:
- name: Accuracy
type: accuracy
value: 0.43
- task:
name: Analogy Questions (U2)
type: multiple-choice-qa
dataset:
name: U2
args: relbert/analogy_questions
type: analogy-questions
metrics:
- name: Accuracy
type: accuracy
value: 0.34649122807017546
- task:
name: Analogy Questions (U4)
type: multiple-choice-qa
dataset:
name: U4
args: relbert/analogy_questions
type: analogy-questions
metrics:
- name: Accuracy
type: accuracy
value: 0.3125
- task:
name: Lexical Relation Classification (BLESS)
type: classification
dataset:
name: BLESS
args: relbert/lexical_relation_classification
type: relation-classification
metrics:
- name: F1
type: f1
value: 0.8142232936567726
- name: F1 (macro)
type: f1_macro
value: 0.7823150685401111
- task:
name: Lexical Relation Classification (CogALexV)
type: classification
dataset:
name: CogALexV
args: relbert/lexical_relation_classification
type: relation-classification
metrics:
- name: F1
type: f1
value: 0.7779342723004695
- name: F1 (macro)
type: f1_macro
value: 0.4495225434483775
- task:
name: Lexical Relation Classification (EVALution)
type: classification
dataset:
name: BLESS
args: relbert/lexical_relation_classification
type: relation-classification
metrics:
- name: F1
type: f1
value: 0.5357529794149513
- name: F1 (macro)
type: f1_macro
value: 0.45418166183928343
- task:
name: Lexical Relation Classification (K&H+N)
type: classification
dataset:
name: K&H+N
args: relbert/lexical_relation_classification
type: relation-classification
metrics:
- name: F1
type: f1
value: 0.8190164846630034
- name: F1 (macro)
type: f1_macro
value: 0.6465234410767566
- task:
name: Lexical Relation Classification (ROOT09)
type: classification
dataset:
name: ROOT09
args: relbert/lexical_relation_classification
type: relation-classification
metrics:
- name: F1
type: f1
value: 0.8834221247257913
- name: F1 (macro)
type: f1_macro
value: 0.8771202456083294
---
# relbert/roberta-large-semeval2012-mask-prompt-c-nce-classification-conceptnet-validated
RelBERT fine-tuned from [roberta-large](https://huggingface.co/roberta-large) on
[relbert/semeval2012_relational_similarity](https://huggingface.co/datasets/relbert/semeval2012_relational_similarity).
Fine-tuning is done via [RelBERT](https://github.com/asahi417/relbert) library (see the repository for more detail).
It achieves the following results on the relation understanding tasks:
- Analogy Question ([dataset](https://huggingface.co/datasets/relbert/analogy_questions), [full result](https://huggingface.co/relbert/roberta-large-semeval2012-mask-prompt-c-nce-classification-conceptnet-validated/raw/main/analogy.json)):
- Accuracy on SAT (full): 0.3235294117647059
- Accuracy on SAT: 0.314540059347181
- Accuracy on BATS: 0.4118954974986103
- Accuracy on U2: 0.34649122807017546
- Accuracy on U4: 0.3125
- Accuracy on Google: 0.43
- Lexical Relation Classification ([dataset](https://huggingface.co/datasets/relbert/lexical_relation_classification), [full result](https://huggingface.co/relbert/roberta-large-semeval2012-mask-prompt-c-nce-classification-conceptnet-validated/raw/main/classification.json)):
- Micro F1 score on BLESS: 0.8142232936567726
- Micro F1 score on CogALexV: 0.7779342723004695
- Micro F1 score on EVALution: 0.5357529794149513
- Micro F1 score on K&H+N: 0.8190164846630034
- Micro F1 score on ROOT09: 0.8834221247257913
- Relation Mapping ([dataset](https://huggingface.co/datasets/relbert/relation_mapping), [full result](https://huggingface.co/relbert/roberta-large-semeval2012-mask-prompt-c-nce-classification-conceptnet-validated/raw/main/relation_mapping.json)):
- Accuracy on Relation Mapping: 0.5911706349206349
### Usage
This model can be used through the [relbert library](https://github.com/asahi417/relbert). Install the library via pip
```shell
pip install relbert
```
and activate model as below.
```python
from relbert import RelBERT
model = RelBERT("relbert/roberta-large-semeval2012-mask-prompt-c-nce-classification-conceptnet-validated")
vector = model.get_embedding(['Tokyo', 'Japan']) # shape of (1024, )
```
### Training hyperparameters
The following hyperparameters were used during training:
- model: roberta-large
- max_length: 64
- mode: mask
- data: relbert/semeval2012_relational_similarity
- split: train
- data_eval: relbert/conceptnet_high_confidence
- split_eval: full
- template_mode: manual
- template: Today, I finally discovered the relation between <subj> and <obj> : <mask>
- loss_function: nce_logout
- classification_loss: True
- temperature_nce_constant: 0.05
- temperature_nce_rank: {'min': 0.01, 'max': 0.05, 'type': 'linear'}
- epoch: 24
- batch: 128
- lr: 5e-06
- lr_decay: False
- lr_warmup: 1
- weight_decay: 0
- random_seed: 0
- exclude_relation: None
- exclude_relation_eval: None
- n_sample: 640
- gradient_accumulation: 8
The full configuration can be found at [fine-tuning parameter file](https://huggingface.co/relbert/roberta-large-semeval2012-mask-prompt-c-nce-classification-conceptnet-validated/raw/main/trainer_config.json).
### Reference
If you use any resource from RelBERT, please consider to cite our [paper](https://aclanthology.org/2021.eacl-demos.7/).
```
@inproceedings{ushio-etal-2021-distilling-relation-embeddings,
title = "{D}istilling {R}elation {E}mbeddings from {P}re-trained {L}anguage {M}odels",
author = "Ushio, Asahi and
Schockaert, Steven and
Camacho-Collados, Jose",
booktitle = "EMNLP 2021",
year = "2021",
address = "Online",
publisher = "Association for Computational Linguistics",
}
```
|
research-backup/roberta-large-semeval2012-average-prompt-d-nce-classification-conceptnet-validated | research-backup | 2022-09-19T19:55:10Z | 103 | 0 | transformers | [
"transformers",
"pytorch",
"roberta",
"feature-extraction",
"dataset:relbert/semeval2012_relational_similarity",
"model-index",
"text-embeddings-inference",
"endpoints_compatible",
"region:us"
]
| feature-extraction | 2022-09-09T12:51:44Z | ---
datasets:
- relbert/semeval2012_relational_similarity
model-index:
- name: relbert/roberta-large-semeval2012-average-prompt-d-nce-classification-conceptnet-validated
results:
- task:
name: Relation Mapping
type: sorting-task
dataset:
name: Relation Mapping
args: relbert/relation_mapping
type: relation-mapping
metrics:
- name: Accuracy
type: accuracy
value: 0.8295436507936508
- task:
name: Analogy Questions (SAT full)
type: multiple-choice-qa
dataset:
name: SAT full
args: relbert/analogy_questions
type: analogy-questions
metrics:
- name: Accuracy
type: accuracy
value: 0.5828877005347594
- task:
name: Analogy Questions (SAT)
type: multiple-choice-qa
dataset:
name: SAT
args: relbert/analogy_questions
type: analogy-questions
metrics:
- name: Accuracy
type: accuracy
value: 0.6023738872403561
- task:
name: Analogy Questions (BATS)
type: multiple-choice-qa
dataset:
name: BATS
args: relbert/analogy_questions
type: analogy-questions
metrics:
- name: Accuracy
type: accuracy
value: 0.6170094496942746
- task:
name: Analogy Questions (Google)
type: multiple-choice-qa
dataset:
name: Google
args: relbert/analogy_questions
type: analogy-questions
metrics:
- name: Accuracy
type: accuracy
value: 0.842
- task:
name: Analogy Questions (U2)
type: multiple-choice-qa
dataset:
name: U2
args: relbert/analogy_questions
type: analogy-questions
metrics:
- name: Accuracy
type: accuracy
value: 0.5219298245614035
- task:
name: Analogy Questions (U4)
type: multiple-choice-qa
dataset:
name: U4
args: relbert/analogy_questions
type: analogy-questions
metrics:
- name: Accuracy
type: accuracy
value: 0.5347222222222222
- task:
name: Lexical Relation Classification (BLESS)
type: classification
dataset:
name: BLESS
args: relbert/lexical_relation_classification
type: relation-classification
metrics:
- name: F1
type: f1
value: 0.9127617899653457
- name: F1 (macro)
type: f1_macro
value: 0.9077484042036353
- task:
name: Lexical Relation Classification (CogALexV)
type: classification
dataset:
name: CogALexV
args: relbert/lexical_relation_classification
type: relation-classification
metrics:
- name: F1
type: f1
value: 0.8523474178403756
- name: F1 (macro)
type: f1_macro
value: 0.6871561847645433
- task:
name: Lexical Relation Classification (EVALution)
type: classification
dataset:
name: BLESS
args: relbert/lexical_relation_classification
type: relation-classification
metrics:
- name: F1
type: f1
value: 0.676056338028169
- name: F1 (macro)
type: f1_macro
value: 0.6699220665498732
- task:
name: Lexical Relation Classification (K&H+N)
type: classification
dataset:
name: K&H+N
args: relbert/lexical_relation_classification
type: relation-classification
metrics:
- name: F1
type: f1
value: 0.9604228976838005
- name: F1 (macro)
type: f1_macro
value: 0.8725502582807458
- task:
name: Lexical Relation Classification (ROOT09)
type: classification
dataset:
name: ROOT09
args: relbert/lexical_relation_classification
type: relation-classification
metrics:
- name: F1
type: f1
value: 0.8865559385772485
- name: F1 (macro)
type: f1_macro
value: 0.8814062245146053
---
# relbert/roberta-large-semeval2012-average-prompt-d-nce-classification-conceptnet-validated
RelBERT fine-tuned from [roberta-large](https://huggingface.co/roberta-large) on
[relbert/semeval2012_relational_similarity](https://huggingface.co/datasets/relbert/semeval2012_relational_similarity).
Fine-tuning is done via [RelBERT](https://github.com/asahi417/relbert) library (see the repository for more detail).
It achieves the following results on the relation understanding tasks:
- Analogy Question ([dataset](https://huggingface.co/datasets/relbert/analogy_questions), [full result](https://huggingface.co/relbert/roberta-large-semeval2012-average-prompt-d-nce-classification-conceptnet-validated/raw/main/analogy.json)):
- Accuracy on SAT (full): 0.5828877005347594
- Accuracy on SAT: 0.6023738872403561
- Accuracy on BATS: 0.6170094496942746
- Accuracy on U2: 0.5219298245614035
- Accuracy on U4: 0.5347222222222222
- Accuracy on Google: 0.842
- Lexical Relation Classification ([dataset](https://huggingface.co/datasets/relbert/lexical_relation_classification), [full result](https://huggingface.co/relbert/roberta-large-semeval2012-average-prompt-d-nce-classification-conceptnet-validated/raw/main/classification.json)):
- Micro F1 score on BLESS: 0.9127617899653457
- Micro F1 score on CogALexV: 0.8523474178403756
- Micro F1 score on EVALution: 0.676056338028169
- Micro F1 score on K&H+N: 0.9604228976838005
- Micro F1 score on ROOT09: 0.8865559385772485
- Relation Mapping ([dataset](https://huggingface.co/datasets/relbert/relation_mapping), [full result](https://huggingface.co/relbert/roberta-large-semeval2012-average-prompt-d-nce-classification-conceptnet-validated/raw/main/relation_mapping.json)):
- Accuracy on Relation Mapping: 0.8295436507936508
### Usage
This model can be used through the [relbert library](https://github.com/asahi417/relbert). Install the library via pip
```shell
pip install relbert
```
and activate model as below.
```python
from relbert import RelBERT
model = RelBERT("relbert/roberta-large-semeval2012-average-prompt-d-nce-classification-conceptnet-validated")
vector = model.get_embedding(['Tokyo', 'Japan']) # shape of (1024, )
```
### Training hyperparameters
The following hyperparameters were used during training:
- model: roberta-large
- max_length: 64
- mode: average
- data: relbert/semeval2012_relational_similarity
- split: train
- data_eval: relbert/conceptnet_high_confidence
- split_eval: full
- template_mode: manual
- template: I wasn’t aware of this relationship, but I just read in the encyclopedia that <subj> is the <mask> of <obj>
- loss_function: nce_logout
- classification_loss: True
- temperature_nce_constant: 0.05
- temperature_nce_rank: {'min': 0.01, 'max': 0.05, 'type': 'linear'}
- epoch: 30
- batch: 128
- lr: 5e-06
- lr_decay: False
- lr_warmup: 1
- weight_decay: 0
- random_seed: 0
- exclude_relation: None
- exclude_relation_eval: None
- n_sample: 640
- gradient_accumulation: 8
The full configuration can be found at [fine-tuning parameter file](https://huggingface.co/relbert/roberta-large-semeval2012-average-prompt-d-nce-classification-conceptnet-validated/raw/main/trainer_config.json).
### Reference
If you use any resource from RelBERT, please consider to cite our [paper](https://aclanthology.org/2021.eacl-demos.7/).
```
@inproceedings{ushio-etal-2021-distilling-relation-embeddings,
title = "{D}istilling {R}elation {E}mbeddings from {P}re-trained {L}anguage {M}odels",
author = "Ushio, Asahi and
Schockaert, Steven and
Camacho-Collados, Jose",
booktitle = "EMNLP 2021",
year = "2021",
address = "Online",
publisher = "Association for Computational Linguistics",
}
```
|
research-backup/roberta-large-semeval2012-average-no-mask-prompt-e-loob | research-backup | 2022-09-19T19:51:19Z | 105 | 0 | transformers | [
"transformers",
"pytorch",
"roberta",
"feature-extraction",
"dataset:relbert/semeval2012_relational_similarity",
"model-index",
"text-embeddings-inference",
"endpoints_compatible",
"region:us"
]
| feature-extraction | 2022-08-30T14:20:05Z | ---
datasets:
- relbert/semeval2012_relational_similarity
model-index:
- name: relbert/roberta-large-semeval2012-average-no-mask-prompt-e-loob
results:
- task:
name: Relation Mapping
type: sorting-task
dataset:
name: Relation Mapping
args: relbert/relation_mapping
type: relation-mapping
metrics:
- name: Accuracy
type: accuracy
value: 0.9032142857142857
- task:
name: Analogy Questions (SAT full)
type: multiple-choice-qa
dataset:
name: SAT full
args: relbert/analogy_questions
type: analogy-questions
metrics:
- name: Accuracy
type: accuracy
value: 0.5721925133689839
- task:
name: Analogy Questions (SAT)
type: multiple-choice-qa
dataset:
name: SAT
args: relbert/analogy_questions
type: analogy-questions
metrics:
- name: Accuracy
type: accuracy
value: 0.5667655786350149
- task:
name: Analogy Questions (BATS)
type: multiple-choice-qa
dataset:
name: BATS
args: relbert/analogy_questions
type: analogy-questions
metrics:
- name: Accuracy
type: accuracy
value: 0.7776542523624236
- task:
name: Analogy Questions (Google)
type: multiple-choice-qa
dataset:
name: Google
args: relbert/analogy_questions
type: analogy-questions
metrics:
- name: Accuracy
type: accuracy
value: 0.872
- task:
name: Analogy Questions (U2)
type: multiple-choice-qa
dataset:
name: U2
args: relbert/analogy_questions
type: analogy-questions
metrics:
- name: Accuracy
type: accuracy
value: 0.5833333333333334
- task:
name: Analogy Questions (U4)
type: multiple-choice-qa
dataset:
name: U4
args: relbert/analogy_questions
type: analogy-questions
metrics:
- name: Accuracy
type: accuracy
value: 0.6203703703703703
- task:
name: Lexical Relation Classification (BLESS)
type: classification
dataset:
name: BLESS
args: relbert/lexical_relation_classification
type: relation-classification
metrics:
- name: F1
type: f1
value: 0.9202953141479584
- name: F1 (macro)
type: f1_macro
value: 0.9155901755002147
- task:
name: Lexical Relation Classification (CogALexV)
type: classification
dataset:
name: CogALexV
args: relbert/lexical_relation_classification
type: relation-classification
metrics:
- name: F1
type: f1
value: 0.8467136150234742
- name: F1 (macro)
type: f1_macro
value: 0.6838421887453545
- task:
name: Lexical Relation Classification (EVALution)
type: classification
dataset:
name: BLESS
args: relbert/lexical_relation_classification
type: relation-classification
metrics:
- name: F1
type: f1
value: 0.6836403033586133
- name: F1 (macro)
type: f1_macro
value: 0.6705678270928033
- task:
name: Lexical Relation Classification (K&H+N)
type: classification
dataset:
name: K&H+N
args: relbert/lexical_relation_classification
type: relation-classification
metrics:
- name: F1
type: f1
value: 0.9625095638867636
- name: F1 (macro)
type: f1_macro
value: 0.8774656452359669
- task:
name: Lexical Relation Classification (ROOT09)
type: classification
dataset:
name: ROOT09
args: relbert/lexical_relation_classification
type: relation-classification
metrics:
- name: F1
type: f1
value: 0.9009714822939517
- name: F1 (macro)
type: f1_macro
value: 0.8985547104456186
---
# relbert/roberta-large-semeval2012-average-no-mask-prompt-e-loob
RelBERT fine-tuned from [roberta-large](https://huggingface.co/roberta-large) on
[relbert/semeval2012_relational_similarity](https://huggingface.co/datasets/relbert/semeval2012_relational_similarity).
Fine-tuning is done via [RelBERT](https://github.com/asahi417/relbert) library (see the repository for more detail).
It achieves the following results on the relation understanding tasks:
- Analogy Question ([dataset](https://huggingface.co/datasets/relbert/analogy_questions), [full result](https://huggingface.co/relbert/roberta-large-semeval2012-average-no-mask-prompt-e-loob/raw/main/analogy.json)):
- Accuracy on SAT (full): 0.5721925133689839
- Accuracy on SAT: 0.5667655786350149
- Accuracy on BATS: 0.7776542523624236
- Accuracy on U2: 0.5833333333333334
- Accuracy on U4: 0.6203703703703703
- Accuracy on Google: 0.872
- Lexical Relation Classification ([dataset](https://huggingface.co/datasets/relbert/lexical_relation_classification), [full result](https://huggingface.co/relbert/roberta-large-semeval2012-average-no-mask-prompt-e-loob/raw/main/classification.json)):
- Micro F1 score on BLESS: 0.9202953141479584
- Micro F1 score on CogALexV: 0.8467136150234742
- Micro F1 score on EVALution: 0.6836403033586133
- Micro F1 score on K&H+N: 0.9625095638867636
- Micro F1 score on ROOT09: 0.9009714822939517
- Relation Mapping ([dataset](https://huggingface.co/datasets/relbert/relation_mapping), [full result](https://huggingface.co/relbert/roberta-large-semeval2012-average-no-mask-prompt-e-loob/raw/main/relation_mapping.json)):
- Accuracy on Relation Mapping: 0.9032142857142857
### Usage
This model can be used through the [relbert library](https://github.com/asahi417/relbert). Install the library via pip
```shell
pip install relbert
```
and activate model as below.
```python
from relbert import RelBERT
model = RelBERT("relbert/roberta-large-semeval2012-average-no-mask-prompt-e-loob")
vector = model.get_embedding(['Tokyo', 'Japan']) # shape of (1024, )
```
### Training hyperparameters
The following hyperparameters were used during training:
- model: roberta-large
- max_length: 64
- mode: average_no_mask
- data: relbert/semeval2012_relational_similarity
- template_mode: manual
- template: I wasn’t aware of this relationship, but I just read in the encyclopedia that <obj> is <subj>’s <mask>
- loss_function: info_loob
- temperature_nce_constant: 0.05
- temperature_nce_rank: {'min': 0.01, 'max': 0.05, 'type': 'linear'}
- epoch: 22
- batch: 128
- lr: 5e-06
- lr_decay: False
- lr_warmup: 1
- weight_decay: 0
- random_seed: 0
- exclude_relation: None
- n_sample: 640
- gradient_accumulation: 8
The full configuration can be found at [fine-tuning parameter file](https://huggingface.co/relbert/roberta-large-semeval2012-average-no-mask-prompt-e-loob/raw/main/trainer_config.json).
### Reference
If you use any resource from RelBERT, please consider to cite our [paper](https://aclanthology.org/2021.eacl-demos.7/).
```
@inproceedings{ushio-etal-2021-distilling-relation-embeddings,
title = "{D}istilling {R}elation {E}mbeddings from {P}re-trained {L}anguage {M}odels",
author = "Ushio, Asahi and
Schockaert, Steven and
Camacho-Collados, Jose",
booktitle = "EMNLP 2021",
year = "2021",
address = "Online",
publisher = "Association for Computational Linguistics",
}
```
|
research-backup/roberta-large-semeval2012-average-no-mask-prompt-d-loob | research-backup | 2022-09-19T19:47:44Z | 105 | 0 | transformers | [
"transformers",
"pytorch",
"roberta",
"feature-extraction",
"dataset:relbert/semeval2012_relational_similarity",
"model-index",
"text-embeddings-inference",
"endpoints_compatible",
"region:us"
]
| feature-extraction | 2022-08-30T06:57:12Z | ---
datasets:
- relbert/semeval2012_relational_similarity
model-index:
- name: relbert/roberta-large-semeval2012-average-no-mask-prompt-d-loob
results:
- task:
name: Relation Mapping
type: sorting-task
dataset:
name: Relation Mapping
args: relbert/relation_mapping
type: relation-mapping
metrics:
- name: Accuracy
type: accuracy
value: 0.8871031746031746
- task:
name: Analogy Questions (SAT full)
type: multiple-choice-qa
dataset:
name: SAT full
args: relbert/analogy_questions
type: analogy-questions
metrics:
- name: Accuracy
type: accuracy
value: 0.6871657754010695
- task:
name: Analogy Questions (SAT)
type: multiple-choice-qa
dataset:
name: SAT
args: relbert/analogy_questions
type: analogy-questions
metrics:
- name: Accuracy
type: accuracy
value: 0.6913946587537092
- task:
name: Analogy Questions (BATS)
type: multiple-choice-qa
dataset:
name: BATS
args: relbert/analogy_questions
type: analogy-questions
metrics:
- name: Accuracy
type: accuracy
value: 0.8148971650917176
- task:
name: Analogy Questions (Google)
type: multiple-choice-qa
dataset:
name: Google
args: relbert/analogy_questions
type: analogy-questions
metrics:
- name: Accuracy
type: accuracy
value: 0.958
- task:
name: Analogy Questions (U2)
type: multiple-choice-qa
dataset:
name: U2
args: relbert/analogy_questions
type: analogy-questions
metrics:
- name: Accuracy
type: accuracy
value: 0.6359649122807017
- task:
name: Analogy Questions (U4)
type: multiple-choice-qa
dataset:
name: U4
args: relbert/analogy_questions
type: analogy-questions
metrics:
- name: Accuracy
type: accuracy
value: 0.6458333333333334
- task:
name: Lexical Relation Classification (BLESS)
type: classification
dataset:
name: BLESS
args: relbert/lexical_relation_classification
type: relation-classification
metrics:
- name: F1
type: f1
value: 0.9153231881874341
- name: F1 (macro)
type: f1_macro
value: 0.909786964934943
- task:
name: Lexical Relation Classification (CogALexV)
type: classification
dataset:
name: CogALexV
args: relbert/lexical_relation_classification
type: relation-classification
metrics:
- name: F1
type: f1
value: 0.8577464788732394
- name: F1 (macro)
type: f1_macro
value: 0.6952254602767576
- task:
name: Lexical Relation Classification (EVALution)
type: classification
dataset:
name: BLESS
args: relbert/lexical_relation_classification
type: relation-classification
metrics:
- name: F1
type: f1
value: 0.6847237269772481
- name: F1 (macro)
type: f1_macro
value: 0.6742659270266346
- task:
name: Lexical Relation Classification (K&H+N)
type: classification
dataset:
name: K&H+N
args: relbert/lexical_relation_classification
type: relation-classification
metrics:
- name: F1
type: f1
value: 0.9634137859080476
- name: F1 (macro)
type: f1_macro
value: 0.8926357349234371
- task:
name: Lexical Relation Classification (ROOT09)
type: classification
dataset:
name: ROOT09
args: relbert/lexical_relation_classification
type: relation-classification
metrics:
- name: F1
type: f1
value: 0.9106863052334692
- name: F1 (macro)
type: f1_macro
value: 0.9093125585829993
---
# relbert/roberta-large-semeval2012-average-no-mask-prompt-d-loob
RelBERT fine-tuned from [roberta-large](https://huggingface.co/roberta-large) on
[relbert/semeval2012_relational_similarity](https://huggingface.co/datasets/relbert/semeval2012_relational_similarity).
Fine-tuning is done via [RelBERT](https://github.com/asahi417/relbert) library (see the repository for more detail).
It achieves the following results on the relation understanding tasks:
- Analogy Question ([dataset](https://huggingface.co/datasets/relbert/analogy_questions), [full result](https://huggingface.co/relbert/roberta-large-semeval2012-average-no-mask-prompt-d-loob/raw/main/analogy.json)):
- Accuracy on SAT (full): 0.6871657754010695
- Accuracy on SAT: 0.6913946587537092
- Accuracy on BATS: 0.8148971650917176
- Accuracy on U2: 0.6359649122807017
- Accuracy on U4: 0.6458333333333334
- Accuracy on Google: 0.958
- Lexical Relation Classification ([dataset](https://huggingface.co/datasets/relbert/lexical_relation_classification), [full result](https://huggingface.co/relbert/roberta-large-semeval2012-average-no-mask-prompt-d-loob/raw/main/classification.json)):
- Micro F1 score on BLESS: 0.9153231881874341
- Micro F1 score on CogALexV: 0.8577464788732394
- Micro F1 score on EVALution: 0.6847237269772481
- Micro F1 score on K&H+N: 0.9634137859080476
- Micro F1 score on ROOT09: 0.9106863052334692
- Relation Mapping ([dataset](https://huggingface.co/datasets/relbert/relation_mapping), [full result](https://huggingface.co/relbert/roberta-large-semeval2012-average-no-mask-prompt-d-loob/raw/main/relation_mapping.json)):
- Accuracy on Relation Mapping: 0.8871031746031746
### Usage
This model can be used through the [relbert library](https://github.com/asahi417/relbert). Install the library via pip
```shell
pip install relbert
```
and activate model as below.
```python
from relbert import RelBERT
model = RelBERT("relbert/roberta-large-semeval2012-average-no-mask-prompt-d-loob")
vector = model.get_embedding(['Tokyo', 'Japan']) # shape of (1024, )
```
### Training hyperparameters
The following hyperparameters were used during training:
- model: roberta-large
- max_length: 64
- mode: average_no_mask
- data: relbert/semeval2012_relational_similarity
- template_mode: manual
- template: I wasn’t aware of this relationship, but I just read in the encyclopedia that <subj> is the <mask> of <obj>
- loss_function: info_loob
- temperature_nce_constant: 0.05
- temperature_nce_rank: {'min': 0.01, 'max': 0.05, 'type': 'linear'}
- epoch: 21
- batch: 128
- lr: 5e-06
- lr_decay: False
- lr_warmup: 1
- weight_decay: 0
- random_seed: 0
- exclude_relation: None
- n_sample: 640
- gradient_accumulation: 8
The full configuration can be found at [fine-tuning parameter file](https://huggingface.co/relbert/roberta-large-semeval2012-average-no-mask-prompt-d-loob/raw/main/trainer_config.json).
### Reference
If you use any resource from RelBERT, please consider to cite our [paper](https://aclanthology.org/2021.eacl-demos.7/).
```
@inproceedings{ushio-etal-2021-distilling-relation-embeddings,
title = "{D}istilling {R}elation {E}mbeddings from {P}re-trained {L}anguage {M}odels",
author = "Ushio, Asahi and
Schockaert, Steven and
Camacho-Collados, Jose",
booktitle = "EMNLP 2021",
year = "2021",
address = "Online",
publisher = "Association for Computational Linguistics",
}
```
|
research-backup/roberta-large-semeval2012-average-prompt-e-loob | research-backup | 2022-09-19T19:33:04Z | 101 | 0 | transformers | [
"transformers",
"pytorch",
"roberta",
"feature-extraction",
"dataset:relbert/semeval2012_relational_similarity",
"model-index",
"text-embeddings-inference",
"endpoints_compatible",
"region:us"
]
| feature-extraction | 2022-08-29T01:31:32Z | ---
datasets:
- relbert/semeval2012_relational_similarity
model-index:
- name: relbert/roberta-large-semeval2012-average-prompt-e-loob
results:
- task:
name: Relation Mapping
type: sorting-task
dataset:
name: Relation Mapping
args: relbert/relation_mapping
type: relation-mapping
metrics:
- name: Accuracy
type: accuracy
value: 0.9121031746031746
- task:
name: Analogy Questions (SAT full)
type: multiple-choice-qa
dataset:
name: SAT full
args: relbert/analogy_questions
type: analogy-questions
metrics:
- name: Accuracy
type: accuracy
value: 0.5909090909090909
- task:
name: Analogy Questions (SAT)
type: multiple-choice-qa
dataset:
name: SAT
args: relbert/analogy_questions
type: analogy-questions
metrics:
- name: Accuracy
type: accuracy
value: 0.5875370919881305
- task:
name: Analogy Questions (BATS)
type: multiple-choice-qa
dataset:
name: BATS
args: relbert/analogy_questions
type: analogy-questions
metrics:
- name: Accuracy
type: accuracy
value: 0.7670928293496387
- task:
name: Analogy Questions (Google)
type: multiple-choice-qa
dataset:
name: Google
args: relbert/analogy_questions
type: analogy-questions
metrics:
- name: Accuracy
type: accuracy
value: 0.912
- task:
name: Analogy Questions (U2)
type: multiple-choice-qa
dataset:
name: U2
args: relbert/analogy_questions
type: analogy-questions
metrics:
- name: Accuracy
type: accuracy
value: 0.5570175438596491
- task:
name: Analogy Questions (U4)
type: multiple-choice-qa
dataset:
name: U4
args: relbert/analogy_questions
type: analogy-questions
metrics:
- name: Accuracy
type: accuracy
value: 0.5879629629629629
- task:
name: Lexical Relation Classification (BLESS)
type: classification
dataset:
name: BLESS
args: relbert/lexical_relation_classification
type: relation-classification
metrics:
- name: F1
type: f1
value: 0.9207473255989151
- name: F1 (macro)
type: f1_macro
value: 0.9149001350257856
- task:
name: Lexical Relation Classification (CogALexV)
type: classification
dataset:
name: CogALexV
args: relbert/lexical_relation_classification
type: relation-classification
metrics:
- name: F1
type: f1
value: 0.8481220657276995
- name: F1 (macro)
type: f1_macro
value: 0.6824179529207882
- task:
name: Lexical Relation Classification (EVALution)
type: classification
dataset:
name: BLESS
args: relbert/lexical_relation_classification
type: relation-classification
metrics:
- name: F1
type: f1
value: 0.6798483206933911
- name: F1 (macro)
type: f1_macro
value: 0.6735513654805187
- task:
name: Lexical Relation Classification (K&H+N)
type: classification
dataset:
name: K&H+N
args: relbert/lexical_relation_classification
type: relation-classification
metrics:
- name: F1
type: f1
value: 0.9589622313417264
- name: F1 (macro)
type: f1_macro
value: 0.872950103232891
- task:
name: Lexical Relation Classification (ROOT09)
type: classification
dataset:
name: ROOT09
args: relbert/lexical_relation_classification
type: relation-classification
metrics:
- name: F1
type: f1
value: 0.895017235976183
- name: F1 (macro)
type: f1_macro
value: 0.8900982680408713
---
# relbert/roberta-large-semeval2012-average-prompt-e-loob
RelBERT fine-tuned from [roberta-large](https://huggingface.co/roberta-large) on
[relbert/semeval2012_relational_similarity](https://huggingface.co/datasets/relbert/semeval2012_relational_similarity).
Fine-tuning is done via [RelBERT](https://github.com/asahi417/relbert) library (see the repository for more detail).
It achieves the following results on the relation understanding tasks:
- Analogy Question ([dataset](https://huggingface.co/datasets/relbert/analogy_questions), [full result](https://huggingface.co/relbert/roberta-large-semeval2012-average-prompt-e-loob/raw/main/analogy.json)):
- Accuracy on SAT (full): 0.5909090909090909
- Accuracy on SAT: 0.5875370919881305
- Accuracy on BATS: 0.7670928293496387
- Accuracy on U2: 0.5570175438596491
- Accuracy on U4: 0.5879629629629629
- Accuracy on Google: 0.912
- Lexical Relation Classification ([dataset](https://huggingface.co/datasets/relbert/lexical_relation_classification), [full result](https://huggingface.co/relbert/roberta-large-semeval2012-average-prompt-e-loob/raw/main/classification.json)):
- Micro F1 score on BLESS: 0.9207473255989151
- Micro F1 score on CogALexV: 0.8481220657276995
- Micro F1 score on EVALution: 0.6798483206933911
- Micro F1 score on K&H+N: 0.9589622313417264
- Micro F1 score on ROOT09: 0.895017235976183
- Relation Mapping ([dataset](https://huggingface.co/datasets/relbert/relation_mapping), [full result](https://huggingface.co/relbert/roberta-large-semeval2012-average-prompt-e-loob/raw/main/relation_mapping.json)):
- Accuracy on Relation Mapping: 0.9121031746031746
### Usage
This model can be used through the [relbert library](https://github.com/asahi417/relbert). Install the library via pip
```shell
pip install relbert
```
and activate model as below.
```python
from relbert import RelBERT
model = RelBERT("relbert/roberta-large-semeval2012-average-prompt-e-loob")
vector = model.get_embedding(['Tokyo', 'Japan']) # shape of (1024, )
```
### Training hyperparameters
The following hyperparameters were used during training:
- model: roberta-large
- max_length: 64
- mode: average
- data: relbert/semeval2012_relational_similarity
- template_mode: manual
- template: I wasn’t aware of this relationship, but I just read in the encyclopedia that <obj> is <subj>’s <mask>
- loss_function: info_loob
- temperature_nce_constant: 0.05
- temperature_nce_rank: {'min': 0.01, 'max': 0.05, 'type': 'linear'}
- epoch: 22
- batch: 128
- lr: 5e-06
- lr_decay: False
- lr_warmup: 1
- weight_decay: 0
- random_seed: 0
- exclude_relation: None
- n_sample: 640
- gradient_accumulation: 8
The full configuration can be found at [fine-tuning parameter file](https://huggingface.co/relbert/roberta-large-semeval2012-average-prompt-e-loob/raw/main/trainer_config.json).
### Reference
If you use any resource from RelBERT, please consider to cite our [paper](https://aclanthology.org/2021.eacl-demos.7/).
```
@inproceedings{ushio-etal-2021-distilling-relation-embeddings,
title = "{D}istilling {R}elation {E}mbeddings from {P}re-trained {L}anguage {M}odels",
author = "Ushio, Asahi and
Schockaert, Steven and
Camacho-Collados, Jose",
booktitle = "EMNLP 2021",
year = "2021",
address = "Online",
publisher = "Association for Computational Linguistics",
}
```
|
research-backup/roberta-large-semeval2012-average-prompt-d-loob | research-backup | 2022-09-19T19:29:19Z | 104 | 0 | transformers | [
"transformers",
"pytorch",
"roberta",
"feature-extraction",
"dataset:relbert/semeval2012_relational_similarity",
"model-index",
"text-embeddings-inference",
"endpoints_compatible",
"region:us"
]
| feature-extraction | 2022-08-28T18:08:47Z | ---
datasets:
- relbert/semeval2012_relational_similarity
model-index:
- name: relbert/roberta-large-semeval2012-average-prompt-d-loob
results:
- task:
name: Relation Mapping
type: sorting-task
dataset:
name: Relation Mapping
args: relbert/relation_mapping
type: relation-mapping
metrics:
- name: Accuracy
type: accuracy
value: 0.8432936507936508
- task:
name: Analogy Questions (SAT full)
type: multiple-choice-qa
dataset:
name: SAT full
args: relbert/analogy_questions
type: analogy-questions
metrics:
- name: Accuracy
type: accuracy
value: 0.7032085561497327
- task:
name: Analogy Questions (SAT)
type: multiple-choice-qa
dataset:
name: SAT
args: relbert/analogy_questions
type: analogy-questions
metrics:
- name: Accuracy
type: accuracy
value: 0.7091988130563798
- task:
name: Analogy Questions (BATS)
type: multiple-choice-qa
dataset:
name: BATS
args: relbert/analogy_questions
type: analogy-questions
metrics:
- name: Accuracy
type: accuracy
value: 0.8182323513062812
- task:
name: Analogy Questions (Google)
type: multiple-choice-qa
dataset:
name: Google
args: relbert/analogy_questions
type: analogy-questions
metrics:
- name: Accuracy
type: accuracy
value: 0.962
- task:
name: Analogy Questions (U2)
type: multiple-choice-qa
dataset:
name: U2
args: relbert/analogy_questions
type: analogy-questions
metrics:
- name: Accuracy
type: accuracy
value: 0.6535087719298246
- task:
name: Analogy Questions (U4)
type: multiple-choice-qa
dataset:
name: U4
args: relbert/analogy_questions
type: analogy-questions
metrics:
- name: Accuracy
type: accuracy
value: 0.6342592592592593
- task:
name: Lexical Relation Classification (BLESS)
type: classification
dataset:
name: BLESS
args: relbert/lexical_relation_classification
type: relation-classification
metrics:
- name: F1
type: f1
value: 0.9154738586710863
- name: F1 (macro)
type: f1_macro
value: 0.9105308478206379
- task:
name: Lexical Relation Classification (CogALexV)
type: classification
dataset:
name: CogALexV
args: relbert/lexical_relation_classification
type: relation-classification
metrics:
- name: F1
type: f1
value: 0.8652582159624412
- name: F1 (macro)
type: f1_macro
value: 0.7157465075284571
- task:
name: Lexical Relation Classification (EVALution)
type: classification
dataset:
name: BLESS
args: relbert/lexical_relation_classification
type: relation-classification
metrics:
- name: F1
type: f1
value: 0.6841820151679306
- name: F1 (macro)
type: f1_macro
value: 0.6652440461492628
- task:
name: Lexical Relation Classification (K&H+N)
type: classification
dataset:
name: K&H+N
args: relbert/lexical_relation_classification
type: relation-classification
metrics:
- name: F1
type: f1
value: 0.9582666759407387
- name: F1 (macro)
type: f1_macro
value: 0.8705160523996387
- task:
name: Lexical Relation Classification (ROOT09)
type: classification
dataset:
name: ROOT09
args: relbert/lexical_relation_classification
type: relation-classification
metrics:
- name: F1
type: f1
value: 0.9078658727671576
- name: F1 (macro)
type: f1_macro
value: 0.9051927463291504
---
# relbert/roberta-large-semeval2012-average-prompt-d-loob
RelBERT fine-tuned from [roberta-large](https://huggingface.co/roberta-large) on
[relbert/semeval2012_relational_similarity](https://huggingface.co/datasets/relbert/semeval2012_relational_similarity).
Fine-tuning is done via [RelBERT](https://github.com/asahi417/relbert) library (see the repository for more detail).
It achieves the following results on the relation understanding tasks:
- Analogy Question ([dataset](https://huggingface.co/datasets/relbert/analogy_questions), [full result](https://huggingface.co/relbert/roberta-large-semeval2012-average-prompt-d-loob/raw/main/analogy.json)):
- Accuracy on SAT (full): 0.7032085561497327
- Accuracy on SAT: 0.7091988130563798
- Accuracy on BATS: 0.8182323513062812
- Accuracy on U2: 0.6535087719298246
- Accuracy on U4: 0.6342592592592593
- Accuracy on Google: 0.962
- Lexical Relation Classification ([dataset](https://huggingface.co/datasets/relbert/lexical_relation_classification), [full result](https://huggingface.co/relbert/roberta-large-semeval2012-average-prompt-d-loob/raw/main/classification.json)):
- Micro F1 score on BLESS: 0.9154738586710863
- Micro F1 score on CogALexV: 0.8652582159624412
- Micro F1 score on EVALution: 0.6841820151679306
- Micro F1 score on K&H+N: 0.9582666759407387
- Micro F1 score on ROOT09: 0.9078658727671576
- Relation Mapping ([dataset](https://huggingface.co/datasets/relbert/relation_mapping), [full result](https://huggingface.co/relbert/roberta-large-semeval2012-average-prompt-d-loob/raw/main/relation_mapping.json)):
- Accuracy on Relation Mapping: 0.8432936507936508
### Usage
This model can be used through the [relbert library](https://github.com/asahi417/relbert). Install the library via pip
```shell
pip install relbert
```
and activate model as below.
```python
from relbert import RelBERT
model = RelBERT("relbert/roberta-large-semeval2012-average-prompt-d-loob")
vector = model.get_embedding(['Tokyo', 'Japan']) # shape of (1024, )
```
### Training hyperparameters
The following hyperparameters were used during training:
- model: roberta-large
- max_length: 64
- mode: average
- data: relbert/semeval2012_relational_similarity
- template_mode: manual
- template: I wasn’t aware of this relationship, but I just read in the encyclopedia that <subj> is the <mask> of <obj>
- loss_function: info_loob
- temperature_nce_constant: 0.05
- temperature_nce_rank: {'min': 0.01, 'max': 0.05, 'type': 'linear'}
- epoch: 22
- batch: 128
- lr: 5e-06
- lr_decay: False
- lr_warmup: 1
- weight_decay: 0
- random_seed: 0
- exclude_relation: None
- n_sample: 640
- gradient_accumulation: 8
The full configuration can be found at [fine-tuning parameter file](https://huggingface.co/relbert/roberta-large-semeval2012-average-prompt-d-loob/raw/main/trainer_config.json).
### Reference
If you use any resource from RelBERT, please consider to cite our [paper](https://aclanthology.org/2021.eacl-demos.7/).
```
@inproceedings{ushio-etal-2021-distilling-relation-embeddings,
title = "{D}istilling {R}elation {E}mbeddings from {P}re-trained {L}anguage {M}odels",
author = "Ushio, Asahi and
Schockaert, Steven and
Camacho-Collados, Jose",
booktitle = "EMNLP 2021",
year = "2021",
address = "Online",
publisher = "Association for Computational Linguistics",
}
```
|
research-backup/roberta-large-semeval2012-average-prompt-a-loob | research-backup | 2022-09-19T19:18:16Z | 107 | 0 | transformers | [
"transformers",
"pytorch",
"roberta",
"feature-extraction",
"dataset:relbert/semeval2012_relational_similarity",
"model-index",
"text-embeddings-inference",
"endpoints_compatible",
"region:us"
]
| feature-extraction | 2022-08-27T20:04:16Z | ---
datasets:
- relbert/semeval2012_relational_similarity
model-index:
- name: relbert/roberta-large-semeval2012-average-prompt-a-loob
results:
- task:
name: Relation Mapping
type: sorting-task
dataset:
name: Relation Mapping
args: relbert/relation_mapping
type: relation-mapping
metrics:
- name: Accuracy
type: accuracy
value: 0.8641666666666666
- task:
name: Analogy Questions (SAT full)
type: multiple-choice-qa
dataset:
name: SAT full
args: relbert/analogy_questions
type: analogy-questions
metrics:
- name: Accuracy
type: accuracy
value: 0.6443850267379679
- task:
name: Analogy Questions (SAT)
type: multiple-choice-qa
dataset:
name: SAT
args: relbert/analogy_questions
type: analogy-questions
metrics:
- name: Accuracy
type: accuracy
value: 0.6468842729970327
- task:
name: Analogy Questions (BATS)
type: multiple-choice-qa
dataset:
name: BATS
args: relbert/analogy_questions
type: analogy-questions
metrics:
- name: Accuracy
type: accuracy
value: 0.7137298499166204
- task:
name: Analogy Questions (Google)
type: multiple-choice-qa
dataset:
name: Google
args: relbert/analogy_questions
type: analogy-questions
metrics:
- name: Accuracy
type: accuracy
value: 0.898
- task:
name: Analogy Questions (U2)
type: multiple-choice-qa
dataset:
name: U2
args: relbert/analogy_questions
type: analogy-questions
metrics:
- name: Accuracy
type: accuracy
value: 0.543859649122807
- task:
name: Analogy Questions (U4)
type: multiple-choice-qa
dataset:
name: U4
args: relbert/analogy_questions
type: analogy-questions
metrics:
- name: Accuracy
type: accuracy
value: 0.5833333333333334
- task:
name: Lexical Relation Classification (BLESS)
type: classification
dataset:
name: BLESS
args: relbert/lexical_relation_classification
type: relation-classification
metrics:
- name: F1
type: f1
value: 0.9153231881874341
- name: F1 (macro)
type: f1_macro
value: 0.910194305368961
- task:
name: Lexical Relation Classification (CogALexV)
type: classification
dataset:
name: CogALexV
args: relbert/lexical_relation_classification
type: relation-classification
metrics:
- name: F1
type: f1
value: 0.854225352112676
- name: F1 (macro)
type: f1_macro
value: 0.6939611644499436
- task:
name: Lexical Relation Classification (EVALution)
type: classification
dataset:
name: BLESS
args: relbert/lexical_relation_classification
type: relation-classification
metrics:
- name: F1
type: f1
value: 0.6603466955579632
- name: F1 (macro)
type: f1_macro
value: 0.6449027403702262
- task:
name: Lexical Relation Classification (K&H+N)
type: classification
dataset:
name: K&H+N
args: relbert/lexical_relation_classification
type: relation-classification
metrics:
- name: F1
type: f1
value: 0.9617444529456771
- name: F1 (macro)
type: f1_macro
value: 0.8891323512830197
- task:
name: Lexical Relation Classification (ROOT09)
type: classification
dataset:
name: ROOT09
args: relbert/lexical_relation_classification
type: relation-classification
metrics:
- name: F1
type: f1
value: 0.902851770604826
- name: F1 (macro)
type: f1_macro
value: 0.9021609534307928
---
# relbert/roberta-large-semeval2012-average-prompt-a-loob
RelBERT fine-tuned from [roberta-large](https://huggingface.co/roberta-large) on
[relbert/semeval2012_relational_similarity](https://huggingface.co/datasets/relbert/semeval2012_relational_similarity).
Fine-tuning is done via [RelBERT](https://github.com/asahi417/relbert) library (see the repository for more detail).
It achieves the following results on the relation understanding tasks:
- Analogy Question ([dataset](https://huggingface.co/datasets/relbert/analogy_questions), [full result](https://huggingface.co/relbert/roberta-large-semeval2012-average-prompt-a-loob/raw/main/analogy.json)):
- Accuracy on SAT (full): 0.6443850267379679
- Accuracy on SAT: 0.6468842729970327
- Accuracy on BATS: 0.7137298499166204
- Accuracy on U2: 0.543859649122807
- Accuracy on U4: 0.5833333333333334
- Accuracy on Google: 0.898
- Lexical Relation Classification ([dataset](https://huggingface.co/datasets/relbert/lexical_relation_classification), [full result](https://huggingface.co/relbert/roberta-large-semeval2012-average-prompt-a-loob/raw/main/classification.json)):
- Micro F1 score on BLESS: 0.9153231881874341
- Micro F1 score on CogALexV: 0.854225352112676
- Micro F1 score on EVALution: 0.6603466955579632
- Micro F1 score on K&H+N: 0.9617444529456771
- Micro F1 score on ROOT09: 0.902851770604826
- Relation Mapping ([dataset](https://huggingface.co/datasets/relbert/relation_mapping), [full result](https://huggingface.co/relbert/roberta-large-semeval2012-average-prompt-a-loob/raw/main/relation_mapping.json)):
- Accuracy on Relation Mapping: 0.8641666666666666
### Usage
This model can be used through the [relbert library](https://github.com/asahi417/relbert). Install the library via pip
```shell
pip install relbert
```
and activate model as below.
```python
from relbert import RelBERT
model = RelBERT("relbert/roberta-large-semeval2012-average-prompt-a-loob")
vector = model.get_embedding(['Tokyo', 'Japan']) # shape of (1024, )
```
### Training hyperparameters
The following hyperparameters were used during training:
- model: roberta-large
- max_length: 64
- mode: average
- data: relbert/semeval2012_relational_similarity
- template_mode: manual
- template: Today, I finally discovered the relation between <subj> and <obj> : <subj> is the <mask> of <obj>
- loss_function: info_loob
- temperature_nce_constant: 0.05
- temperature_nce_rank: {'min': 0.01, 'max': 0.05, 'type': 'linear'}
- epoch: 22
- batch: 128
- lr: 5e-06
- lr_decay: False
- lr_warmup: 1
- weight_decay: 0
- random_seed: 0
- exclude_relation: None
- n_sample: 640
- gradient_accumulation: 8
The full configuration can be found at [fine-tuning parameter file](https://huggingface.co/relbert/roberta-large-semeval2012-average-prompt-a-loob/raw/main/trainer_config.json).
### Reference
If you use any resource from RelBERT, please consider to cite our [paper](https://aclanthology.org/2021.eacl-demos.7/).
```
@inproceedings{ushio-etal-2021-distilling-relation-embeddings,
title = "{D}istilling {R}elation {E}mbeddings from {P}re-trained {L}anguage {M}odels",
author = "Ushio, Asahi and
Schockaert, Steven and
Camacho-Collados, Jose",
booktitle = "EMNLP 2021",
year = "2021",
address = "Online",
publisher = "Association for Computational Linguistics",
}
```
|
research-backup/roberta-large-semeval2012-mask-prompt-e-loob | research-backup | 2022-09-19T19:14:37Z | 107 | 0 | transformers | [
"transformers",
"pytorch",
"roberta",
"feature-extraction",
"dataset:relbert/semeval2012_relational_similarity",
"model-index",
"text-embeddings-inference",
"endpoints_compatible",
"region:us"
]
| feature-extraction | 2022-08-27T12:42:28Z | ---
datasets:
- relbert/semeval2012_relational_similarity
model-index:
- name: relbert/roberta-large-semeval2012-mask-prompt-e-loob
results:
- task:
name: Relation Mapping
type: sorting-task
dataset:
name: Relation Mapping
args: relbert/relation_mapping
type: relation-mapping
metrics:
- name: Accuracy
type: accuracy
value: 0.8682936507936508
- task:
name: Analogy Questions (SAT full)
type: multiple-choice-qa
dataset:
name: SAT full
args: relbert/analogy_questions
type: analogy-questions
metrics:
- name: Accuracy
type: accuracy
value: 0.6176470588235294
- task:
name: Analogy Questions (SAT)
type: multiple-choice-qa
dataset:
name: SAT
args: relbert/analogy_questions
type: analogy-questions
metrics:
- name: Accuracy
type: accuracy
value: 0.6231454005934718
- task:
name: Analogy Questions (BATS)
type: multiple-choice-qa
dataset:
name: BATS
args: relbert/analogy_questions
type: analogy-questions
metrics:
- name: Accuracy
type: accuracy
value: 0.7570872707059477
- task:
name: Analogy Questions (Google)
type: multiple-choice-qa
dataset:
name: Google
args: relbert/analogy_questions
type: analogy-questions
metrics:
- name: Accuracy
type: accuracy
value: 0.874
- task:
name: Analogy Questions (U2)
type: multiple-choice-qa
dataset:
name: U2
args: relbert/analogy_questions
type: analogy-questions
metrics:
- name: Accuracy
type: accuracy
value: 0.6008771929824561
- task:
name: Analogy Questions (U4)
type: multiple-choice-qa
dataset:
name: U4
args: relbert/analogy_questions
type: analogy-questions
metrics:
- name: Accuracy
type: accuracy
value: 0.6226851851851852
- task:
name: Lexical Relation Classification (BLESS)
type: classification
dataset:
name: BLESS
args: relbert/lexical_relation_classification
type: relation-classification
metrics:
- name: F1
type: f1
value: 0.9311435889709206
- name: F1 (macro)
type: f1_macro
value: 0.9268380061574883
- task:
name: Lexical Relation Classification (CogALexV)
type: classification
dataset:
name: CogALexV
args: relbert/lexical_relation_classification
type: relation-classification
metrics:
- name: F1
type: f1
value: 0.8744131455399061
- name: F1 (macro)
type: f1_macro
value: 0.7267491613759859
- task:
name: Lexical Relation Classification (EVALution)
type: classification
dataset:
name: BLESS
args: relbert/lexical_relation_classification
type: relation-classification
metrics:
- name: F1
type: f1
value: 0.7053087757313109
- name: F1 (macro)
type: f1_macro
value: 0.694918491135901
- task:
name: Lexical Relation Classification (K&H+N)
type: classification
dataset:
name: K&H+N
args: relbert/lexical_relation_classification
type: relation-classification
metrics:
- name: F1
type: f1
value: 0.9652222299506156
- name: F1 (macro)
type: f1_macro
value: 0.8967485289493923
- task:
name: Lexical Relation Classification (ROOT09)
type: classification
dataset:
name: ROOT09
args: relbert/lexical_relation_classification
type: relation-classification
metrics:
- name: F1
type: f1
value: 0.8965841429019116
- name: F1 (macro)
type: f1_macro
value: 0.8952392246946669
---
# relbert/roberta-large-semeval2012-mask-prompt-e-loob
RelBERT fine-tuned from [roberta-large](https://huggingface.co/roberta-large) on
[relbert/semeval2012_relational_similarity](https://huggingface.co/datasets/relbert/semeval2012_relational_similarity).
Fine-tuning is done via [RelBERT](https://github.com/asahi417/relbert) library (see the repository for more detail).
It achieves the following results on the relation understanding tasks:
- Analogy Question ([dataset](https://huggingface.co/datasets/relbert/analogy_questions), [full result](https://huggingface.co/relbert/roberta-large-semeval2012-mask-prompt-e-loob/raw/main/analogy.json)):
- Accuracy on SAT (full): 0.6176470588235294
- Accuracy on SAT: 0.6231454005934718
- Accuracy on BATS: 0.7570872707059477
- Accuracy on U2: 0.6008771929824561
- Accuracy on U4: 0.6226851851851852
- Accuracy on Google: 0.874
- Lexical Relation Classification ([dataset](https://huggingface.co/datasets/relbert/lexical_relation_classification), [full result](https://huggingface.co/relbert/roberta-large-semeval2012-mask-prompt-e-loob/raw/main/classification.json)):
- Micro F1 score on BLESS: 0.9311435889709206
- Micro F1 score on CogALexV: 0.8744131455399061
- Micro F1 score on EVALution: 0.7053087757313109
- Micro F1 score on K&H+N: 0.9652222299506156
- Micro F1 score on ROOT09: 0.8965841429019116
- Relation Mapping ([dataset](https://huggingface.co/datasets/relbert/relation_mapping), [full result](https://huggingface.co/relbert/roberta-large-semeval2012-mask-prompt-e-loob/raw/main/relation_mapping.json)):
- Accuracy on Relation Mapping: 0.8682936507936508
### Usage
This model can be used through the [relbert library](https://github.com/asahi417/relbert). Install the library via pip
```shell
pip install relbert
```
and activate model as below.
```python
from relbert import RelBERT
model = RelBERT("relbert/roberta-large-semeval2012-mask-prompt-e-loob")
vector = model.get_embedding(['Tokyo', 'Japan']) # shape of (1024, )
```
### Training hyperparameters
The following hyperparameters were used during training:
- model: roberta-large
- max_length: 64
- mode: mask
- data: relbert/semeval2012_relational_similarity
- template_mode: manual
- template: I wasn’t aware of this relationship, but I just read in the encyclopedia that <obj> is <subj>’s <mask>
- loss_function: info_loob
- temperature_nce_constant: 0.05
- temperature_nce_rank: {'min': 0.01, 'max': 0.05, 'type': 'linear'}
- epoch: 22
- batch: 128
- lr: 5e-06
- lr_decay: False
- lr_warmup: 1
- weight_decay: 0
- random_seed: 0
- exclude_relation: None
- n_sample: 640
- gradient_accumulation: 8
The full configuration can be found at [fine-tuning parameter file](https://huggingface.co/relbert/roberta-large-semeval2012-mask-prompt-e-loob/raw/main/trainer_config.json).
### Reference
If you use any resource from RelBERT, please consider to cite our [paper](https://aclanthology.org/2021.eacl-demos.7/).
```
@inproceedings{ushio-etal-2021-distilling-relation-embeddings,
title = "{D}istilling {R}elation {E}mbeddings from {P}re-trained {L}anguage {M}odels",
author = "Ushio, Asahi and
Schockaert, Steven and
Camacho-Collados, Jose",
booktitle = "EMNLP 2021",
year = "2021",
address = "Online",
publisher = "Association for Computational Linguistics",
}
```
|
research-backup/roberta-large-semeval2012-mask-prompt-a-loob | research-backup | 2022-09-19T18:59:55Z | 107 | 0 | transformers | [
"transformers",
"pytorch",
"roberta",
"feature-extraction",
"dataset:relbert/semeval2012_relational_similarity",
"model-index",
"text-embeddings-inference",
"endpoints_compatible",
"region:us"
]
| feature-extraction | 2022-08-26T06:59:38Z | ---
datasets:
- relbert/semeval2012_relational_similarity
model-index:
- name: relbert/roberta-large-semeval2012-mask-prompt-a-loob
results:
- task:
name: Relation Mapping
type: sorting-task
dataset:
name: Relation Mapping
args: relbert/relation_mapping
type: relation-mapping
metrics:
- name: Accuracy
type: accuracy
value: 0.9060317460317461
- task:
name: Analogy Questions (SAT full)
type: multiple-choice-qa
dataset:
name: SAT full
args: relbert/analogy_questions
type: analogy-questions
metrics:
- name: Accuracy
type: accuracy
value: 0.6550802139037433
- task:
name: Analogy Questions (SAT)
type: multiple-choice-qa
dataset:
name: SAT
args: relbert/analogy_questions
type: analogy-questions
metrics:
- name: Accuracy
type: accuracy
value: 0.655786350148368
- task:
name: Analogy Questions (BATS)
type: multiple-choice-qa
dataset:
name: BATS
args: relbert/analogy_questions
type: analogy-questions
metrics:
- name: Accuracy
type: accuracy
value: 0.8043357420789328
- task:
name: Analogy Questions (Google)
type: multiple-choice-qa
dataset:
name: Google
args: relbert/analogy_questions
type: analogy-questions
metrics:
- name: Accuracy
type: accuracy
value: 0.95
- task:
name: Analogy Questions (U2)
type: multiple-choice-qa
dataset:
name: U2
args: relbert/analogy_questions
type: analogy-questions
metrics:
- name: Accuracy
type: accuracy
value: 0.631578947368421
- task:
name: Analogy Questions (U4)
type: multiple-choice-qa
dataset:
name: U4
args: relbert/analogy_questions
type: analogy-questions
metrics:
- name: Accuracy
type: accuracy
value: 0.6412037037037037
- task:
name: Lexical Relation Classification (BLESS)
type: classification
dataset:
name: BLESS
args: relbert/lexical_relation_classification
type: relation-classification
metrics:
- name: F1
type: f1
value: 0.9245140876902215
- name: F1 (macro)
type: f1_macro
value: 0.9208294548760101
- task:
name: Lexical Relation Classification (CogALexV)
type: classification
dataset:
name: CogALexV
args: relbert/lexical_relation_classification
type: relation-classification
metrics:
- name: F1
type: f1
value: 0.8814553990610329
- name: F1 (macro)
type: f1_macro
value: 0.7355497663400952
- task:
name: Lexical Relation Classification (EVALution)
type: classification
dataset:
name: BLESS
args: relbert/lexical_relation_classification
type: relation-classification
metrics:
- name: F1
type: f1
value: 0.7128927410617552
- name: F1 (macro)
type: f1_macro
value: 0.7065924774146382
- task:
name: Lexical Relation Classification (K&H+N)
type: classification
dataset:
name: K&H+N
args: relbert/lexical_relation_classification
type: relation-classification
metrics:
- name: F1
type: f1
value: 0.9646657856298254
- name: F1 (macro)
type: f1_macro
value: 0.8945677578632619
- task:
name: Lexical Relation Classification (ROOT09)
type: classification
dataset:
name: ROOT09
args: relbert/lexical_relation_classification
type: relation-classification
metrics:
- name: F1
type: f1
value: 0.9081792541523034
- name: F1 (macro)
type: f1_macro
value: 0.906414518159255
---
# relbert/roberta-large-semeval2012-mask-prompt-a-loob
RelBERT fine-tuned from [roberta-large](https://huggingface.co/roberta-large) on
[relbert/semeval2012_relational_similarity](https://huggingface.co/datasets/relbert/semeval2012_relational_similarity).
Fine-tuning is done via [RelBERT](https://github.com/asahi417/relbert) library (see the repository for more detail).
It achieves the following results on the relation understanding tasks:
- Analogy Question ([dataset](https://huggingface.co/datasets/relbert/analogy_questions), [full result](https://huggingface.co/relbert/roberta-large-semeval2012-mask-prompt-a-loob/raw/main/analogy.json)):
- Accuracy on SAT (full): 0.6550802139037433
- Accuracy on SAT: 0.655786350148368
- Accuracy on BATS: 0.8043357420789328
- Accuracy on U2: 0.631578947368421
- Accuracy on U4: 0.6412037037037037
- Accuracy on Google: 0.95
- Lexical Relation Classification ([dataset](https://huggingface.co/datasets/relbert/lexical_relation_classification), [full result](https://huggingface.co/relbert/roberta-large-semeval2012-mask-prompt-a-loob/raw/main/classification.json)):
- Micro F1 score on BLESS: 0.9245140876902215
- Micro F1 score on CogALexV: 0.8814553990610329
- Micro F1 score on EVALution: 0.7128927410617552
- Micro F1 score on K&H+N: 0.9646657856298254
- Micro F1 score on ROOT09: 0.9081792541523034
- Relation Mapping ([dataset](https://huggingface.co/datasets/relbert/relation_mapping), [full result](https://huggingface.co/relbert/roberta-large-semeval2012-mask-prompt-a-loob/raw/main/relation_mapping.json)):
- Accuracy on Relation Mapping: 0.9060317460317461
### Usage
This model can be used through the [relbert library](https://github.com/asahi417/relbert). Install the library via pip
```shell
pip install relbert
```
and activate model as below.
```python
from relbert import RelBERT
model = RelBERT("relbert/roberta-large-semeval2012-mask-prompt-a-loob")
vector = model.get_embedding(['Tokyo', 'Japan']) # shape of (1024, )
```
### Training hyperparameters
The following hyperparameters were used during training:
- model: roberta-large
- max_length: 64
- mode: mask
- data: relbert/semeval2012_relational_similarity
- template_mode: manual
- template: Today, I finally discovered the relation between <subj> and <obj> : <subj> is the <mask> of <obj>
- loss_function: info_loob
- temperature_nce_constant: 0.05
- temperature_nce_rank: {'min': 0.01, 'max': 0.05, 'type': 'linear'}
- epoch: 21
- batch: 128
- lr: 5e-06
- lr_decay: False
- lr_warmup: 1
- weight_decay: 0
- random_seed: 0
- exclude_relation: None
- n_sample: 640
- gradient_accumulation: 8
The full configuration can be found at [fine-tuning parameter file](https://huggingface.co/relbert/roberta-large-semeval2012-mask-prompt-a-loob/raw/main/trainer_config.json).
### Reference
If you use any resource from RelBERT, please consider to cite our [paper](https://aclanthology.org/2021.eacl-demos.7/).
```
@inproceedings{ushio-etal-2021-distilling-relation-embeddings,
title = "{D}istilling {R}elation {E}mbeddings from {P}re-trained {L}anguage {M}odels",
author = "Ushio, Asahi and
Schockaert, Steven and
Camacho-Collados, Jose",
booktitle = "EMNLP 2021",
year = "2021",
address = "Online",
publisher = "Association for Computational Linguistics",
}
```
|
sd-concepts-library/ikea-fabler | sd-concepts-library | 2022-09-19T18:58:11Z | 0 | 0 | null | [
"license:mit",
"region:us"
]
| null | 2022-09-19T18:58:07Z | ---
license: mit
---
### ikea-fabler on Stable Diffusion
This is the `<ikea-fabler>` concept taught to Stable Diffusion via Textual Inversion. You can load this concept into the [Stable Conceptualizer](https://colab.research.google.com/github/huggingface/notebooks/blob/main/diffusers/stable_conceptualizer_inference.ipynb) notebook. You can also train your own concepts and load them into the concept libraries using [this notebook](https://colab.research.google.com/github/huggingface/notebooks/blob/main/diffusers/sd_textual_inversion_training.ipynb).
Here is the new concept you will be able to use as an `object`:
|
research-backup/roberta-large-semeval2012-v2-average-no-mask-prompt-e-nce | research-backup | 2022-09-19T18:56:16Z | 103 | 0 | transformers | [
"transformers",
"pytorch",
"roberta",
"feature-extraction",
"dataset:relbert/semeval2012_relational_similarity_v2",
"model-index",
"text-embeddings-inference",
"endpoints_compatible",
"region:us"
]
| feature-extraction | 2022-08-20T14:24:21Z | ---
datasets:
- relbert/semeval2012_relational_similarity_v2
model-index:
- name: relbert/roberta-large-semeval2012-v2-average-no-mask-prompt-e-nce
results:
- task:
name: Relation Mapping
type: sorting-task
dataset:
name: Relation Mapping
args: relbert/relation_mapping
type: relation-mapping
metrics:
- name: Accuracy
type: accuracy
value: 0.911547619047619
- task:
name: Analogy Questions (SAT full)
type: multiple-choice-qa
dataset:
name: SAT full
args: relbert/analogy_questions
type: analogy-questions
metrics:
- name: Accuracy
type: accuracy
value: 0.5935828877005348
- task:
name: Analogy Questions (SAT)
type: multiple-choice-qa
dataset:
name: SAT
args: relbert/analogy_questions
type: analogy-questions
metrics:
- name: Accuracy
type: accuracy
value: 0.6023738872403561
- task:
name: Analogy Questions (BATS)
type: multiple-choice-qa
dataset:
name: BATS
args: relbert/analogy_questions
type: analogy-questions
metrics:
- name: Accuracy
type: accuracy
value: 0.7498610339077265
- task:
name: Analogy Questions (Google)
type: multiple-choice-qa
dataset:
name: Google
args: relbert/analogy_questions
type: analogy-questions
metrics:
- name: Accuracy
type: accuracy
value: 0.868
- task:
name: Analogy Questions (U2)
type: multiple-choice-qa
dataset:
name: U2
args: relbert/analogy_questions
type: analogy-questions
metrics:
- name: Accuracy
type: accuracy
value: 0.618421052631579
- task:
name: Analogy Questions (U4)
type: multiple-choice-qa
dataset:
name: U4
args: relbert/analogy_questions
type: analogy-questions
metrics:
- name: Accuracy
type: accuracy
value: 0.6365740740740741
- task:
name: Lexical Relation Classification (BLESS)
type: classification
dataset:
name: BLESS
args: relbert/lexical_relation_classification
type: relation-classification
metrics:
- name: F1
type: f1
value: 0.9190899502787404
- name: F1 (macro)
type: f1_macro
value: 0.9137760457433256
- task:
name: Lexical Relation Classification (CogALexV)
type: classification
dataset:
name: CogALexV
args: relbert/lexical_relation_classification
type: relation-classification
metrics:
- name: F1
type: f1
value: 0.854225352112676
- name: F1 (macro)
type: f1_macro
value: 0.6960792498811619
- task:
name: Lexical Relation Classification (EVALution)
type: classification
dataset:
name: BLESS
args: relbert/lexical_relation_classification
type: relation-classification
metrics:
- name: F1
type: f1
value: 0.6738894907908992
- name: F1 (macro)
type: f1_macro
value: 0.6683142084374337
- task:
name: Lexical Relation Classification (K&H+N)
type: classification
dataset:
name: K&H+N
args: relbert/lexical_relation_classification
type: relation-classification
metrics:
- name: F1
type: f1
value: 0.9637615636085414
- name: F1 (macro)
type: f1_macro
value: 0.890107974704234
- task:
name: Lexical Relation Classification (ROOT09)
type: classification
dataset:
name: ROOT09
args: relbert/lexical_relation_classification
type: relation-classification
metrics:
- name: F1
type: f1
value: 0.9053588216859918
- name: F1 (macro)
type: f1_macro
value: 0.9023263285944801
---
# relbert/roberta-large-semeval2012-v2-average-no-mask-prompt-e-nce
RelBERT fine-tuned from [roberta-large](https://huggingface.co/roberta-large) on
[relbert/semeval2012_relational_similarity_v2](https://huggingface.co/datasets/relbert/semeval2012_relational_similarity_v2).
Fine-tuning is done via [RelBERT](https://github.com/asahi417/relbert) library (see the repository for more detail).
It achieves the following results on the relation understanding tasks:
- Analogy Question ([dataset](https://huggingface.co/datasets/relbert/analogy_questions), [full result](https://huggingface.co/relbert/roberta-large-semeval2012-v2-average-no-mask-prompt-e-nce/raw/main/analogy.json)):
- Accuracy on SAT (full): 0.5935828877005348
- Accuracy on SAT: 0.6023738872403561
- Accuracy on BATS: 0.7498610339077265
- Accuracy on U2: 0.618421052631579
- Accuracy on U4: 0.6365740740740741
- Accuracy on Google: 0.868
- Lexical Relation Classification ([dataset](https://huggingface.co/datasets/relbert/lexical_relation_classification), [full result](https://huggingface.co/relbert/roberta-large-semeval2012-v2-average-no-mask-prompt-e-nce/raw/main/classification.json)):
- Micro F1 score on BLESS: 0.9190899502787404
- Micro F1 score on CogALexV: 0.854225352112676
- Micro F1 score on EVALution: 0.6738894907908992
- Micro F1 score on K&H+N: 0.9637615636085414
- Micro F1 score on ROOT09: 0.9053588216859918
- Relation Mapping ([dataset](https://huggingface.co/datasets/relbert/relation_mapping), [full result](https://huggingface.co/relbert/roberta-large-semeval2012-v2-average-no-mask-prompt-e-nce/raw/main/relation_mapping.json)):
- Accuracy on Relation Mapping: 0.911547619047619
### Usage
This model can be used through the [relbert library](https://github.com/asahi417/relbert). Install the library via pip
```shell
pip install relbert
```
and activate model as below.
```python
from relbert import RelBERT
model = RelBERT("relbert/roberta-large-semeval2012-v2-average-no-mask-prompt-e-nce")
vector = model.get_embedding(['Tokyo', 'Japan']) # shape of (1024, )
```
### Training hyperparameters
The following hyperparameters were used during training:
- model: roberta-large
- max_length: 64
- mode: average_no_mask
- data: relbert/semeval2012_relational_similarity_v2
- template_mode: manual
- template: I wasn’t aware of this relationship, but I just read in the encyclopedia that <obj> is <subj>’s <mask>
- loss_function: nce_logout
- temperature_nce_constant: 0.05
- temperature_nce_rank: {'min': 0.01, 'max': 0.05, 'type': 'linear'}
- epoch: 21
- batch: 128
- lr: 5e-06
- lr_decay: False
- lr_warmup: 1
- weight_decay: 0
- random_seed: 0
- exclude_relation: None
- n_sample: 640
- gradient_accumulation: 8
The full configuration can be found at [fine-tuning parameter file](https://huggingface.co/relbert/roberta-large-semeval2012-v2-average-no-mask-prompt-e-nce/raw/main/trainer_config.json).
### Reference
If you use any resource from RelBERT, please consider to cite our [paper](https://aclanthology.org/2021.eacl-demos.7/).
```
@inproceedings{ushio-etal-2021-distilling-relation-embeddings,
title = "{D}istilling {R}elation {E}mbeddings from {P}re-trained {L}anguage {M}odels",
author = "Ushio, Asahi and
Schockaert, Steven and
Camacho-Collados, Jose",
booktitle = "EMNLP 2021",
year = "2021",
address = "Online",
publisher = "Association for Computational Linguistics",
}
```
|
research-backup/roberta-large-semeval2012-v2-average-no-mask-prompt-c-nce | research-backup | 2022-09-19T18:49:11Z | 104 | 0 | transformers | [
"transformers",
"pytorch",
"roberta",
"feature-extraction",
"dataset:relbert/semeval2012_relational_similarity_v2",
"model-index",
"text-embeddings-inference",
"endpoints_compatible",
"region:us"
]
| feature-extraction | 2022-08-20T04:15:11Z | ---
datasets:
- relbert/semeval2012_relational_similarity_v2
model-index:
- name: relbert/roberta-large-semeval2012-v2-average-no-mask-prompt-c-nce
results:
- task:
name: Relation Mapping
type: sorting-task
dataset:
name: Relation Mapping
args: relbert/relation_mapping
type: relation-mapping
metrics:
- name: Accuracy
type: accuracy
value: 0.8667460317460317
- task:
name: Analogy Questions (SAT full)
type: multiple-choice-qa
dataset:
name: SAT full
args: relbert/analogy_questions
type: analogy-questions
metrics:
- name: Accuracy
type: accuracy
value: 0.6283422459893048
- task:
name: Analogy Questions (SAT)
type: multiple-choice-qa
dataset:
name: SAT
args: relbert/analogy_questions
type: analogy-questions
metrics:
- name: Accuracy
type: accuracy
value: 0.6320474777448071
- task:
name: Analogy Questions (BATS)
type: multiple-choice-qa
dataset:
name: BATS
args: relbert/analogy_questions
type: analogy-questions
metrics:
- name: Accuracy
type: accuracy
value: 0.7726514730405781
- task:
name: Analogy Questions (Google)
type: multiple-choice-qa
dataset:
name: Google
args: relbert/analogy_questions
type: analogy-questions
metrics:
- name: Accuracy
type: accuracy
value: 0.92
- task:
name: Analogy Questions (U2)
type: multiple-choice-qa
dataset:
name: U2
args: relbert/analogy_questions
type: analogy-questions
metrics:
- name: Accuracy
type: accuracy
value: 0.631578947368421
- task:
name: Analogy Questions (U4)
type: multiple-choice-qa
dataset:
name: U4
args: relbert/analogy_questions
type: analogy-questions
metrics:
- name: Accuracy
type: accuracy
value: 0.6365740740740741
- task:
name: Lexical Relation Classification (BLESS)
type: classification
dataset:
name: BLESS
args: relbert/lexical_relation_classification
type: relation-classification
metrics:
- name: F1
type: f1
value: 0.9218020189844809
- name: F1 (macro)
type: f1_macro
value: 0.9177802775241372
- task:
name: Lexical Relation Classification (CogALexV)
type: classification
dataset:
name: CogALexV
args: relbert/lexical_relation_classification
type: relation-classification
metrics:
- name: F1
type: f1
value: 0.8694835680751174
- name: F1 (macro)
type: f1_macro
value: 0.7143050763634161
- task:
name: Lexical Relation Classification (EVALution)
type: classification
dataset:
name: BLESS
args: relbert/lexical_relation_classification
type: relation-classification
metrics:
- name: F1
type: f1
value: 0.6852654387865655
- name: F1 (macro)
type: f1_macro
value: 0.6719168721915184
- task:
name: Lexical Relation Classification (K&H+N)
type: classification
dataset:
name: K&H+N
args: relbert/lexical_relation_classification
type: relation-classification
metrics:
- name: F1
type: f1
value: 0.9568755651387633
- name: F1 (macro)
type: f1_macro
value: 0.8813368476884477
- task:
name: Lexical Relation Classification (ROOT09)
type: classification
dataset:
name: ROOT09
args: relbert/lexical_relation_classification
type: relation-classification
metrics:
- name: F1
type: f1
value: 0.911939830774052
- name: F1 (macro)
type: f1_macro
value: 0.9100147353897433
---
# relbert/roberta-large-semeval2012-v2-average-no-mask-prompt-c-nce
RelBERT fine-tuned from [roberta-large](https://huggingface.co/roberta-large) on
[relbert/semeval2012_relational_similarity_v2](https://huggingface.co/datasets/relbert/semeval2012_relational_similarity_v2).
Fine-tuning is done via [RelBERT](https://github.com/asahi417/relbert) library (see the repository for more detail).
It achieves the following results on the relation understanding tasks:
- Analogy Question ([dataset](https://huggingface.co/datasets/relbert/analogy_questions), [full result](https://huggingface.co/relbert/roberta-large-semeval2012-v2-average-no-mask-prompt-c-nce/raw/main/analogy.json)):
- Accuracy on SAT (full): 0.6283422459893048
- Accuracy on SAT: 0.6320474777448071
- Accuracy on BATS: 0.7726514730405781
- Accuracy on U2: 0.631578947368421
- Accuracy on U4: 0.6365740740740741
- Accuracy on Google: 0.92
- Lexical Relation Classification ([dataset](https://huggingface.co/datasets/relbert/lexical_relation_classification), [full result](https://huggingface.co/relbert/roberta-large-semeval2012-v2-average-no-mask-prompt-c-nce/raw/main/classification.json)):
- Micro F1 score on BLESS: 0.9218020189844809
- Micro F1 score on CogALexV: 0.8694835680751174
- Micro F1 score on EVALution: 0.6852654387865655
- Micro F1 score on K&H+N: 0.9568755651387633
- Micro F1 score on ROOT09: 0.911939830774052
- Relation Mapping ([dataset](https://huggingface.co/datasets/relbert/relation_mapping), [full result](https://huggingface.co/relbert/roberta-large-semeval2012-v2-average-no-mask-prompt-c-nce/raw/main/relation_mapping.json)):
- Accuracy on Relation Mapping: 0.8667460317460317
### Usage
This model can be used through the [relbert library](https://github.com/asahi417/relbert). Install the library via pip
```shell
pip install relbert
```
and activate model as below.
```python
from relbert import RelBERT
model = RelBERT("relbert/roberta-large-semeval2012-v2-average-no-mask-prompt-c-nce")
vector = model.get_embedding(['Tokyo', 'Japan']) # shape of (1024, )
```
### Training hyperparameters
The following hyperparameters were used during training:
- model: roberta-large
- max_length: 64
- mode: average_no_mask
- data: relbert/semeval2012_relational_similarity_v2
- template_mode: manual
- template: Today, I finally discovered the relation between <subj> and <obj> : <mask>
- loss_function: nce_logout
- temperature_nce_constant: 0.05
- temperature_nce_rank: {'min': 0.01, 'max': 0.05, 'type': 'linear'}
- epoch: 29
- batch: 128
- lr: 5e-06
- lr_decay: False
- lr_warmup: 1
- weight_decay: 0
- random_seed: 0
- exclude_relation: None
- n_sample: 640
- gradient_accumulation: 8
The full configuration can be found at [fine-tuning parameter file](https://huggingface.co/relbert/roberta-large-semeval2012-v2-average-no-mask-prompt-c-nce/raw/main/trainer_config.json).
### Reference
If you use any resource from RelBERT, please consider to cite our [paper](https://aclanthology.org/2021.eacl-demos.7/).
```
@inproceedings{ushio-etal-2021-distilling-relation-embeddings,
title = "{D}istilling {R}elation {E}mbeddings from {P}re-trained {L}anguage {M}odels",
author = "Ushio, Asahi and
Schockaert, Steven and
Camacho-Collados, Jose",
booktitle = "EMNLP 2021",
year = "2021",
address = "Online",
publisher = "Association for Computational Linguistics",
}
```
|
adil-o/dqn-SpaceInvadersNoFrameskip-v4 | adil-o | 2022-09-19T18:46:41Z | 1 | 0 | stable-baselines3 | [
"stable-baselines3",
"SpaceInvadersNoFrameskip-v4",
"deep-reinforcement-learning",
"reinforcement-learning",
"model-index",
"region:us"
]
| reinforcement-learning | 2022-09-19T18:46:06Z | ---
library_name: stable-baselines3
tags:
- SpaceInvadersNoFrameskip-v4
- deep-reinforcement-learning
- reinforcement-learning
- stable-baselines3
model-index:
- name: DQN
results:
- metrics:
- type: mean_reward
value: 425.50 +/- 151.35
name: mean_reward
task:
type: reinforcement-learning
name: reinforcement-learning
dataset:
name: SpaceInvadersNoFrameskip-v4
type: SpaceInvadersNoFrameskip-v4
---
# **DQN** Agent playing **SpaceInvadersNoFrameskip-v4**
This is a trained model of a **DQN** agent playing **SpaceInvadersNoFrameskip-v4**
using the [stable-baselines3 library](https://github.com/DLR-RM/stable-baselines3)
and the [RL Zoo](https://github.com/DLR-RM/rl-baselines3-zoo).
The RL Zoo is a training framework for Stable Baselines3
reinforcement learning agents,
with hyperparameter optimization and pre-trained agents included.
## Usage (with SB3 RL Zoo)
RL Zoo: https://github.com/DLR-RM/rl-baselines3-zoo<br/>
SB3: https://github.com/DLR-RM/stable-baselines3<br/>
SB3 Contrib: https://github.com/Stable-Baselines-Team/stable-baselines3-contrib
```
# Download model and save it into the logs/ folder
python -m utils.load_from_hub --algo dqn --env SpaceInvadersNoFrameskip-v4 -orga adil-o -f logs/
python enjoy.py --algo dqn --env SpaceInvadersNoFrameskip-v4 -f logs/
```
## Training (with the RL Zoo)
```
python train.py --algo dqn --env SpaceInvadersNoFrameskip-v4 -f logs/
# Upload the model and generate video (when possible)
python -m utils.push_to_hub --algo dqn --env SpaceInvadersNoFrameskip-v4 -f logs/ -orga adil-o
```
## Hyperparameters
```python
OrderedDict([('batch_size', 32),
('buffer_size', 100000),
('env_wrapper',
['stable_baselines3.common.atari_wrappers.AtariWrapper']),
('exploration_final_eps', 0.01),
('exploration_fraction', 0.1),
('frame_stack', 3),
('gradient_steps', 1),
('learning_rate', 0.0001),
('learning_starts', 100000),
('n_timesteps', 1000000.0),
('optimize_memory_usage', False),
('policy', 'CnnPolicy'),
('target_update_interval', 1000),
('train_freq', 4),
('normalize', False)])
```
|
research-backup/roberta-large-semeval2012-v2-average-no-mask-prompt-b-nce | research-backup | 2022-09-19T18:45:31Z | 106 | 0 | transformers | [
"transformers",
"pytorch",
"roberta",
"feature-extraction",
"dataset:relbert/semeval2012_relational_similarity_v2",
"model-index",
"text-embeddings-inference",
"endpoints_compatible",
"region:us"
]
| feature-extraction | 2022-08-19T23:10:39Z | ---
datasets:
- relbert/semeval2012_relational_similarity_v2
model-index:
- name: relbert/roberta-large-semeval2012-v2-average-no-mask-prompt-b-nce
results:
- task:
name: Relation Mapping
type: sorting-task
dataset:
name: Relation Mapping
args: relbert/relation_mapping
type: relation-mapping
metrics:
- name: Accuracy
type: accuracy
value: 0.8426984126984127
- task:
name: Analogy Questions (SAT full)
type: multiple-choice-qa
dataset:
name: SAT full
args: relbert/analogy_questions
type: analogy-questions
metrics:
- name: Accuracy
type: accuracy
value: 0.5989304812834224
- task:
name: Analogy Questions (SAT)
type: multiple-choice-qa
dataset:
name: SAT
args: relbert/analogy_questions
type: analogy-questions
metrics:
- name: Accuracy
type: accuracy
value: 0.599406528189911
- task:
name: Analogy Questions (BATS)
type: multiple-choice-qa
dataset:
name: BATS
args: relbert/analogy_questions
type: analogy-questions
metrics:
- name: Accuracy
type: accuracy
value: 0.7570872707059477
- task:
name: Analogy Questions (Google)
type: multiple-choice-qa
dataset:
name: Google
args: relbert/analogy_questions
type: analogy-questions
metrics:
- name: Accuracy
type: accuracy
value: 0.9
- task:
name: Analogy Questions (U2)
type: multiple-choice-qa
dataset:
name: U2
args: relbert/analogy_questions
type: analogy-questions
metrics:
- name: Accuracy
type: accuracy
value: 0.5175438596491229
- task:
name: Analogy Questions (U4)
type: multiple-choice-qa
dataset:
name: U4
args: relbert/analogy_questions
type: analogy-questions
metrics:
- name: Accuracy
type: accuracy
value: 0.6203703703703703
- task:
name: Lexical Relation Classification (BLESS)
type: classification
dataset:
name: BLESS
args: relbert/lexical_relation_classification
type: relation-classification
metrics:
- name: F1
type: f1
value: 0.9142684948018682
- name: F1 (macro)
type: f1_macro
value: 0.9110348384096337
- task:
name: Lexical Relation Classification (CogALexV)
type: classification
dataset:
name: CogALexV
args: relbert/lexical_relation_classification
type: relation-classification
metrics:
- name: F1
type: f1
value: 0.8523474178403756
- name: F1 (macro)
type: f1_macro
value: 0.6846296909617913
- task:
name: Lexical Relation Classification (EVALution)
type: classification
dataset:
name: BLESS
args: relbert/lexical_relation_classification
type: relation-classification
metrics:
- name: F1
type: f1
value: 0.6728060671722643
- name: F1 (macro)
type: f1_macro
value: 0.6659943041357158
- task:
name: Lexical Relation Classification (K&H+N)
type: classification
dataset:
name: K&H+N
args: relbert/lexical_relation_classification
type: relation-classification
metrics:
- name: F1
type: f1
value: 0.9463031230437504
- name: F1 (macro)
type: f1_macro
value: 0.8467358173655017
- task:
name: Lexical Relation Classification (ROOT09)
type: classification
dataset:
name: ROOT09
args: relbert/lexical_relation_classification
type: relation-classification
metrics:
- name: F1
type: f1
value: 0.8968975242870574
- name: F1 (macro)
type: f1_macro
value: 0.8954832383340866
---
# relbert/roberta-large-semeval2012-v2-average-no-mask-prompt-b-nce
RelBERT fine-tuned from [roberta-large](https://huggingface.co/roberta-large) on
[relbert/semeval2012_relational_similarity_v2](https://huggingface.co/datasets/relbert/semeval2012_relational_similarity_v2).
Fine-tuning is done via [RelBERT](https://github.com/asahi417/relbert) library (see the repository for more detail).
It achieves the following results on the relation understanding tasks:
- Analogy Question ([dataset](https://huggingface.co/datasets/relbert/analogy_questions), [full result](https://huggingface.co/relbert/roberta-large-semeval2012-v2-average-no-mask-prompt-b-nce/raw/main/analogy.json)):
- Accuracy on SAT (full): 0.5989304812834224
- Accuracy on SAT: 0.599406528189911
- Accuracy on BATS: 0.7570872707059477
- Accuracy on U2: 0.5175438596491229
- Accuracy on U4: 0.6203703703703703
- Accuracy on Google: 0.9
- Lexical Relation Classification ([dataset](https://huggingface.co/datasets/relbert/lexical_relation_classification), [full result](https://huggingface.co/relbert/roberta-large-semeval2012-v2-average-no-mask-prompt-b-nce/raw/main/classification.json)):
- Micro F1 score on BLESS: 0.9142684948018682
- Micro F1 score on CogALexV: 0.8523474178403756
- Micro F1 score on EVALution: 0.6728060671722643
- Micro F1 score on K&H+N: 0.9463031230437504
- Micro F1 score on ROOT09: 0.8968975242870574
- Relation Mapping ([dataset](https://huggingface.co/datasets/relbert/relation_mapping), [full result](https://huggingface.co/relbert/roberta-large-semeval2012-v2-average-no-mask-prompt-b-nce/raw/main/relation_mapping.json)):
- Accuracy on Relation Mapping: 0.8426984126984127
### Usage
This model can be used through the [relbert library](https://github.com/asahi417/relbert). Install the library via pip
```shell
pip install relbert
```
and activate model as below.
```python
from relbert import RelBERT
model = RelBERT("relbert/roberta-large-semeval2012-v2-average-no-mask-prompt-b-nce")
vector = model.get_embedding(['Tokyo', 'Japan']) # shape of (1024, )
```
### Training hyperparameters
The following hyperparameters were used during training:
- model: roberta-large
- max_length: 64
- mode: average_no_mask
- data: relbert/semeval2012_relational_similarity_v2
- template_mode: manual
- template: Today, I finally discovered the relation between <subj> and <obj> : <obj> is <subj>'s <mask>
- loss_function: nce_logout
- temperature_nce_constant: 0.05
- temperature_nce_rank: {'min': 0.01, 'max': 0.05, 'type': 'linear'}
- epoch: 28
- batch: 128
- lr: 5e-06
- lr_decay: False
- lr_warmup: 1
- weight_decay: 0
- random_seed: 0
- exclude_relation: None
- n_sample: 640
- gradient_accumulation: 8
The full configuration can be found at [fine-tuning parameter file](https://huggingface.co/relbert/roberta-large-semeval2012-v2-average-no-mask-prompt-b-nce/raw/main/trainer_config.json).
### Reference
If you use any resource from RelBERT, please consider to cite our [paper](https://aclanthology.org/2021.eacl-demos.7/).
```
@inproceedings{ushio-etal-2021-distilling-relation-embeddings,
title = "{D}istilling {R}elation {E}mbeddings from {P}re-trained {L}anguage {M}odels",
author = "Ushio, Asahi and
Schockaert, Steven and
Camacho-Collados, Jose",
booktitle = "EMNLP 2021",
year = "2021",
address = "Online",
publisher = "Association for Computational Linguistics",
}
```
|
research-backup/roberta-large-semeval2012-v2-average-no-mask-prompt-a-nce | research-backup | 2022-09-19T18:41:57Z | 104 | 0 | transformers | [
"transformers",
"pytorch",
"roberta",
"feature-extraction",
"dataset:relbert/semeval2012_relational_similarity_v2",
"model-index",
"text-embeddings-inference",
"endpoints_compatible",
"region:us"
]
| feature-extraction | 2022-08-19T18:06:50Z | ---
datasets:
- relbert/semeval2012_relational_similarity_v2
model-index:
- name: relbert/roberta-large-semeval2012-v2-average-no-mask-prompt-a-nce
results:
- task:
name: Relation Mapping
type: sorting-task
dataset:
name: Relation Mapping
args: relbert/relation_mapping
type: relation-mapping
metrics:
- name: Accuracy
type: accuracy
value: 0.8451190476190477
- task:
name: Analogy Questions (SAT full)
type: multiple-choice-qa
dataset:
name: SAT full
args: relbert/analogy_questions
type: analogy-questions
metrics:
- name: Accuracy
type: accuracy
value: 0.6096256684491979
- task:
name: Analogy Questions (SAT)
type: multiple-choice-qa
dataset:
name: SAT
args: relbert/analogy_questions
type: analogy-questions
metrics:
- name: Accuracy
type: accuracy
value: 0.6142433234421365
- task:
name: Analogy Questions (BATS)
type: multiple-choice-qa
dataset:
name: BATS
args: relbert/analogy_questions
type: analogy-questions
metrics:
- name: Accuracy
type: accuracy
value: 0.7504168982768205
- task:
name: Analogy Questions (Google)
type: multiple-choice-qa
dataset:
name: Google
args: relbert/analogy_questions
type: analogy-questions
metrics:
- name: Accuracy
type: accuracy
value: 0.902
- task:
name: Analogy Questions (U2)
type: multiple-choice-qa
dataset:
name: U2
args: relbert/analogy_questions
type: analogy-questions
metrics:
- name: Accuracy
type: accuracy
value: 0.5701754385964912
- task:
name: Analogy Questions (U4)
type: multiple-choice-qa
dataset:
name: U4
args: relbert/analogy_questions
type: analogy-questions
metrics:
- name: Accuracy
type: accuracy
value: 0.6342592592592593
- task:
name: Lexical Relation Classification (BLESS)
type: classification
dataset:
name: BLESS
args: relbert/lexical_relation_classification
type: relation-classification
metrics:
- name: F1
type: f1
value: 0.9174325749585657
- name: F1 (macro)
type: f1_macro
value: 0.9147052349582953
- task:
name: Lexical Relation Classification (CogALexV)
type: classification
dataset:
name: CogALexV
args: relbert/lexical_relation_classification
type: relation-classification
metrics:
- name: F1
type: f1
value: 0.8525821596244132
- name: F1 (macro)
type: f1_macro
value: 0.6857226427858921
- task:
name: Lexical Relation Classification (EVALution)
type: classification
dataset:
name: BLESS
args: relbert/lexical_relation_classification
type: relation-classification
metrics:
- name: F1
type: f1
value: 0.6771397616468039
- name: F1 (macro)
type: f1_macro
value: 0.6712704719484096
- task:
name: Lexical Relation Classification (K&H+N)
type: classification
dataset:
name: K&H+N
args: relbert/lexical_relation_classification
type: relation-classification
metrics:
- name: F1
type: f1
value: 0.9540933435348126
- name: F1 (macro)
type: f1_macro
value: 0.8681826742269192
- task:
name: Lexical Relation Classification (ROOT09)
type: classification
dataset:
name: ROOT09
args: relbert/lexical_relation_classification
type: relation-classification
metrics:
- name: F1
type: f1
value: 0.9106863052334692
- name: F1 (macro)
type: f1_macro
value: 0.9083078769735016
---
# relbert/roberta-large-semeval2012-v2-average-no-mask-prompt-a-nce
RelBERT fine-tuned from [roberta-large](https://huggingface.co/roberta-large) on
[relbert/semeval2012_relational_similarity_v2](https://huggingface.co/datasets/relbert/semeval2012_relational_similarity_v2).
Fine-tuning is done via [RelBERT](https://github.com/asahi417/relbert) library (see the repository for more detail).
It achieves the following results on the relation understanding tasks:
- Analogy Question ([dataset](https://huggingface.co/datasets/relbert/analogy_questions), [full result](https://huggingface.co/relbert/roberta-large-semeval2012-v2-average-no-mask-prompt-a-nce/raw/main/analogy.json)):
- Accuracy on SAT (full): 0.6096256684491979
- Accuracy on SAT: 0.6142433234421365
- Accuracy on BATS: 0.7504168982768205
- Accuracy on U2: 0.5701754385964912
- Accuracy on U4: 0.6342592592592593
- Accuracy on Google: 0.902
- Lexical Relation Classification ([dataset](https://huggingface.co/datasets/relbert/lexical_relation_classification), [full result](https://huggingface.co/relbert/roberta-large-semeval2012-v2-average-no-mask-prompt-a-nce/raw/main/classification.json)):
- Micro F1 score on BLESS: 0.9174325749585657
- Micro F1 score on CogALexV: 0.8525821596244132
- Micro F1 score on EVALution: 0.6771397616468039
- Micro F1 score on K&H+N: 0.9540933435348126
- Micro F1 score on ROOT09: 0.9106863052334692
- Relation Mapping ([dataset](https://huggingface.co/datasets/relbert/relation_mapping), [full result](https://huggingface.co/relbert/roberta-large-semeval2012-v2-average-no-mask-prompt-a-nce/raw/main/relation_mapping.json)):
- Accuracy on Relation Mapping: 0.8451190476190477
### Usage
This model can be used through the [relbert library](https://github.com/asahi417/relbert). Install the library via pip
```shell
pip install relbert
```
and activate model as below.
```python
from relbert import RelBERT
model = RelBERT("relbert/roberta-large-semeval2012-v2-average-no-mask-prompt-a-nce")
vector = model.get_embedding(['Tokyo', 'Japan']) # shape of (1024, )
```
### Training hyperparameters
The following hyperparameters were used during training:
- model: roberta-large
- max_length: 64
- mode: average_no_mask
- data: relbert/semeval2012_relational_similarity_v2
- template_mode: manual
- template: Today, I finally discovered the relation between <subj> and <obj> : <subj> is the <mask> of <obj>
- loss_function: nce_logout
- temperature_nce_constant: 0.05
- temperature_nce_rank: {'min': 0.01, 'max': 0.05, 'type': 'linear'}
- epoch: 29
- batch: 128
- lr: 5e-06
- lr_decay: False
- lr_warmup: 1
- weight_decay: 0
- random_seed: 0
- exclude_relation: None
- n_sample: 640
- gradient_accumulation: 8
The full configuration can be found at [fine-tuning parameter file](https://huggingface.co/relbert/roberta-large-semeval2012-v2-average-no-mask-prompt-a-nce/raw/main/trainer_config.json).
### Reference
If you use any resource from RelBERT, please consider to cite our [paper](https://aclanthology.org/2021.eacl-demos.7/).
```
@inproceedings{ushio-etal-2021-distilling-relation-embeddings,
title = "{D}istilling {R}elation {E}mbeddings from {P}re-trained {L}anguage {M}odels",
author = "Ushio, Asahi and
Schockaert, Steven and
Camacho-Collados, Jose",
booktitle = "EMNLP 2021",
year = "2021",
address = "Online",
publisher = "Association for Computational Linguistics",
}
```
|
research-backup/roberta-large-semeval2012-v2-average-prompt-e-nce | research-backup | 2022-09-19T18:38:22Z | 104 | 0 | transformers | [
"transformers",
"pytorch",
"roberta",
"feature-extraction",
"dataset:relbert/semeval2012_relational_similarity_v2",
"model-index",
"text-embeddings-inference",
"endpoints_compatible",
"region:us"
]
| feature-extraction | 2022-08-19T13:03:12Z | ---
datasets:
- relbert/semeval2012_relational_similarity_v2
model-index:
- name: relbert/roberta-large-semeval2012-v2-average-prompt-e-nce
results:
- task:
name: Relation Mapping
type: sorting-task
dataset:
name: Relation Mapping
args: relbert/relation_mapping
type: relation-mapping
metrics:
- name: Accuracy
type: accuracy
value: 0.8316269841269841
- task:
name: Analogy Questions (SAT full)
type: multiple-choice-qa
dataset:
name: SAT full
args: relbert/analogy_questions
type: analogy-questions
metrics:
- name: Accuracy
type: accuracy
value: 0.6096256684491979
- task:
name: Analogy Questions (SAT)
type: multiple-choice-qa
dataset:
name: SAT
args: relbert/analogy_questions
type: analogy-questions
metrics:
- name: Accuracy
type: accuracy
value: 0.6053412462908012
- task:
name: Analogy Questions (BATS)
type: multiple-choice-qa
dataset:
name: BATS
args: relbert/analogy_questions
type: analogy-questions
metrics:
- name: Accuracy
type: accuracy
value: 0.754863813229572
- task:
name: Analogy Questions (Google)
type: multiple-choice-qa
dataset:
name: Google
args: relbert/analogy_questions
type: analogy-questions
metrics:
- name: Accuracy
type: accuracy
value: 0.886
- task:
name: Analogy Questions (U2)
type: multiple-choice-qa
dataset:
name: U2
args: relbert/analogy_questions
type: analogy-questions
metrics:
- name: Accuracy
type: accuracy
value: 0.5570175438596491
- task:
name: Analogy Questions (U4)
type: multiple-choice-qa
dataset:
name: U4
args: relbert/analogy_questions
type: analogy-questions
metrics:
- name: Accuracy
type: accuracy
value: 0.6273148148148148
- task:
name: Lexical Relation Classification (BLESS)
type: classification
dataset:
name: BLESS
args: relbert/lexical_relation_classification
type: relation-classification
metrics:
- name: F1
type: f1
value: 0.9255687810757872
- name: F1 (macro)
type: f1_macro
value: 0.9223987715298315
- task:
name: Lexical Relation Classification (CogALexV)
type: classification
dataset:
name: CogALexV
args: relbert/lexical_relation_classification
type: relation-classification
metrics:
- name: F1
type: f1
value: 0.8396713615023474
- name: F1 (macro)
type: f1_macro
value: 0.6696257351833762
- task:
name: Lexical Relation Classification (EVALution)
type: classification
dataset:
name: BLESS
args: relbert/lexical_relation_classification
type: relation-classification
metrics:
- name: F1
type: f1
value: 0.6803900325027086
- name: F1 (macro)
type: f1_macro
value: 0.6668423343702814
- task:
name: Lexical Relation Classification (K&H+N)
type: classification
dataset:
name: K&H+N
args: relbert/lexical_relation_classification
type: relation-classification
metrics:
- name: F1
type: f1
value: 0.9586840091813313
- name: F1 (macro)
type: f1_macro
value: 0.876302193283041
- task:
name: Lexical Relation Classification (ROOT09)
type: classification
dataset:
name: ROOT09
args: relbert/lexical_relation_classification
type: relation-classification
metrics:
- name: F1
type: f1
value: 0.9019116264493889
- name: F1 (macro)
type: f1_macro
value: 0.9004696301895015
---
# relbert/roberta-large-semeval2012-v2-average-prompt-e-nce
RelBERT fine-tuned from [roberta-large](https://huggingface.co/roberta-large) on
[relbert/semeval2012_relational_similarity_v2](https://huggingface.co/datasets/relbert/semeval2012_relational_similarity_v2).
Fine-tuning is done via [RelBERT](https://github.com/asahi417/relbert) library (see the repository for more detail).
It achieves the following results on the relation understanding tasks:
- Analogy Question ([dataset](https://huggingface.co/datasets/relbert/analogy_questions), [full result](https://huggingface.co/relbert/roberta-large-semeval2012-v2-average-prompt-e-nce/raw/main/analogy.json)):
- Accuracy on SAT (full): 0.6096256684491979
- Accuracy on SAT: 0.6053412462908012
- Accuracy on BATS: 0.754863813229572
- Accuracy on U2: 0.5570175438596491
- Accuracy on U4: 0.6273148148148148
- Accuracy on Google: 0.886
- Lexical Relation Classification ([dataset](https://huggingface.co/datasets/relbert/lexical_relation_classification), [full result](https://huggingface.co/relbert/roberta-large-semeval2012-v2-average-prompt-e-nce/raw/main/classification.json)):
- Micro F1 score on BLESS: 0.9255687810757872
- Micro F1 score on CogALexV: 0.8396713615023474
- Micro F1 score on EVALution: 0.6803900325027086
- Micro F1 score on K&H+N: 0.9586840091813313
- Micro F1 score on ROOT09: 0.9019116264493889
- Relation Mapping ([dataset](https://huggingface.co/datasets/relbert/relation_mapping), [full result](https://huggingface.co/relbert/roberta-large-semeval2012-v2-average-prompt-e-nce/raw/main/relation_mapping.json)):
- Accuracy on Relation Mapping: 0.8316269841269841
### Usage
This model can be used through the [relbert library](https://github.com/asahi417/relbert). Install the library via pip
```shell
pip install relbert
```
and activate model as below.
```python
from relbert import RelBERT
model = RelBERT("relbert/roberta-large-semeval2012-v2-average-prompt-e-nce")
vector = model.get_embedding(['Tokyo', 'Japan']) # shape of (1024, )
```
### Training hyperparameters
The following hyperparameters were used during training:
- model: roberta-large
- max_length: 64
- mode: average
- data: relbert/semeval2012_relational_similarity_v2
- template_mode: manual
- template: I wasn’t aware of this relationship, but I just read in the encyclopedia that <obj> is <subj>’s <mask>
- loss_function: nce_logout
- temperature_nce_constant: 0.05
- temperature_nce_rank: {'min': 0.01, 'max': 0.05, 'type': 'linear'}
- epoch: 29
- batch: 128
- lr: 5e-06
- lr_decay: False
- lr_warmup: 1
- weight_decay: 0
- random_seed: 0
- exclude_relation: None
- n_sample: 640
- gradient_accumulation: 8
The full configuration can be found at [fine-tuning parameter file](https://huggingface.co/relbert/roberta-large-semeval2012-v2-average-prompt-e-nce/raw/main/trainer_config.json).
### Reference
If you use any resource from RelBERT, please consider to cite our [paper](https://aclanthology.org/2021.eacl-demos.7/).
```
@inproceedings{ushio-etal-2021-distilling-relation-embeddings,
title = "{D}istilling {R}elation {E}mbeddings from {P}re-trained {L}anguage {M}odels",
author = "Ushio, Asahi and
Schockaert, Steven and
Camacho-Collados, Jose",
booktitle = "EMNLP 2021",
year = "2021",
address = "Online",
publisher = "Association for Computational Linguistics",
}
```
|
research-backup/roberta-large-semeval2012-v2-average-prompt-d-nce | research-backup | 2022-09-19T18:34:49Z | 103 | 0 | transformers | [
"transformers",
"pytorch",
"roberta",
"feature-extraction",
"dataset:relbert/semeval2012_relational_similarity_v2",
"model-index",
"text-embeddings-inference",
"endpoints_compatible",
"region:us"
]
| feature-extraction | 2022-08-19T07:58:25Z | ---
datasets:
- relbert/semeval2012_relational_similarity_v2
model-index:
- name: relbert/roberta-large-semeval2012-v2-average-prompt-d-nce
results:
- task:
name: Relation Mapping
type: sorting-task
dataset:
name: Relation Mapping
args: relbert/relation_mapping
type: relation-mapping
metrics:
- name: Accuracy
type: accuracy
value: 0.8500793650793651
- task:
name: Analogy Questions (SAT full)
type: multiple-choice-qa
dataset:
name: SAT full
args: relbert/analogy_questions
type: analogy-questions
metrics:
- name: Accuracy
type: accuracy
value: 0.6871657754010695
- task:
name: Analogy Questions (SAT)
type: multiple-choice-qa
dataset:
name: SAT
args: relbert/analogy_questions
type: analogy-questions
metrics:
- name: Accuracy
type: accuracy
value: 0.685459940652819
- task:
name: Analogy Questions (BATS)
type: multiple-choice-qa
dataset:
name: BATS
args: relbert/analogy_questions
type: analogy-questions
metrics:
- name: Accuracy
type: accuracy
value: 0.7715397443023903
- task:
name: Analogy Questions (Google)
type: multiple-choice-qa
dataset:
name: Google
args: relbert/analogy_questions
type: analogy-questions
metrics:
- name: Accuracy
type: accuracy
value: 0.93
- task:
name: Analogy Questions (U2)
type: multiple-choice-qa
dataset:
name: U2
args: relbert/analogy_questions
type: analogy-questions
metrics:
- name: Accuracy
type: accuracy
value: 0.631578947368421
- task:
name: Analogy Questions (U4)
type: multiple-choice-qa
dataset:
name: U4
args: relbert/analogy_questions
type: analogy-questions
metrics:
- name: Accuracy
type: accuracy
value: 0.6527777777777778
- task:
name: Lexical Relation Classification (BLESS)
type: classification
dataset:
name: BLESS
args: relbert/lexical_relation_classification
type: relation-classification
metrics:
- name: F1
type: f1
value: 0.9205966551152629
- name: F1 (macro)
type: f1_macro
value: 0.9178786648785846
- task:
name: Lexical Relation Classification (CogALexV)
type: classification
dataset:
name: CogALexV
args: relbert/lexical_relation_classification
type: relation-classification
metrics:
- name: F1
type: f1
value: 0.8643192488262911
- name: F1 (macro)
type: f1_macro
value: 0.7176774180702856
- task:
name: Lexical Relation Classification (EVALution)
type: classification
dataset:
name: BLESS
args: relbert/lexical_relation_classification
type: relation-classification
metrics:
- name: F1
type: f1
value: 0.6814734561213435
- name: F1 (macro)
type: f1_macro
value: 0.6668276311830647
- task:
name: Lexical Relation Classification (K&H+N)
type: classification
dataset:
name: K&H+N
args: relbert/lexical_relation_classification
type: relation-classification
metrics:
- name: F1
type: f1
value: 0.9590317868818251
- name: F1 (macro)
type: f1_macro
value: 0.865524335026249
- task:
name: Lexical Relation Classification (ROOT09)
type: classification
dataset:
name: ROOT09
args: relbert/lexical_relation_classification
type: relation-classification
metrics:
- name: F1
type: f1
value: 0.8968975242870574
- name: F1 (macro)
type: f1_macro
value: 0.8920343126786042
---
# relbert/roberta-large-semeval2012-v2-average-prompt-d-nce
RelBERT fine-tuned from [roberta-large](https://huggingface.co/roberta-large) on
[relbert/semeval2012_relational_similarity_v2](https://huggingface.co/datasets/relbert/semeval2012_relational_similarity_v2).
Fine-tuning is done via [RelBERT](https://github.com/asahi417/relbert) library (see the repository for more detail).
It achieves the following results on the relation understanding tasks:
- Analogy Question ([dataset](https://huggingface.co/datasets/relbert/analogy_questions), [full result](https://huggingface.co/relbert/roberta-large-semeval2012-v2-average-prompt-d-nce/raw/main/analogy.json)):
- Accuracy on SAT (full): 0.6871657754010695
- Accuracy on SAT: 0.685459940652819
- Accuracy on BATS: 0.7715397443023903
- Accuracy on U2: 0.631578947368421
- Accuracy on U4: 0.6527777777777778
- Accuracy on Google: 0.93
- Lexical Relation Classification ([dataset](https://huggingface.co/datasets/relbert/lexical_relation_classification), [full result](https://huggingface.co/relbert/roberta-large-semeval2012-v2-average-prompt-d-nce/raw/main/classification.json)):
- Micro F1 score on BLESS: 0.9205966551152629
- Micro F1 score on CogALexV: 0.8643192488262911
- Micro F1 score on EVALution: 0.6814734561213435
- Micro F1 score on K&H+N: 0.9590317868818251
- Micro F1 score on ROOT09: 0.8968975242870574
- Relation Mapping ([dataset](https://huggingface.co/datasets/relbert/relation_mapping), [full result](https://huggingface.co/relbert/roberta-large-semeval2012-v2-average-prompt-d-nce/raw/main/relation_mapping.json)):
- Accuracy on Relation Mapping: 0.8500793650793651
### Usage
This model can be used through the [relbert library](https://github.com/asahi417/relbert). Install the library via pip
```shell
pip install relbert
```
and activate model as below.
```python
from relbert import RelBERT
model = RelBERT("relbert/roberta-large-semeval2012-v2-average-prompt-d-nce")
vector = model.get_embedding(['Tokyo', 'Japan']) # shape of (1024, )
```
### Training hyperparameters
The following hyperparameters were used during training:
- model: roberta-large
- max_length: 64
- mode: average
- data: relbert/semeval2012_relational_similarity_v2
- template_mode: manual
- template: I wasn’t aware of this relationship, but I just read in the encyclopedia that <subj> is the <mask> of <obj>
- loss_function: nce_logout
- temperature_nce_constant: 0.05
- temperature_nce_rank: {'min': 0.01, 'max': 0.05, 'type': 'linear'}
- epoch: 22
- batch: 128
- lr: 5e-06
- lr_decay: False
- lr_warmup: 1
- weight_decay: 0
- random_seed: 0
- exclude_relation: None
- n_sample: 640
- gradient_accumulation: 8
The full configuration can be found at [fine-tuning parameter file](https://huggingface.co/relbert/roberta-large-semeval2012-v2-average-prompt-d-nce/raw/main/trainer_config.json).
### Reference
If you use any resource from RelBERT, please consider to cite our [paper](https://aclanthology.org/2021.eacl-demos.7/).
```
@inproceedings{ushio-etal-2021-distilling-relation-embeddings,
title = "{D}istilling {R}elation {E}mbeddings from {P}re-trained {L}anguage {M}odels",
author = "Ushio, Asahi and
Schockaert, Steven and
Camacho-Collados, Jose",
booktitle = "EMNLP 2021",
year = "2021",
address = "Online",
publisher = "Association for Computational Linguistics",
}
```
|
research-backup/roberta-large-semeval2012-v2-average-prompt-c-nce | research-backup | 2022-09-19T18:31:18Z | 103 | 0 | transformers | [
"transformers",
"pytorch",
"roberta",
"feature-extraction",
"dataset:relbert/semeval2012_relational_similarity_v2",
"model-index",
"text-embeddings-inference",
"endpoints_compatible",
"region:us"
]
| feature-extraction | 2022-08-19T02:53:49Z | ---
datasets:
- relbert/semeval2012_relational_similarity_v2
model-index:
- name: relbert/roberta-large-semeval2012-v2-average-prompt-c-nce
results:
- task:
name: Relation Mapping
type: sorting-task
dataset:
name: Relation Mapping
args: relbert/relation_mapping
type: relation-mapping
metrics:
- name: Accuracy
type: accuracy
value: 0.8846428571428572
- task:
name: Analogy Questions (SAT full)
type: multiple-choice-qa
dataset:
name: SAT full
args: relbert/analogy_questions
type: analogy-questions
metrics:
- name: Accuracy
type: accuracy
value: 0.6818181818181818
- task:
name: Analogy Questions (SAT)
type: multiple-choice-qa
dataset:
name: SAT
args: relbert/analogy_questions
type: analogy-questions
metrics:
- name: Accuracy
type: accuracy
value: 0.6735905044510386
- task:
name: Analogy Questions (BATS)
type: multiple-choice-qa
dataset:
name: BATS
args: relbert/analogy_questions
type: analogy-questions
metrics:
- name: Accuracy
type: accuracy
value: 0.811561978877154
- task:
name: Analogy Questions (Google)
type: multiple-choice-qa
dataset:
name: Google
args: relbert/analogy_questions
type: analogy-questions
metrics:
- name: Accuracy
type: accuracy
value: 0.924
- task:
name: Analogy Questions (U2)
type: multiple-choice-qa
dataset:
name: U2
args: relbert/analogy_questions
type: analogy-questions
metrics:
- name: Accuracy
type: accuracy
value: 0.631578947368421
- task:
name: Analogy Questions (U4)
type: multiple-choice-qa
dataset:
name: U4
args: relbert/analogy_questions
type: analogy-questions
metrics:
- name: Accuracy
type: accuracy
value: 0.6504629629629629
- task:
name: Lexical Relation Classification (BLESS)
type: classification
dataset:
name: BLESS
args: relbert/lexical_relation_classification
type: relation-classification
metrics:
- name: F1
type: f1
value: 0.924815428657526
- name: F1 (macro)
type: f1_macro
value: 0.9212115289556371
- task:
name: Lexical Relation Classification (CogALexV)
type: classification
dataset:
name: CogALexV
args: relbert/lexical_relation_classification
type: relation-classification
metrics:
- name: F1
type: f1
value: 0.8720657276995305
- name: F1 (macro)
type: f1_macro
value: 0.7245215538948597
- task:
name: Lexical Relation Classification (EVALution)
type: classification
dataset:
name: BLESS
args: relbert/lexical_relation_classification
type: relation-classification
metrics:
- name: F1
type: f1
value: 0.6841820151679306
- name: F1 (macro)
type: f1_macro
value: 0.6787204202080052
- task:
name: Lexical Relation Classification (K&H+N)
type: classification
dataset:
name: K&H+N
args: relbert/lexical_relation_classification
type: relation-classification
metrics:
- name: F1
type: f1
value: 0.9559713431174793
- name: F1 (macro)
type: f1_macro
value: 0.8722517438133693
- task:
name: Lexical Relation Classification (ROOT09)
type: classification
dataset:
name: ROOT09
args: relbert/lexical_relation_classification
type: relation-classification
metrics:
- name: F1
type: f1
value: 0.9088060169225948
- name: F1 (macro)
type: f1_macro
value: 0.9066857579930224
---
# relbert/roberta-large-semeval2012-v2-average-prompt-c-nce
RelBERT fine-tuned from [roberta-large](https://huggingface.co/roberta-large) on
[relbert/semeval2012_relational_similarity_v2](https://huggingface.co/datasets/relbert/semeval2012_relational_similarity_v2).
Fine-tuning is done via [RelBERT](https://github.com/asahi417/relbert) library (see the repository for more detail).
It achieves the following results on the relation understanding tasks:
- Analogy Question ([dataset](https://huggingface.co/datasets/relbert/analogy_questions), [full result](https://huggingface.co/relbert/roberta-large-semeval2012-v2-average-prompt-c-nce/raw/main/analogy.json)):
- Accuracy on SAT (full): 0.6818181818181818
- Accuracy on SAT: 0.6735905044510386
- Accuracy on BATS: 0.811561978877154
- Accuracy on U2: 0.631578947368421
- Accuracy on U4: 0.6504629629629629
- Accuracy on Google: 0.924
- Lexical Relation Classification ([dataset](https://huggingface.co/datasets/relbert/lexical_relation_classification), [full result](https://huggingface.co/relbert/roberta-large-semeval2012-v2-average-prompt-c-nce/raw/main/classification.json)):
- Micro F1 score on BLESS: 0.924815428657526
- Micro F1 score on CogALexV: 0.8720657276995305
- Micro F1 score on EVALution: 0.6841820151679306
- Micro F1 score on K&H+N: 0.9559713431174793
- Micro F1 score on ROOT09: 0.9088060169225948
- Relation Mapping ([dataset](https://huggingface.co/datasets/relbert/relation_mapping), [full result](https://huggingface.co/relbert/roberta-large-semeval2012-v2-average-prompt-c-nce/raw/main/relation_mapping.json)):
- Accuracy on Relation Mapping: 0.8846428571428572
### Usage
This model can be used through the [relbert library](https://github.com/asahi417/relbert). Install the library via pip
```shell
pip install relbert
```
and activate model as below.
```python
from relbert import RelBERT
model = RelBERT("relbert/roberta-large-semeval2012-v2-average-prompt-c-nce")
vector = model.get_embedding(['Tokyo', 'Japan']) # shape of (1024, )
```
### Training hyperparameters
The following hyperparameters were used during training:
- model: roberta-large
- max_length: 64
- mode: average
- data: relbert/semeval2012_relational_similarity_v2
- template_mode: manual
- template: Today, I finally discovered the relation between <subj> and <obj> : <mask>
- loss_function: nce_logout
- temperature_nce_constant: 0.05
- temperature_nce_rank: {'min': 0.01, 'max': 0.05, 'type': 'linear'}
- epoch: 29
- batch: 128
- lr: 5e-06
- lr_decay: False
- lr_warmup: 1
- weight_decay: 0
- random_seed: 0
- exclude_relation: None
- n_sample: 640
- gradient_accumulation: 8
The full configuration can be found at [fine-tuning parameter file](https://huggingface.co/relbert/roberta-large-semeval2012-v2-average-prompt-c-nce/raw/main/trainer_config.json).
### Reference
If you use any resource from RelBERT, please consider to cite our [paper](https://aclanthology.org/2021.eacl-demos.7/).
```
@inproceedings{ushio-etal-2021-distilling-relation-embeddings,
title = "{D}istilling {R}elation {E}mbeddings from {P}re-trained {L}anguage {M}odels",
author = "Ushio, Asahi and
Schockaert, Steven and
Camacho-Collados, Jose",
booktitle = "EMNLP 2021",
year = "2021",
address = "Online",
publisher = "Association for Computational Linguistics",
}
```
|
research-backup/roberta-large-semeval2012-v2-average-prompt-a-nce | research-backup | 2022-09-19T18:24:10Z | 105 | 0 | transformers | [
"transformers",
"pytorch",
"roberta",
"feature-extraction",
"dataset:relbert/semeval2012_relational_similarity_v2",
"model-index",
"text-embeddings-inference",
"endpoints_compatible",
"region:us"
]
| feature-extraction | 2022-08-18T16:46:29Z | ---
datasets:
- relbert/semeval2012_relational_similarity_v2
model-index:
- name: relbert/roberta-large-semeval2012-v2-average-prompt-a-nce
results:
- task:
name: Relation Mapping
type: sorting-task
dataset:
name: Relation Mapping
args: relbert/relation_mapping
type: relation-mapping
metrics:
- name: Accuracy
type: accuracy
value: 0.8308333333333333
- task:
name: Analogy Questions (SAT full)
type: multiple-choice-qa
dataset:
name: SAT full
args: relbert/analogy_questions
type: analogy-questions
metrics:
- name: Accuracy
type: accuracy
value: 0.6390374331550802
- task:
name: Analogy Questions (SAT)
type: multiple-choice-qa
dataset:
name: SAT
args: relbert/analogy_questions
type: analogy-questions
metrics:
- name: Accuracy
type: accuracy
value: 0.6409495548961425
- task:
name: Analogy Questions (BATS)
type: multiple-choice-qa
dataset:
name: BATS
args: relbert/analogy_questions
type: analogy-questions
metrics:
- name: Accuracy
type: accuracy
value: 0.7570872707059477
- task:
name: Analogy Questions (Google)
type: multiple-choice-qa
dataset:
name: Google
args: relbert/analogy_questions
type: analogy-questions
metrics:
- name: Accuracy
type: accuracy
value: 0.93
- task:
name: Analogy Questions (U2)
type: multiple-choice-qa
dataset:
name: U2
args: relbert/analogy_questions
type: analogy-questions
metrics:
- name: Accuracy
type: accuracy
value: 0.618421052631579
- task:
name: Analogy Questions (U4)
type: multiple-choice-qa
dataset:
name: U4
args: relbert/analogy_questions
type: analogy-questions
metrics:
- name: Accuracy
type: accuracy
value: 0.6388888888888888
- task:
name: Lexical Relation Classification (BLESS)
type: classification
dataset:
name: BLESS
args: relbert/lexical_relation_classification
type: relation-classification
metrics:
- name: F1
type: f1
value: 0.9154738586710863
- name: F1 (macro)
type: f1_macro
value: 0.9102917119981933
- task:
name: Lexical Relation Classification (CogALexV)
type: classification
dataset:
name: CogALexV
args: relbert/lexical_relation_classification
type: relation-classification
metrics:
- name: F1
type: f1
value: 0.852112676056338
- name: F1 (macro)
type: f1_macro
value: 0.6892409688901546
- task:
name: Lexical Relation Classification (EVALution)
type: classification
dataset:
name: BLESS
args: relbert/lexical_relation_classification
type: relation-classification
metrics:
- name: F1
type: f1
value: 0.6852654387865655
- name: F1 (macro)
type: f1_macro
value: 0.6726667668087644
- task:
name: Lexical Relation Classification (K&H+N)
type: classification
dataset:
name: K&H+N
args: relbert/lexical_relation_classification
type: relation-classification
metrics:
- name: F1
type: f1
value: 0.9533282325937261
- name: F1 (macro)
type: f1_macro
value: 0.862481874668915
- task:
name: Lexical Relation Classification (ROOT09)
type: classification
dataset:
name: ROOT09
args: relbert/lexical_relation_classification
type: relation-classification
metrics:
- name: F1
type: f1
value: 0.9078658727671576
- name: F1 (macro)
type: f1_macro
value: 0.9075386153074033
---
# relbert/roberta-large-semeval2012-v2-average-prompt-a-nce
RelBERT fine-tuned from [roberta-large](https://huggingface.co/roberta-large) on
[relbert/semeval2012_relational_similarity_v2](https://huggingface.co/datasets/relbert/semeval2012_relational_similarity_v2).
Fine-tuning is done via [RelBERT](https://github.com/asahi417/relbert) library (see the repository for more detail).
It achieves the following results on the relation understanding tasks:
- Analogy Question ([dataset](https://huggingface.co/datasets/relbert/analogy_questions), [full result](https://huggingface.co/relbert/roberta-large-semeval2012-v2-average-prompt-a-nce/raw/main/analogy.json)):
- Accuracy on SAT (full): 0.6390374331550802
- Accuracy on SAT: 0.6409495548961425
- Accuracy on BATS: 0.7570872707059477
- Accuracy on U2: 0.618421052631579
- Accuracy on U4: 0.6388888888888888
- Accuracy on Google: 0.93
- Lexical Relation Classification ([dataset](https://huggingface.co/datasets/relbert/lexical_relation_classification), [full result](https://huggingface.co/relbert/roberta-large-semeval2012-v2-average-prompt-a-nce/raw/main/classification.json)):
- Micro F1 score on BLESS: 0.9154738586710863
- Micro F1 score on CogALexV: 0.852112676056338
- Micro F1 score on EVALution: 0.6852654387865655
- Micro F1 score on K&H+N: 0.9533282325937261
- Micro F1 score on ROOT09: 0.9078658727671576
- Relation Mapping ([dataset](https://huggingface.co/datasets/relbert/relation_mapping), [full result](https://huggingface.co/relbert/roberta-large-semeval2012-v2-average-prompt-a-nce/raw/main/relation_mapping.json)):
- Accuracy on Relation Mapping: 0.8308333333333333
### Usage
This model can be used through the [relbert library](https://github.com/asahi417/relbert). Install the library via pip
```shell
pip install relbert
```
and activate model as below.
```python
from relbert import RelBERT
model = RelBERT("relbert/roberta-large-semeval2012-v2-average-prompt-a-nce")
vector = model.get_embedding(['Tokyo', 'Japan']) # shape of (1024, )
```
### Training hyperparameters
The following hyperparameters were used during training:
- model: roberta-large
- max_length: 64
- mode: average
- data: relbert/semeval2012_relational_similarity_v2
- template_mode: manual
- template: Today, I finally discovered the relation between <subj> and <obj> : <subj> is the <mask> of <obj>
- loss_function: nce_logout
- temperature_nce_constant: 0.05
- temperature_nce_rank: {'min': 0.01, 'max': 0.05, 'type': 'linear'}
- epoch: 29
- batch: 128
- lr: 5e-06
- lr_decay: False
- lr_warmup: 1
- weight_decay: 0
- random_seed: 0
- exclude_relation: None
- n_sample: 640
- gradient_accumulation: 8
The full configuration can be found at [fine-tuning parameter file](https://huggingface.co/relbert/roberta-large-semeval2012-v2-average-prompt-a-nce/raw/main/trainer_config.json).
### Reference
If you use any resource from RelBERT, please consider to cite our [paper](https://aclanthology.org/2021.eacl-demos.7/).
```
@inproceedings{ushio-etal-2021-distilling-relation-embeddings,
title = "{D}istilling {R}elation {E}mbeddings from {P}re-trained {L}anguage {M}odels",
author = "Ushio, Asahi and
Schockaert, Steven and
Camacho-Collados, Jose",
booktitle = "EMNLP 2021",
year = "2021",
address = "Online",
publisher = "Association for Computational Linguistics",
}
```
|
research-backup/roberta-large-semeval2012-v2-mask-prompt-e-nce | research-backup | 2022-09-19T18:20:37Z | 107 | 0 | transformers | [
"transformers",
"pytorch",
"roberta",
"feature-extraction",
"dataset:relbert/semeval2012_relational_similarity_v2",
"model-index",
"text-embeddings-inference",
"endpoints_compatible",
"region:us"
]
| feature-extraction | 2022-08-18T11:42:44Z | ---
datasets:
- relbert/semeval2012_relational_similarity_v2
model-index:
- name: relbert/roberta-large-semeval2012-v2-mask-prompt-e-nce
results:
- task:
name: Relation Mapping
type: sorting-task
dataset:
name: Relation Mapping
args: relbert/relation_mapping
type: relation-mapping
metrics:
- name: Accuracy
type: accuracy
value: 0.8457142857142858
- task:
name: Analogy Questions (SAT full)
type: multiple-choice-qa
dataset:
name: SAT full
args: relbert/analogy_questions
type: analogy-questions
metrics:
- name: Accuracy
type: accuracy
value: 0.6096256684491979
- task:
name: Analogy Questions (SAT)
type: multiple-choice-qa
dataset:
name: SAT
args: relbert/analogy_questions
type: analogy-questions
metrics:
- name: Accuracy
type: accuracy
value: 0.6112759643916914
- task:
name: Analogy Questions (BATS)
type: multiple-choice-qa
dataset:
name: BATS
args: relbert/analogy_questions
type: analogy-questions
metrics:
- name: Accuracy
type: accuracy
value: 0.7576431350750417
- task:
name: Analogy Questions (Google)
type: multiple-choice-qa
dataset:
name: Google
args: relbert/analogy_questions
type: analogy-questions
metrics:
- name: Accuracy
type: accuracy
value: 0.878
- task:
name: Analogy Questions (U2)
type: multiple-choice-qa
dataset:
name: U2
args: relbert/analogy_questions
type: analogy-questions
metrics:
- name: Accuracy
type: accuracy
value: 0.5964912280701754
- task:
name: Analogy Questions (U4)
type: multiple-choice-qa
dataset:
name: U4
args: relbert/analogy_questions
type: analogy-questions
metrics:
- name: Accuracy
type: accuracy
value: 0.6087962962962963
- task:
name: Lexical Relation Classification (BLESS)
type: classification
dataset:
name: BLESS
args: relbert/lexical_relation_classification
type: relation-classification
metrics:
- name: F1
type: f1
value: 0.9264728039777008
- name: F1 (macro)
type: f1_macro
value: 0.9231888761944194
- task:
name: Lexical Relation Classification (CogALexV)
type: classification
dataset:
name: CogALexV
args: relbert/lexical_relation_classification
type: relation-classification
metrics:
- name: F1
type: f1
value: 0.8720657276995305
- name: F1 (macro)
type: f1_macro
value: 0.7203249423895846
- task:
name: Lexical Relation Classification (EVALution)
type: classification
dataset:
name: BLESS
args: relbert/lexical_relation_classification
type: relation-classification
metrics:
- name: F1
type: f1
value: 0.7074756229685807
- name: F1 (macro)
type: f1_macro
value: 0.7003587066174993
- task:
name: Lexical Relation Classification (K&H+N)
type: classification
dataset:
name: K&H+N
args: relbert/lexical_relation_classification
type: relation-classification
metrics:
- name: F1
type: f1
value: 0.9625095638867636
- name: F1 (macro)
type: f1_macro
value: 0.8943198093953978
- task:
name: Lexical Relation Classification (ROOT09)
type: classification
dataset:
name: ROOT09
args: relbert/lexical_relation_classification
type: relation-classification
metrics:
- name: F1
type: f1
value: 0.9022250078345346
- name: F1 (macro)
type: f1_macro
value: 0.9008228707899653
---
# relbert/roberta-large-semeval2012-v2-mask-prompt-e-nce
RelBERT fine-tuned from [roberta-large](https://huggingface.co/roberta-large) on
[relbert/semeval2012_relational_similarity_v2](https://huggingface.co/datasets/relbert/semeval2012_relational_similarity_v2).
Fine-tuning is done via [RelBERT](https://github.com/asahi417/relbert) library (see the repository for more detail).
It achieves the following results on the relation understanding tasks:
- Analogy Question ([dataset](https://huggingface.co/datasets/relbert/analogy_questions), [full result](https://huggingface.co/relbert/roberta-large-semeval2012-v2-mask-prompt-e-nce/raw/main/analogy.json)):
- Accuracy on SAT (full): 0.6096256684491979
- Accuracy on SAT: 0.6112759643916914
- Accuracy on BATS: 0.7576431350750417
- Accuracy on U2: 0.5964912280701754
- Accuracy on U4: 0.6087962962962963
- Accuracy on Google: 0.878
- Lexical Relation Classification ([dataset](https://huggingface.co/datasets/relbert/lexical_relation_classification), [full result](https://huggingface.co/relbert/roberta-large-semeval2012-v2-mask-prompt-e-nce/raw/main/classification.json)):
- Micro F1 score on BLESS: 0.9264728039777008
- Micro F1 score on CogALexV: 0.8720657276995305
- Micro F1 score on EVALution: 0.7074756229685807
- Micro F1 score on K&H+N: 0.9625095638867636
- Micro F1 score on ROOT09: 0.9022250078345346
- Relation Mapping ([dataset](https://huggingface.co/datasets/relbert/relation_mapping), [full result](https://huggingface.co/relbert/roberta-large-semeval2012-v2-mask-prompt-e-nce/raw/main/relation_mapping.json)):
- Accuracy on Relation Mapping: 0.8457142857142858
### Usage
This model can be used through the [relbert library](https://github.com/asahi417/relbert). Install the library via pip
```shell
pip install relbert
```
and activate model as below.
```python
from relbert import RelBERT
model = RelBERT("relbert/roberta-large-semeval2012-v2-mask-prompt-e-nce")
vector = model.get_embedding(['Tokyo', 'Japan']) # shape of (1024, )
```
### Training hyperparameters
The following hyperparameters were used during training:
- model: roberta-large
- max_length: 64
- mode: mask
- data: relbert/semeval2012_relational_similarity_v2
- template_mode: manual
- template: I wasn’t aware of this relationship, but I just read in the encyclopedia that <obj> is <subj>’s <mask>
- loss_function: nce_logout
- temperature_nce_constant: 0.05
- temperature_nce_rank: {'min': 0.01, 'max': 0.05, 'type': 'linear'}
- epoch: 29
- batch: 128
- lr: 5e-06
- lr_decay: False
- lr_warmup: 1
- weight_decay: 0
- random_seed: 0
- exclude_relation: None
- n_sample: 640
- gradient_accumulation: 8
The full configuration can be found at [fine-tuning parameter file](https://huggingface.co/relbert/roberta-large-semeval2012-v2-mask-prompt-e-nce/raw/main/trainer_config.json).
### Reference
If you use any resource from RelBERT, please consider to cite our [paper](https://aclanthology.org/2021.eacl-demos.7/).
```
@inproceedings{ushio-etal-2021-distilling-relation-embeddings,
title = "{D}istilling {R}elation {E}mbeddings from {P}re-trained {L}anguage {M}odels",
author = "Ushio, Asahi and
Schockaert, Steven and
Camacho-Collados, Jose",
booktitle = "EMNLP 2021",
year = "2021",
address = "Online",
publisher = "Association for Computational Linguistics",
}
```
|
sd-concepts-library/grisstyle | sd-concepts-library | 2022-09-19T18:17:52Z | 0 | 9 | null | [
"license:mit",
"region:us"
]
| null | 2022-09-19T18:17:47Z | ---
license: mit
---
### GrisStyle on Stable Diffusion
This is the `<gris>` concept taught to Stable Diffusion via Textual Inversion. You can load this concept into the [Stable Conceptualizer](https://colab.research.google.com/github/huggingface/notebooks/blob/main/diffusers/stable_conceptualizer_inference.ipynb) notebook. You can also train your own concepts and load them into the concept libraries using [this notebook](https://colab.research.google.com/github/huggingface/notebooks/blob/main/diffusers/sd_textual_inversion_training.ipynb).
Here is the new concept you will be able to use as a `style`:
|
sd-concepts-library/crested-gecko | sd-concepts-library | 2022-09-19T18:17:08Z | 0 | 0 | null | [
"license:mit",
"region:us"
]
| null | 2022-09-19T18:17:04Z | ---
license: mit
---
### crested gecko on Stable Diffusion
This is the `<crested-gecko>` concept taught to Stable Diffusion via Textual Inversion. You can load this concept into the [Stable Conceptualizer](https://colab.research.google.com/github/huggingface/notebooks/blob/main/diffusers/stable_conceptualizer_inference.ipynb) notebook. You can also train your own concepts and load them into the concept libraries using [this notebook](https://colab.research.google.com/github/huggingface/notebooks/blob/main/diffusers/sd_textual_inversion_training.ipynb).
Here is the new concept you will be able to use as an `object`:
|
research-backup/roberta-large-semeval2012-v2-mask-prompt-d-nce | research-backup | 2022-09-19T18:17:02Z | 97 | 0 | transformers | [
"transformers",
"pytorch",
"roberta",
"feature-extraction",
"dataset:relbert/semeval2012_relational_similarity_v2",
"model-index",
"text-embeddings-inference",
"endpoints_compatible",
"region:us"
]
| feature-extraction | 2022-08-18T06:38:11Z | ---
datasets:
- relbert/semeval2012_relational_similarity_v2
model-index:
- name: relbert/roberta-large-semeval2012-v2-mask-prompt-d-nce
results:
- task:
name: Relation Mapping
type: sorting-task
dataset:
name: Relation Mapping
args: relbert/relation_mapping
type: relation-mapping
metrics:
- name: Accuracy
type: accuracy
value: 0.8847619047619047
- task:
name: Analogy Questions (SAT full)
type: multiple-choice-qa
dataset:
name: SAT full
args: relbert/analogy_questions
type: analogy-questions
metrics:
- name: Accuracy
type: accuracy
value: 0.6925133689839572
- task:
name: Analogy Questions (SAT)
type: multiple-choice-qa
dataset:
name: SAT
args: relbert/analogy_questions
type: analogy-questions
metrics:
- name: Accuracy
type: accuracy
value: 0.6913946587537092
- task:
name: Analogy Questions (BATS)
type: multiple-choice-qa
dataset:
name: BATS
args: relbert/analogy_questions
type: analogy-questions
metrics:
- name: Accuracy
type: accuracy
value: 0.7871039466370205
- task:
name: Analogy Questions (Google)
type: multiple-choice-qa
dataset:
name: Google
args: relbert/analogy_questions
type: analogy-questions
metrics:
- name: Accuracy
type: accuracy
value: 0.938
- task:
name: Analogy Questions (U2)
type: multiple-choice-qa
dataset:
name: U2
args: relbert/analogy_questions
type: analogy-questions
metrics:
- name: Accuracy
type: accuracy
value: 0.6403508771929824
- task:
name: Analogy Questions (U4)
type: multiple-choice-qa
dataset:
name: U4
args: relbert/analogy_questions
type: analogy-questions
metrics:
- name: Accuracy
type: accuracy
value: 0.6504629629629629
- task:
name: Lexical Relation Classification (BLESS)
type: classification
dataset:
name: BLESS
args: relbert/lexical_relation_classification
type: relation-classification
metrics:
- name: F1
type: f1
value: 0.9242127467229169
- name: F1 (macro)
type: f1_macro
value: 0.9194459817003556
- task:
name: Lexical Relation Classification (CogALexV)
type: classification
dataset:
name: CogALexV
args: relbert/lexical_relation_classification
type: relation-classification
metrics:
- name: F1
type: f1
value: 0.8776995305164319
- name: F1 (macro)
type: f1_macro
value: 0.7319038272561456
- task:
name: Lexical Relation Classification (EVALution)
type: classification
dataset:
name: BLESS
args: relbert/lexical_relation_classification
type: relation-classification
metrics:
- name: F1
type: f1
value: 0.7215601300108341
- name: F1 (macro)
type: f1_macro
value: 0.7132048463591607
- task:
name: Lexical Relation Classification (K&H+N)
type: classification
dataset:
name: K&H+N
args: relbert/lexical_relation_classification
type: relation-classification
metrics:
- name: F1
type: f1
value: 0.9652222299506156
- name: F1 (macro)
type: f1_macro
value: 0.8926432179055038
- task:
name: Lexical Relation Classification (ROOT09)
type: classification
dataset:
name: ROOT09
args: relbert/lexical_relation_classification
type: relation-classification
metrics:
- name: F1
type: f1
value: 0.9147602632403635
- name: F1 (macro)
type: f1_macro
value: 0.9123157220224268
---
# relbert/roberta-large-semeval2012-v2-mask-prompt-d-nce
RelBERT fine-tuned from [roberta-large](https://huggingface.co/roberta-large) on
[relbert/semeval2012_relational_similarity_v2](https://huggingface.co/datasets/relbert/semeval2012_relational_similarity_v2).
Fine-tuning is done via [RelBERT](https://github.com/asahi417/relbert) library (see the repository for more detail).
It achieves the following results on the relation understanding tasks:
- Analogy Question ([dataset](https://huggingface.co/datasets/relbert/analogy_questions), [full result](https://huggingface.co/relbert/roberta-large-semeval2012-v2-mask-prompt-d-nce/raw/main/analogy.json)):
- Accuracy on SAT (full): 0.6925133689839572
- Accuracy on SAT: 0.6913946587537092
- Accuracy on BATS: 0.7871039466370205
- Accuracy on U2: 0.6403508771929824
- Accuracy on U4: 0.6504629629629629
- Accuracy on Google: 0.938
- Lexical Relation Classification ([dataset](https://huggingface.co/datasets/relbert/lexical_relation_classification), [full result](https://huggingface.co/relbert/roberta-large-semeval2012-v2-mask-prompt-d-nce/raw/main/classification.json)):
- Micro F1 score on BLESS: 0.9242127467229169
- Micro F1 score on CogALexV: 0.8776995305164319
- Micro F1 score on EVALution: 0.7215601300108341
- Micro F1 score on K&H+N: 0.9652222299506156
- Micro F1 score on ROOT09: 0.9147602632403635
- Relation Mapping ([dataset](https://huggingface.co/datasets/relbert/relation_mapping), [full result](https://huggingface.co/relbert/roberta-large-semeval2012-v2-mask-prompt-d-nce/raw/main/relation_mapping.json)):
- Accuracy on Relation Mapping: 0.8847619047619047
### Usage
This model can be used through the [relbert library](https://github.com/asahi417/relbert). Install the library via pip
```shell
pip install relbert
```
and activate model as below.
```python
from relbert import RelBERT
model = RelBERT("relbert/roberta-large-semeval2012-v2-mask-prompt-d-nce")
vector = model.get_embedding(['Tokyo', 'Japan']) # shape of (1024, )
```
### Training hyperparameters
The following hyperparameters were used during training:
- model: roberta-large
- max_length: 64
- mode: mask
- data: relbert/semeval2012_relational_similarity_v2
- template_mode: manual
- template: I wasn’t aware of this relationship, but I just read in the encyclopedia that <subj> is the <mask> of <obj>
- loss_function: nce_logout
- temperature_nce_constant: 0.05
- temperature_nce_rank: {'min': 0.01, 'max': 0.05, 'type': 'linear'}
- epoch: 28
- batch: 128
- lr: 5e-06
- lr_decay: False
- lr_warmup: 1
- weight_decay: 0
- random_seed: 0
- exclude_relation: None
- n_sample: 640
- gradient_accumulation: 8
The full configuration can be found at [fine-tuning parameter file](https://huggingface.co/relbert/roberta-large-semeval2012-v2-mask-prompt-d-nce/raw/main/trainer_config.json).
### Reference
If you use any resource from RelBERT, please consider to cite our [paper](https://aclanthology.org/2021.eacl-demos.7/).
```
@inproceedings{ushio-etal-2021-distilling-relation-embeddings,
title = "{D}istilling {R}elation {E}mbeddings from {P}re-trained {L}anguage {M}odels",
author = "Ushio, Asahi and
Schockaert, Steven and
Camacho-Collados, Jose",
booktitle = "EMNLP 2021",
year = "2021",
address = "Online",
publisher = "Association for Computational Linguistics",
}
```
|
research-backup/roberta-large-semeval2012-v2-mask-prompt-b-nce | research-backup | 2022-09-19T18:09:52Z | 104 | 0 | transformers | [
"transformers",
"pytorch",
"roberta",
"feature-extraction",
"dataset:relbert/semeval2012_relational_similarity_v2",
"model-index",
"text-embeddings-inference",
"endpoints_compatible",
"region:us"
]
| feature-extraction | 2022-08-17T20:31:16Z | ---
datasets:
- relbert/semeval2012_relational_similarity_v2
model-index:
- name: relbert/roberta-large-semeval2012-v2-mask-prompt-b-nce
results:
- task:
name: Relation Mapping
type: sorting-task
dataset:
name: Relation Mapping
args: relbert/relation_mapping
type: relation-mapping
metrics:
- name: Accuracy
type: accuracy
value: 0.8893650793650794
- task:
name: Analogy Questions (SAT full)
type: multiple-choice-qa
dataset:
name: SAT full
args: relbert/analogy_questions
type: analogy-questions
metrics:
- name: Accuracy
type: accuracy
value: 0.5909090909090909
- task:
name: Analogy Questions (SAT)
type: multiple-choice-qa
dataset:
name: SAT
args: relbert/analogy_questions
type: analogy-questions
metrics:
- name: Accuracy
type: accuracy
value: 0.5934718100890207
- task:
name: Analogy Questions (BATS)
type: multiple-choice-qa
dataset:
name: BATS
args: relbert/analogy_questions
type: analogy-questions
metrics:
- name: Accuracy
type: accuracy
value: 0.7626459143968871
- task:
name: Analogy Questions (Google)
type: multiple-choice-qa
dataset:
name: Google
args: relbert/analogy_questions
type: analogy-questions
metrics:
- name: Accuracy
type: accuracy
value: 0.892
- task:
name: Analogy Questions (U2)
type: multiple-choice-qa
dataset:
name: U2
args: relbert/analogy_questions
type: analogy-questions
metrics:
- name: Accuracy
type: accuracy
value: 0.5526315789473685
- task:
name: Analogy Questions (U4)
type: multiple-choice-qa
dataset:
name: U4
args: relbert/analogy_questions
type: analogy-questions
metrics:
- name: Accuracy
type: accuracy
value: 0.6064814814814815
- task:
name: Lexical Relation Classification (BLESS)
type: classification
dataset:
name: BLESS
args: relbert/lexical_relation_classification
type: relation-classification
metrics:
- name: F1
type: f1
value: 0.9249660991411782
- name: F1 (macro)
type: f1_macro
value: 0.9227342891403616
- task:
name: Lexical Relation Classification (CogALexV)
type: classification
dataset:
name: CogALexV
args: relbert/lexical_relation_classification
type: relation-classification
metrics:
- name: F1
type: f1
value: 0.8685446009389672
- name: F1 (macro)
type: f1_macro
value: 0.7221315620076061
- task:
name: Lexical Relation Classification (EVALution)
type: classification
dataset:
name: BLESS
args: relbert/lexical_relation_classification
type: relation-classification
metrics:
- name: F1
type: f1
value: 0.6895991332611051
- name: F1 (macro)
type: f1_macro
value: 0.6823504904547306
- task:
name: Lexical Relation Classification (K&H+N)
type: classification
dataset:
name: K&H+N
args: relbert/lexical_relation_classification
type: relation-classification
metrics:
- name: F1
type: f1
value: 0.9632051192877513
- name: F1 (macro)
type: f1_macro
value: 0.8871487887884136
- task:
name: Lexical Relation Classification (ROOT09)
type: classification
dataset:
name: ROOT09
args: relbert/lexical_relation_classification
type: relation-classification
metrics:
- name: F1
type: f1
value: 0.9015982450642431
- name: F1 (macro)
type: f1_macro
value: 0.9009961240438994
---
# relbert/roberta-large-semeval2012-v2-mask-prompt-b-nce
RelBERT fine-tuned from [roberta-large](https://huggingface.co/roberta-large) on
[relbert/semeval2012_relational_similarity_v2](https://huggingface.co/datasets/relbert/semeval2012_relational_similarity_v2).
Fine-tuning is done via [RelBERT](https://github.com/asahi417/relbert) library (see the repository for more detail).
It achieves the following results on the relation understanding tasks:
- Analogy Question ([dataset](https://huggingface.co/datasets/relbert/analogy_questions), [full result](https://huggingface.co/relbert/roberta-large-semeval2012-v2-mask-prompt-b-nce/raw/main/analogy.json)):
- Accuracy on SAT (full): 0.5909090909090909
- Accuracy on SAT: 0.5934718100890207
- Accuracy on BATS: 0.7626459143968871
- Accuracy on U2: 0.5526315789473685
- Accuracy on U4: 0.6064814814814815
- Accuracy on Google: 0.892
- Lexical Relation Classification ([dataset](https://huggingface.co/datasets/relbert/lexical_relation_classification), [full result](https://huggingface.co/relbert/roberta-large-semeval2012-v2-mask-prompt-b-nce/raw/main/classification.json)):
- Micro F1 score on BLESS: 0.9249660991411782
- Micro F1 score on CogALexV: 0.8685446009389672
- Micro F1 score on EVALution: 0.6895991332611051
- Micro F1 score on K&H+N: 0.9632051192877513
- Micro F1 score on ROOT09: 0.9015982450642431
- Relation Mapping ([dataset](https://huggingface.co/datasets/relbert/relation_mapping), [full result](https://huggingface.co/relbert/roberta-large-semeval2012-v2-mask-prompt-b-nce/raw/main/relation_mapping.json)):
- Accuracy on Relation Mapping: 0.8893650793650794
### Usage
This model can be used through the [relbert library](https://github.com/asahi417/relbert). Install the library via pip
```shell
pip install relbert
```
and activate model as below.
```python
from relbert import RelBERT
model = RelBERT("relbert/roberta-large-semeval2012-v2-mask-prompt-b-nce")
vector = model.get_embedding(['Tokyo', 'Japan']) # shape of (1024, )
```
### Training hyperparameters
The following hyperparameters were used during training:
- model: roberta-large
- max_length: 64
- mode: mask
- data: relbert/semeval2012_relational_similarity_v2
- template_mode: manual
- template: Today, I finally discovered the relation between <subj> and <obj> : <obj> is <subj>'s <mask>
- loss_function: nce_logout
- temperature_nce_constant: 0.05
- temperature_nce_rank: {'min': 0.01, 'max': 0.05, 'type': 'linear'}
- epoch: 27
- batch: 128
- lr: 5e-06
- lr_decay: False
- lr_warmup: 1
- weight_decay: 0
- random_seed: 0
- exclude_relation: None
- n_sample: 640
- gradient_accumulation: 8
The full configuration can be found at [fine-tuning parameter file](https://huggingface.co/relbert/roberta-large-semeval2012-v2-mask-prompt-b-nce/raw/main/trainer_config.json).
### Reference
If you use any resource from RelBERT, please consider to cite our [paper](https://aclanthology.org/2021.eacl-demos.7/).
```
@inproceedings{ushio-etal-2021-distilling-relation-embeddings,
title = "{D}istilling {R}elation {E}mbeddings from {P}re-trained {L}anguage {M}odels",
author = "Ushio, Asahi and
Schockaert, Steven and
Camacho-Collados, Jose",
booktitle = "EMNLP 2021",
year = "2021",
address = "Online",
publisher = "Association for Computational Linguistics",
}
```
|
AlexanderD/test_cats | AlexanderD | 2022-09-19T18:08:02Z | 0 | 0 | keras | [
"keras",
"region:us"
]
| null | 2022-09-19T17:11:54Z | ---
thumbnail: "url to a thumbnail used in social sharing"
library_name: keras
tags:
- keras
widget:
- src: https://huggingface.co/datasets/test_cats/cifar1.jpg
example_title: Tiger
--- |
research-backup/roberta-large-semeval2012-v2-mask-prompt-a-nce | research-backup | 2022-09-19T18:06:14Z | 106 | 0 | transformers | [
"transformers",
"pytorch",
"roberta",
"feature-extraction",
"dataset:relbert/semeval2012_relational_similarity_v2",
"model-index",
"text-embeddings-inference",
"endpoints_compatible",
"region:us"
]
| feature-extraction | 2022-08-17T15:27:11Z | ---
datasets:
- relbert/semeval2012_relational_similarity_v2
model-index:
- name: relbert/roberta-large-semeval2012-v2-mask-prompt-a-nce
results:
- task:
name: Relation Mapping
type: sorting-task
dataset:
name: Relation Mapping
args: relbert/relation_mapping
type: relation-mapping
metrics:
- name: Accuracy
type: accuracy
value: 0.8338095238095238
- task:
name: Analogy Questions (SAT full)
type: multiple-choice-qa
dataset:
name: SAT full
args: relbert/analogy_questions
type: analogy-questions
metrics:
- name: Accuracy
type: accuracy
value: 0.7165775401069518
- task:
name: Analogy Questions (SAT)
type: multiple-choice-qa
dataset:
name: SAT
args: relbert/analogy_questions
type: analogy-questions
metrics:
- name: Accuracy
type: accuracy
value: 0.7181008902077152
- task:
name: Analogy Questions (BATS)
type: multiple-choice-qa
dataset:
name: BATS
args: relbert/analogy_questions
type: analogy-questions
metrics:
- name: Accuracy
type: accuracy
value: 0.7626459143968871
- task:
name: Analogy Questions (Google)
type: multiple-choice-qa
dataset:
name: Google
args: relbert/analogy_questions
type: analogy-questions
metrics:
- name: Accuracy
type: accuracy
value: 0.946
- task:
name: Analogy Questions (U2)
type: multiple-choice-qa
dataset:
name: U2
args: relbert/analogy_questions
type: analogy-questions
metrics:
- name: Accuracy
type: accuracy
value: 0.6359649122807017
- task:
name: Analogy Questions (U4)
type: multiple-choice-qa
dataset:
name: U4
args: relbert/analogy_questions
type: analogy-questions
metrics:
- name: Accuracy
type: accuracy
value: 0.6435185185185185
- task:
name: Lexical Relation Classification (BLESS)
type: classification
dataset:
name: BLESS
args: relbert/lexical_relation_classification
type: relation-classification
metrics:
- name: F1
type: f1
value: 0.9264728039777008
- name: F1 (macro)
type: f1_macro
value: 0.9236796530968624
- task:
name: Lexical Relation Classification (CogALexV)
type: classification
dataset:
name: CogALexV
args: relbert/lexical_relation_classification
type: relation-classification
metrics:
- name: F1
type: f1
value: 0.8760563380281691
- name: F1 (macro)
type: f1_macro
value: 0.7316819468333253
- task:
name: Lexical Relation Classification (EVALution)
type: classification
dataset:
name: BLESS
args: relbert/lexical_relation_classification
type: relation-classification
metrics:
- name: F1
type: f1
value: 0.7085590465872156
- name: F1 (macro)
type: f1_macro
value: 0.6972629880144019
- task:
name: Lexical Relation Classification (K&H+N)
type: classification
dataset:
name: K&H+N
args: relbert/lexical_relation_classification
type: relation-classification
metrics:
- name: F1
type: f1
value: 0.9585448981011337
- name: F1 (macro)
type: f1_macro
value: 0.8754227812726614
- task:
name: Lexical Relation Classification (ROOT09)
type: classification
dataset:
name: ROOT09
args: relbert/lexical_relation_classification
type: relation-classification
metrics:
- name: F1
type: f1
value: 0.9088060169225948
- name: F1 (macro)
type: f1_macro
value: 0.9075082674855798
---
# relbert/roberta-large-semeval2012-v2-mask-prompt-a-nce
RelBERT fine-tuned from [roberta-large](https://huggingface.co/roberta-large) on
[relbert/semeval2012_relational_similarity_v2](https://huggingface.co/datasets/relbert/semeval2012_relational_similarity_v2).
Fine-tuning is done via [RelBERT](https://github.com/asahi417/relbert) library (see the repository for more detail).
It achieves the following results on the relation understanding tasks:
- Analogy Question ([dataset](https://huggingface.co/datasets/relbert/analogy_questions), [full result](https://huggingface.co/relbert/roberta-large-semeval2012-v2-mask-prompt-a-nce/raw/main/analogy.json)):
- Accuracy on SAT (full): 0.7165775401069518
- Accuracy on SAT: 0.7181008902077152
- Accuracy on BATS: 0.7626459143968871
- Accuracy on U2: 0.6359649122807017
- Accuracy on U4: 0.6435185185185185
- Accuracy on Google: 0.946
- Lexical Relation Classification ([dataset](https://huggingface.co/datasets/relbert/lexical_relation_classification), [full result](https://huggingface.co/relbert/roberta-large-semeval2012-v2-mask-prompt-a-nce/raw/main/classification.json)):
- Micro F1 score on BLESS: 0.9264728039777008
- Micro F1 score on CogALexV: 0.8760563380281691
- Micro F1 score on EVALution: 0.7085590465872156
- Micro F1 score on K&H+N: 0.9585448981011337
- Micro F1 score on ROOT09: 0.9088060169225948
- Relation Mapping ([dataset](https://huggingface.co/datasets/relbert/relation_mapping), [full result](https://huggingface.co/relbert/roberta-large-semeval2012-v2-mask-prompt-a-nce/raw/main/relation_mapping.json)):
- Accuracy on Relation Mapping: 0.8338095238095238
### Usage
This model can be used through the [relbert library](https://github.com/asahi417/relbert). Install the library via pip
```shell
pip install relbert
```
and activate model as below.
```python
from relbert import RelBERT
model = RelBERT("relbert/roberta-large-semeval2012-v2-mask-prompt-a-nce")
vector = model.get_embedding(['Tokyo', 'Japan']) # shape of (1024, )
```
### Training hyperparameters
The following hyperparameters were used during training:
- model: roberta-large
- max_length: 64
- mode: mask
- data: relbert/semeval2012_relational_similarity_v2
- template_mode: manual
- template: Today, I finally discovered the relation between <subj> and <obj> : <subj> is the <mask> of <obj>
- loss_function: nce_logout
- temperature_nce_constant: 0.05
- temperature_nce_rank: {'min': 0.01, 'max': 0.05, 'type': 'linear'}
- epoch: 29
- batch: 128
- lr: 5e-06
- lr_decay: False
- lr_warmup: 1
- weight_decay: 0
- random_seed: 0
- exclude_relation: None
- n_sample: 640
- gradient_accumulation: 8
The full configuration can be found at [fine-tuning parameter file](https://huggingface.co/relbert/roberta-large-semeval2012-v2-mask-prompt-a-nce/raw/main/trainer_config.json).
### Reference
If you use any resource from RelBERT, please consider to cite our [paper](https://aclanthology.org/2021.eacl-demos.7/).
```
@inproceedings{ushio-etal-2021-distilling-relation-embeddings,
title = "{D}istilling {R}elation {E}mbeddings from {P}re-trained {L}anguage {M}odels",
author = "Ushio, Asahi and
Schockaert, Steven and
Camacho-Collados, Jose",
booktitle = "EMNLP 2021",
year = "2021",
address = "Online",
publisher = "Association for Computational Linguistics",
}
```
|
sd-concepts-library/depthmap-style | sd-concepts-library | 2022-09-19T17:59:18Z | 0 | 15 | null | [
"license:mit",
"region:us"
]
| null | 2022-09-19T17:59:14Z | ---
license: mit
---
### Depthmap Style on Stable Diffusion
This is the `<depthmap>` concept taught to Stable Diffusion via Textual Inversion. You can load this concept into the [Stable Conceptualizer](https://colab.research.google.com/github/huggingface/notebooks/blob/main/diffusers/stable_conceptualizer_inference.ipynb) notebook. You can also train your own concepts and load them into the concept libraries using [this notebook](https://colab.research.google.com/github/huggingface/notebooks/blob/main/diffusers/sd_textual_inversion_training.ipynb).
Here is the new concept you will be able to use as a `style`:
|
huggingtweets/chriscantino | huggingtweets | 2022-09-19T17:50:31Z | 109 | 0 | transformers | [
"transformers",
"pytorch",
"gpt2",
"text-generation",
"huggingtweets",
"en",
"autotrain_compatible",
"text-generation-inference",
"endpoints_compatible",
"region:us"
]
| text-generation | 2022-09-19T17:49:30Z | ---
language: en
thumbnail: http://www.huggingtweets.com/chriscantino/1663609825906/predictions.png
tags:
- huggingtweets
widget:
- text: "My dream is"
---
<div class="inline-flex flex-col" style="line-height: 1.5;">
<div class="flex">
<div
style="display:inherit; margin-left: 4px; margin-right: 4px; width: 92px; height:92px; border-radius: 50%; background-size: cover; background-image: url('https://pbs.twimg.com/profile_images/1554673291570212864/RWwZGZ1E_400x400.png')">
</div>
<div
style="display:none; margin-left: 4px; margin-right: 4px; width: 92px; height:92px; border-radius: 50%; background-size: cover; background-image: url('')">
</div>
<div
style="display:none; margin-left: 4px; margin-right: 4px; width: 92px; height:92px; border-radius: 50%; background-size: cover; background-image: url('')">
</div>
</div>
<div style="text-align: center; margin-top: 3px; font-size: 16px; font-weight: 800">🤖 AI BOT 🤖</div>
<div style="text-align: center; font-size: 16px; font-weight: 800">cantino.eth</div>
<div style="text-align: center; font-size: 14px;">@chriscantino</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 tweets from cantino.eth.
| Data | cantino.eth |
| --- | --- |
| Tweets downloaded | 3250 |
| Retweets | 38 |
| Short tweets | 666 |
| Tweets kept | 2546 |
[Explore the data](https://wandb.ai/wandb/huggingtweets/runs/3owatwcf/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 @chriscantino's tweets.
Hyperparameters and metrics are recorded in the [W&B training run](https://wandb.ai/wandb/huggingtweets/runs/2b5rm6g9) for full transparency and reproducibility.
At the end of training, [the final model](https://wandb.ai/wandb/huggingtweets/runs/2b5rm6g9/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/chriscantino')
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)
|
sd-concepts-library/f-22 | sd-concepts-library | 2022-09-19T17:46:19Z | 0 | 0 | null | [
"license:mit",
"region:us"
]
| null | 2022-09-19T17:46:12Z | ---
license: mit
---
### F-22 on Stable Diffusion
This is the `<f-22>` concept taught to Stable Diffusion via Textual Inversion. You can load this concept into the [Stable Conceptualizer](https://colab.research.google.com/github/huggingface/notebooks/blob/main/diffusers/stable_conceptualizer_inference.ipynb) notebook. You can also train your own concepts and load them into the concept libraries using [this notebook](https://colab.research.google.com/github/huggingface/notebooks/blob/main/diffusers/sd_textual_inversion_training.ipynb).
Here is the new concept you will be able to use as an `object`:
|
research-backup/roberta-large-semeval2012-mask-prompt-c-triplet | research-backup | 2022-09-19T17:00:24Z | 104 | 0 | transformers | [
"transformers",
"pytorch",
"roberta",
"feature-extraction",
"dataset:semeval2012",
"model-index",
"text-embeddings-inference",
"endpoints_compatible",
"region:us"
]
| feature-extraction | 2022-07-24T13:34:08Z | ---
datasets:
- semeval2012
model-index:
- name: relbert/roberta-large-semeval2012-mask-prompt-c-triplet
results:
- task:
name: Relation Mapping
type: sorting-task
dataset:
name: Relation Mapping
args: relbert/relation_mapping
type: relation-mapping
metrics:
- name: Accuracy
type: accuracy
value: 0.7999801587301587
- task:
name: Analogy Questions (SAT full)
type: multiple-choice-qa
dataset:
name: SAT full
args: relbert/analogy_questions
type: analogy-questions
metrics:
- name: Accuracy
type: accuracy
value: 0.4999999999999999
- task:
name: Analogy Questions (SAT)
type: multiple-choice-qa
dataset:
name: SAT
args: relbert/analogy_questions
type: analogy-questions
metrics:
- name: Accuracy
type: accuracy
value: 0.516320474777448
- task:
name: Analogy Questions (BATS)
type: multiple-choice-qa
dataset:
name: BATS
args: relbert/analogy_questions
type: analogy-questions
metrics:
- name: Accuracy
type: accuracy
value: 0.6842690383546415
- task:
name: Analogy Questions (Google)
type: multiple-choice-qa
dataset:
name: Google
args: relbert/analogy_questions
type: analogy-questions
metrics:
- name: Accuracy
type: accuracy
value: 0.866
- task:
name: Analogy Questions (U2)
type: multiple-choice-qa
dataset:
name: U2
args: relbert/analogy_questions
type: analogy-questions
metrics:
- name: Accuracy
type: accuracy
value: 0.4605263157894737
- task:
name: Analogy Questions (U4)
type: multiple-choice-qa
dataset:
name: U4
args: relbert/analogy_questions
type: analogy-questions
metrics:
- name: Accuracy
type: accuracy
value: 0.5138888888888888
- task:
name: Lexical Relation Classification (BLESS)
type: classification
dataset:
name: BLESS
args: relbert/lexical_relation_classification
type: relation-classification
metrics:
- name: F1
type: f1
value: 0.9273768268796143
- name: F1 (macro)
type: f1_macro
value: 0.9225150711749914
- task:
name: Lexical Relation Classification (CogALexV)
type: classification
dataset:
name: CogALexV
args: relbert/lexical_relation_classification
type: relation-classification
metrics:
- name: F1
type: f1
value: 0.8720657276995305
- name: F1 (macro)
type: f1_macro
value: 0.7091947759445287
- task:
name: Lexical Relation Classification (EVALution)
type: classification
dataset:
name: BLESS
args: relbert/lexical_relation_classification
type: relation-classification
metrics:
- name: F1
type: f1
value: 0.7215601300108341
- name: F1 (macro)
type: f1_macro
value: 0.70570510560382
- task:
name: Lexical Relation Classification (K&H+N)
type: classification
dataset:
name: K&H+N
args: relbert/lexical_relation_classification
type: relation-classification
metrics:
- name: F1
type: f1
value: 0.9681435626347639
- name: F1 (macro)
type: f1_macro
value: 0.9054090342043304
- task:
name: Lexical Relation Classification (ROOT09)
type: classification
dataset:
name: ROOT09
args: relbert/lexical_relation_classification
type: relation-classification
metrics:
- name: F1
type: f1
value: 0.9185208398621121
- name: F1 (macro)
type: f1_macro
value: 0.917840745705091
---
# relbert/roberta-large-semeval2012-mask-prompt-c-triplet
RelBERT fine-tuned from [roberta-large](https://huggingface.co/roberta-large) on
[semeval2012](https://huggingface.co/datasets/semeval2012).
Fine-tuning is done via [RelBERT](https://github.com/asahi417/relbert) library (see the repository for more detail).
It achieves the following results on the relation understanding tasks:
- Analogy Question ([dataset](https://huggingface.co/datasets/relbert/analogy_questions), [full result](https://huggingface.co/relbert/roberta-large-semeval2012-mask-prompt-c-triplet/raw/main/analogy.json)):
- Accuracy on SAT (full): 0.4999999999999999
- Accuracy on SAT: 0.516320474777448
- Accuracy on BATS: 0.6842690383546415
- Accuracy on U2: 0.4605263157894737
- Accuracy on U4: 0.5138888888888888
- Accuracy on Google: 0.866
- Lexical Relation Classification ([dataset](https://huggingface.co/datasets/relbert/lexical_relation_classification), [full result](https://huggingface.co/relbert/roberta-large-semeval2012-mask-prompt-c-triplet/raw/main/classification.json)):
- Micro F1 score on BLESS: 0.9273768268796143
- Micro F1 score on CogALexV: 0.8720657276995305
- Micro F1 score on EVALution: 0.7215601300108341
- Micro F1 score on K&H+N: 0.9681435626347639
- Micro F1 score on ROOT09: 0.9185208398621121
- Relation Mapping ([dataset](https://huggingface.co/datasets/relbert/relation_mapping), [full result](https://huggingface.co/relbert/roberta-large-semeval2012-mask-prompt-c-triplet/raw/main/relation_mapping.json)):
- Accuracy on Relation Mapping: 0.7999801587301587
### Usage
This model can be used through the [relbert library](https://github.com/asahi417/relbert). Install the library via pip
```shell
pip install relbert
```
and activate model as below.
```python
from relbert import RelBERT
model = RelBERT("relbert/roberta-large-semeval2012-mask-prompt-c-triplet")
vector = model.get_embedding(['Tokyo', 'Japan']) # shape of (1024, )
```
### Training hyperparameters
The following hyperparameters were used during training:
- model: roberta-large
- max_length: 64
- mode: mask
- data: semeval2012
- n_sample: 10
- custom_template: Today, I finally discovered the relation between <subj> and <obj> : <mask>
- template: None
- softmax_loss: True
- in_batch_negative: True
- parent_contrast: True
- mse_margin: 1
- epoch: 1
- lr_warmup: 10
- batch: 64
- lr: 2e-05
- lr_decay: False
- weight_decay: 0
- optimizer: adam
- momentum: 0.9
- fp16: False
- random_seed: 0
The full configuration can be found at [fine-tuning parameter file](https://huggingface.co/relbert/roberta-large-semeval2012-mask-prompt-c-triplet/raw/main/trainer_config.json).
### Reference
If you use any resource from RelBERT, please consider to cite our [paper](https://aclanthology.org/2021.eacl-demos.7/).
```
@inproceedings{ushio-etal-2021-distilling-relation-embeddings,
title = "{D}istilling {R}elation {E}mbeddings from {P}re-trained {L}anguage {M}odels",
author = "Ushio, Asahi and
Schockaert, Steven and
Camacho-Collados, Jose",
booktitle = "EMNLP 2021",
year = "2021",
address = "Online",
publisher = "Association for Computational Linguistics",
}
```
|
research-backup/roberta-large-semeval2012-average-prompt-d-triplet | research-backup | 2022-09-19T16:47:34Z | 105 | 0 | transformers | [
"transformers",
"pytorch",
"roberta",
"feature-extraction",
"dataset:semeval2012",
"model-index",
"text-embeddings-inference",
"endpoints_compatible",
"region:us"
]
| feature-extraction | 2022-07-24T10:45:27Z | ---
datasets:
- semeval2012
model-index:
- name: relbert/roberta-large-semeval2012-average-prompt-d-triplet
results:
- task:
name: Relation Mapping
type: sorting-task
dataset:
name: Relation Mapping
args: relbert/relation_mapping
type: relation-mapping
metrics:
- name: Accuracy
type: accuracy
value: 0.8641666666666666
- task:
name: Analogy Questions (SAT full)
type: multiple-choice-qa
dataset:
name: SAT full
args: relbert/analogy_questions
type: analogy-questions
metrics:
- name: Accuracy
type: accuracy
value: 0.6737967914438503
- task:
name: Analogy Questions (SAT)
type: multiple-choice-qa
dataset:
name: SAT
args: relbert/analogy_questions
type: analogy-questions
metrics:
- name: Accuracy
type: accuracy
value: 0.6795252225519288
- task:
name: Analogy Questions (BATS)
type: multiple-choice-qa
dataset:
name: BATS
args: relbert/analogy_questions
type: analogy-questions
metrics:
- name: Accuracy
type: accuracy
value: 0.7859922178988327
- task:
name: Analogy Questions (Google)
type: multiple-choice-qa
dataset:
name: Google
args: relbert/analogy_questions
type: analogy-questions
metrics:
- name: Accuracy
type: accuracy
value: 0.938
- task:
name: Analogy Questions (U2)
type: multiple-choice-qa
dataset:
name: U2
args: relbert/analogy_questions
type: analogy-questions
metrics:
- name: Accuracy
type: accuracy
value: 0.6535087719298246
- task:
name: Analogy Questions (U4)
type: multiple-choice-qa
dataset:
name: U4
args: relbert/analogy_questions
type: analogy-questions
metrics:
- name: Accuracy
type: accuracy
value: 0.6550925925925926
- task:
name: Lexical Relation Classification (BLESS)
type: classification
dataset:
name: BLESS
args: relbert/lexical_relation_classification
type: relation-classification
metrics:
- name: F1
type: f1
value: 0.9252674401084827
- name: F1 (macro)
type: f1_macro
value: 0.9199095894826357
- task:
name: Lexical Relation Classification (CogALexV)
type: classification
dataset:
name: CogALexV
args: relbert/lexical_relation_classification
type: relation-classification
metrics:
- name: F1
type: f1
value: 0.8692488262910798
- name: F1 (macro)
type: f1_macro
value: 0.7160574925535569
- task:
name: Lexical Relation Classification (EVALution)
type: classification
dataset:
name: BLESS
args: relbert/lexical_relation_classification
type: relation-classification
metrics:
- name: F1
type: f1
value: 0.7009750812567714
- name: F1 (macro)
type: f1_macro
value: 0.6878313451654309
- task:
name: Lexical Relation Classification (K&H+N)
type: classification
dataset:
name: K&H+N
args: relbert/lexical_relation_classification
type: relation-classification
metrics:
- name: F1
type: f1
value: 0.9683522292550601
- name: F1 (macro)
type: f1_macro
value: 0.8957380868540458
- task:
name: Lexical Relation Classification (ROOT09)
type: classification
dataset:
name: ROOT09
args: relbert/lexical_relation_classification
type: relation-classification
metrics:
- name: F1
type: f1
value: 0.9009714822939517
- name: F1 (macro)
type: f1_macro
value: 0.9036194719551146
---
# relbert/roberta-large-semeval2012-average-prompt-d-triplet
RelBERT fine-tuned from [roberta-large](https://huggingface.co/roberta-large) on
[semeval2012](https://huggingface.co/datasets/semeval2012).
Fine-tuning is done via [RelBERT](https://github.com/asahi417/relbert) library (see the repository for more detail).
It achieves the following results on the relation understanding tasks:
- Analogy Question ([dataset](https://huggingface.co/datasets/relbert/analogy_questions), [full result](https://huggingface.co/relbert/roberta-large-semeval2012-average-prompt-d-triplet/raw/main/analogy.json)):
- Accuracy on SAT (full): 0.6737967914438503
- Accuracy on SAT: 0.6795252225519288
- Accuracy on BATS: 0.7859922178988327
- Accuracy on U2: 0.6535087719298246
- Accuracy on U4: 0.6550925925925926
- Accuracy on Google: 0.938
- Lexical Relation Classification ([dataset](https://huggingface.co/datasets/relbert/lexical_relation_classification), [full result](https://huggingface.co/relbert/roberta-large-semeval2012-average-prompt-d-triplet/raw/main/classification.json)):
- Micro F1 score on BLESS: 0.9252674401084827
- Micro F1 score on CogALexV: 0.8692488262910798
- Micro F1 score on EVALution: 0.7009750812567714
- Micro F1 score on K&H+N: 0.9683522292550601
- Micro F1 score on ROOT09: 0.9009714822939517
- Relation Mapping ([dataset](https://huggingface.co/datasets/relbert/relation_mapping), [full result](https://huggingface.co/relbert/roberta-large-semeval2012-average-prompt-d-triplet/raw/main/relation_mapping.json)):
- Accuracy on Relation Mapping: 0.8641666666666666
### Usage
This model can be used through the [relbert library](https://github.com/asahi417/relbert). Install the library via pip
```shell
pip install relbert
```
and activate model as below.
```python
from relbert import RelBERT
model = RelBERT("relbert/roberta-large-semeval2012-average-prompt-d-triplet")
vector = model.get_embedding(['Tokyo', 'Japan']) # shape of (1024, )
```
### Training hyperparameters
The following hyperparameters were used during training:
- model: roberta-large
- max_length: 64
- mode: average
- data: semeval2012
- n_sample: 10
- custom_template: I wasn’t aware of this relationship, but I just read in the encyclopedia that <subj> is the <mask> of <obj>
- template: None
- softmax_loss: True
- in_batch_negative: True
- parent_contrast: True
- mse_margin: 1
- epoch: 1
- lr_warmup: 10
- batch: 64
- lr: 2e-05
- lr_decay: False
- weight_decay: 0
- optimizer: adam
- momentum: 0.9
- fp16: False
- random_seed: 0
The full configuration can be found at [fine-tuning parameter file](https://huggingface.co/relbert/roberta-large-semeval2012-average-prompt-d-triplet/raw/main/trainer_config.json).
### Reference
If you use any resource from RelBERT, please consider to cite our [paper](https://aclanthology.org/2021.eacl-demos.7/).
```
@inproceedings{ushio-etal-2021-distilling-relation-embeddings,
title = "{D}istilling {R}elation {E}mbeddings from {P}re-trained {L}anguage {M}odels",
author = "Ushio, Asahi and
Schockaert, Steven and
Camacho-Collados, Jose",
booktitle = "EMNLP 2021",
year = "2021",
address = "Online",
publisher = "Association for Computational Linguistics",
}
```
|
clboetticher-school/xlm-roberta-base-finetuned-panx-fr | clboetticher-school | 2022-09-19T16:46:23Z | 114 | 0 | transformers | [
"transformers",
"pytorch",
"xlm-roberta",
"token-classification",
"generated_from_trainer",
"dataset:xtreme",
"license:mit",
"model-index",
"autotrain_compatible",
"endpoints_compatible",
"region:us"
]
| token-classification | 2022-09-19T16:27:36Z | ---
license: mit
tags:
- generated_from_trainer
datasets:
- xtreme
metrics:
- f1
model-index:
- name: xlm-roberta-base-finetuned-panx-fr
results:
- task:
name: Token Classification
type: token-classification
dataset:
name: xtreme
type: xtreme
args: PAN-X.fr
metrics:
- name: F1
type: f1
value: 0.9213082901554404
---
<!-- This model card has been generated automatically according to the information the Trainer had access to. You
should probably proofread and complete it, then remove this comment. -->
# xlm-roberta-base-finetuned-panx-fr
This model is a fine-tuned version of [xlm-roberta-base](https://huggingface.co/xlm-roberta-base) on the xtreme dataset.
It achieves the following results on the evaluation set:
- Loss: 0.1117
- F1: 0.9213
## Model description
More information needed
## Intended uses & limitations
More information needed
## Training and evaluation data
More information needed
## Training procedure
### Training hyperparameters
The following hyperparameters were used during training:
- learning_rate: 5e-05
- train_batch_size: 24
- eval_batch_size: 24
- seed: 42
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
- lr_scheduler_type: linear
- num_epochs: 3
### Training results
| Training Loss | Epoch | Step | Validation Loss | F1 |
|:-------------:|:-----:|:----:|:---------------:|:------:|
| 0.5779 | 1.0 | 191 | 0.2832 | 0.8091 |
| 0.2735 | 2.0 | 382 | 0.1570 | 0.8943 |
| 0.1769 | 3.0 | 573 | 0.1117 | 0.9213 |
### Framework versions
- Transformers 4.11.3
- Pytorch 1.12.1+cu113
- Datasets 1.16.1
- Tokenizers 0.10.3
|
research-backup/roberta-large-semeval2012-average-prompt-c-triplet | research-backup | 2022-09-19T16:43:19Z | 103 | 0 | transformers | [
"transformers",
"pytorch",
"roberta",
"feature-extraction",
"dataset:semeval2012",
"model-index",
"text-embeddings-inference",
"endpoints_compatible",
"region:us"
]
| feature-extraction | 2022-07-24T10:43:59Z | ---
datasets:
- semeval2012
model-index:
- name: relbert/roberta-large-semeval2012-average-prompt-c-triplet
results:
- task:
name: Relation Mapping
type: sorting-task
dataset:
name: Relation Mapping
args: relbert/relation_mapping
type: relation-mapping
metrics:
- name: Accuracy
type: accuracy
value: 0.8742857142857143
- task:
name: Analogy Questions (SAT full)
type: multiple-choice-qa
dataset:
name: SAT full
args: relbert/analogy_questions
type: analogy-questions
metrics:
- name: Accuracy
type: accuracy
value: 0.5748663101604278
- task:
name: Analogy Questions (SAT)
type: multiple-choice-qa
dataset:
name: SAT
args: relbert/analogy_questions
type: analogy-questions
metrics:
- name: Accuracy
type: accuracy
value: 0.5786350148367952
- task:
name: Analogy Questions (BATS)
type: multiple-choice-qa
dataset:
name: BATS
args: relbert/analogy_questions
type: analogy-questions
metrics:
- name: Accuracy
type: accuracy
value: 0.820455808782657
- task:
name: Analogy Questions (Google)
type: multiple-choice-qa
dataset:
name: Google
args: relbert/analogy_questions
type: analogy-questions
metrics:
- name: Accuracy
type: accuracy
value: 0.918
- task:
name: Analogy Questions (U2)
type: multiple-choice-qa
dataset:
name: U2
args: relbert/analogy_questions
type: analogy-questions
metrics:
- name: Accuracy
type: accuracy
value: 0.6140350877192983
- task:
name: Analogy Questions (U4)
type: multiple-choice-qa
dataset:
name: U4
args: relbert/analogy_questions
type: analogy-questions
metrics:
- name: Accuracy
type: accuracy
value: 0.6319444444444444
- task:
name: Lexical Relation Classification (BLESS)
type: classification
dataset:
name: BLESS
args: relbert/lexical_relation_classification
type: relation-classification
metrics:
- name: F1
type: f1
value: 0.924815428657526
- name: F1 (macro)
type: f1_macro
value: 0.9202612346118308
- task:
name: Lexical Relation Classification (CogALexV)
type: classification
dataset:
name: CogALexV
args: relbert/lexical_relation_classification
type: relation-classification
metrics:
- name: F1
type: f1
value: 0.8718309859154929
- name: F1 (macro)
type: f1_macro
value: 0.7152177972947781
- task:
name: Lexical Relation Classification (EVALution)
type: classification
dataset:
name: BLESS
args: relbert/lexical_relation_classification
type: relation-classification
metrics:
- name: F1
type: f1
value: 0.704225352112676
- name: F1 (macro)
type: f1_macro
value: 0.6846186699994758
- task:
name: Lexical Relation Classification (K&H+N)
type: classification
dataset:
name: K&H+N
args: relbert/lexical_relation_classification
type: relation-classification
metrics:
- name: F1
type: f1
value: 0.9667524518327885
- name: F1 (macro)
type: f1_macro
value: 0.8963571819720165
- task:
name: Lexical Relation Classification (ROOT09)
type: classification
dataset:
name: ROOT09
args: relbert/lexical_relation_classification
type: relation-classification
metrics:
- name: F1
type: f1
value: 0.9015982450642431
- name: F1 (macro)
type: f1_macro
value: 0.8974114781326906
---
# relbert/roberta-large-semeval2012-average-prompt-c-triplet
RelBERT fine-tuned from [roberta-large](https://huggingface.co/roberta-large) on
[semeval2012](https://huggingface.co/datasets/semeval2012).
Fine-tuning is done via [RelBERT](https://github.com/asahi417/relbert) library (see the repository for more detail).
It achieves the following results on the relation understanding tasks:
- Analogy Question ([dataset](https://huggingface.co/datasets/relbert/analogy_questions), [full result](https://huggingface.co/relbert/roberta-large-semeval2012-average-prompt-c-triplet/raw/main/analogy.json)):
- Accuracy on SAT (full): 0.5748663101604278
- Accuracy on SAT: 0.5786350148367952
- Accuracy on BATS: 0.820455808782657
- Accuracy on U2: 0.6140350877192983
- Accuracy on U4: 0.6319444444444444
- Accuracy on Google: 0.918
- Lexical Relation Classification ([dataset](https://huggingface.co/datasets/relbert/lexical_relation_classification), [full result](https://huggingface.co/relbert/roberta-large-semeval2012-average-prompt-c-triplet/raw/main/classification.json)):
- Micro F1 score on BLESS: 0.924815428657526
- Micro F1 score on CogALexV: 0.8718309859154929
- Micro F1 score on EVALution: 0.704225352112676
- Micro F1 score on K&H+N: 0.9667524518327885
- Micro F1 score on ROOT09: 0.9015982450642431
- Relation Mapping ([dataset](https://huggingface.co/datasets/relbert/relation_mapping), [full result](https://huggingface.co/relbert/roberta-large-semeval2012-average-prompt-c-triplet/raw/main/relation_mapping.json)):
- Accuracy on Relation Mapping: 0.8742857142857143
### Usage
This model can be used through the [relbert library](https://github.com/asahi417/relbert). Install the library via pip
```shell
pip install relbert
```
and activate model as below.
```python
from relbert import RelBERT
model = RelBERT("relbert/roberta-large-semeval2012-average-prompt-c-triplet")
vector = model.get_embedding(['Tokyo', 'Japan']) # shape of (1024, )
```
### Training hyperparameters
The following hyperparameters were used during training:
- model: roberta-large
- max_length: 64
- mode: average
- data: semeval2012
- n_sample: 10
- custom_template: Today, I finally discovered the relation between <subj> and <obj> : <mask>
- template: None
- softmax_loss: True
- in_batch_negative: True
- parent_contrast: True
- mse_margin: 1
- epoch: 1
- lr_warmup: 10
- batch: 64
- lr: 2e-05
- lr_decay: False
- weight_decay: 0
- optimizer: adam
- momentum: 0.9
- fp16: False
- random_seed: 0
The full configuration can be found at [fine-tuning parameter file](https://huggingface.co/relbert/roberta-large-semeval2012-average-prompt-c-triplet/raw/main/trainer_config.json).
### Reference
If you use any resource from RelBERT, please consider to cite our [paper](https://aclanthology.org/2021.eacl-demos.7/).
```
@inproceedings{ushio-etal-2021-distilling-relation-embeddings,
title = "{D}istilling {R}elation {E}mbeddings from {P}re-trained {L}anguage {M}odels",
author = "Ushio, Asahi and
Schockaert, Steven and
Camacho-Collados, Jose",
booktitle = "EMNLP 2021",
year = "2021",
address = "Online",
publisher = "Association for Computational Linguistics",
}
```
|
research-backup/roberta-large-semeval2012-average-prompt-b-triplet | research-backup | 2022-09-19T16:39:47Z | 103 | 0 | transformers | [
"transformers",
"pytorch",
"roberta",
"feature-extraction",
"dataset:semeval2012",
"model-index",
"text-embeddings-inference",
"endpoints_compatible",
"region:us"
]
| feature-extraction | 2022-07-24T10:42:35Z | ---
datasets:
- semeval2012
model-index:
- name: relbert/roberta-large-semeval2012-average-prompt-b-triplet
results:
- task:
name: Relation Mapping
type: sorting-task
dataset:
name: Relation Mapping
args: relbert/relation_mapping
type: relation-mapping
metrics:
- name: Accuracy
type: accuracy
value: 0.815952380952381
- task:
name: Analogy Questions (SAT full)
type: multiple-choice-qa
dataset:
name: SAT full
args: relbert/analogy_questions
type: analogy-questions
metrics:
- name: Accuracy
type: accuracy
value: 0.5748663101604278
- task:
name: Analogy Questions (SAT)
type: multiple-choice-qa
dataset:
name: SAT
args: relbert/analogy_questions
type: analogy-questions
metrics:
- name: Accuracy
type: accuracy
value: 0.5756676557863502
- task:
name: Analogy Questions (BATS)
type: multiple-choice-qa
dataset:
name: BATS
args: relbert/analogy_questions
type: analogy-questions
metrics:
- name: Accuracy
type: accuracy
value: 0.7809894385769872
- task:
name: Analogy Questions (Google)
type: multiple-choice-qa
dataset:
name: Google
args: relbert/analogy_questions
type: analogy-questions
metrics:
- name: Accuracy
type: accuracy
value: 0.87
- task:
name: Analogy Questions (U2)
type: multiple-choice-qa
dataset:
name: U2
args: relbert/analogy_questions
type: analogy-questions
metrics:
- name: Accuracy
type: accuracy
value: 0.5570175438596491
- task:
name: Analogy Questions (U4)
type: multiple-choice-qa
dataset:
name: U4
args: relbert/analogy_questions
type: analogy-questions
metrics:
- name: Accuracy
type: accuracy
value: 0.5763888888888888
- task:
name: Lexical Relation Classification (BLESS)
type: classification
dataset:
name: BLESS
args: relbert/lexical_relation_classification
type: relation-classification
metrics:
- name: F1
type: f1
value: 0.9156245291547386
- name: F1 (macro)
type: f1_macro
value: 0.9123138480377561
- task:
name: Lexical Relation Classification (CogALexV)
type: classification
dataset:
name: CogALexV
args: relbert/lexical_relation_classification
type: relation-classification
metrics:
- name: F1
type: f1
value: 0.8652582159624412
- name: F1 (macro)
type: f1_macro
value: 0.7098768153077847
- task:
name: Lexical Relation Classification (EVALution)
type: classification
dataset:
name: BLESS
args: relbert/lexical_relation_classification
type: relation-classification
metrics:
- name: F1
type: f1
value: 0.6765980498374865
- name: F1 (macro)
type: f1_macro
value: 0.667723188418867
- task:
name: Lexical Relation Classification (K&H+N)
type: classification
dataset:
name: K&H+N
args: relbert/lexical_relation_classification
type: relation-classification
metrics:
- name: F1
type: f1
value: 0.9621617861862697
- name: F1 (macro)
type: f1_macro
value: 0.8800726259971795
- task:
name: Lexical Relation Classification (ROOT09)
type: classification
dataset:
name: ROOT09
args: relbert/lexical_relation_classification
type: relation-classification
metrics:
- name: F1
type: f1
value: 0.8925101848950172
- name: F1 (macro)
type: f1_macro
value: 0.8890641447568232
---
# relbert/roberta-large-semeval2012-average-prompt-b-triplet
RelBERT fine-tuned from [roberta-large](https://huggingface.co/roberta-large) on
[semeval2012](https://huggingface.co/datasets/semeval2012).
Fine-tuning is done via [RelBERT](https://github.com/asahi417/relbert) library (see the repository for more detail).
It achieves the following results on the relation understanding tasks:
- Analogy Question ([dataset](https://huggingface.co/datasets/relbert/analogy_questions), [full result](https://huggingface.co/relbert/roberta-large-semeval2012-average-prompt-b-triplet/raw/main/analogy.json)):
- Accuracy on SAT (full): 0.5748663101604278
- Accuracy on SAT: 0.5756676557863502
- Accuracy on BATS: 0.7809894385769872
- Accuracy on U2: 0.5570175438596491
- Accuracy on U4: 0.5763888888888888
- Accuracy on Google: 0.87
- Lexical Relation Classification ([dataset](https://huggingface.co/datasets/relbert/lexical_relation_classification), [full result](https://huggingface.co/relbert/roberta-large-semeval2012-average-prompt-b-triplet/raw/main/classification.json)):
- Micro F1 score on BLESS: 0.9156245291547386
- Micro F1 score on CogALexV: 0.8652582159624412
- Micro F1 score on EVALution: 0.6765980498374865
- Micro F1 score on K&H+N: 0.9621617861862697
- Micro F1 score on ROOT09: 0.8925101848950172
- Relation Mapping ([dataset](https://huggingface.co/datasets/relbert/relation_mapping), [full result](https://huggingface.co/relbert/roberta-large-semeval2012-average-prompt-b-triplet/raw/main/relation_mapping.json)):
- Accuracy on Relation Mapping: 0.815952380952381
### Usage
This model can be used through the [relbert library](https://github.com/asahi417/relbert). Install the library via pip
```shell
pip install relbert
```
and activate model as below.
```python
from relbert import RelBERT
model = RelBERT("relbert/roberta-large-semeval2012-average-prompt-b-triplet")
vector = model.get_embedding(['Tokyo', 'Japan']) # shape of (1024, )
```
### Training hyperparameters
The following hyperparameters were used during training:
- model: roberta-large
- max_length: 64
- mode: average
- data: semeval2012
- n_sample: 10
- custom_template: Today, I finally discovered the relation between <subj> and <obj> : <obj> is <subj>'s <mask>
- template: None
- softmax_loss: True
- in_batch_negative: True
- parent_contrast: True
- mse_margin: 1
- epoch: 1
- lr_warmup: 10
- batch: 64
- lr: 2e-05
- lr_decay: False
- weight_decay: 0
- optimizer: adam
- momentum: 0.9
- fp16: False
- random_seed: 0
The full configuration can be found at [fine-tuning parameter file](https://huggingface.co/relbert/roberta-large-semeval2012-average-prompt-b-triplet/raw/main/trainer_config.json).
### Reference
If you use any resource from RelBERT, please consider to cite our [paper](https://aclanthology.org/2021.eacl-demos.7/).
```
@inproceedings{ushio-etal-2021-distilling-relation-embeddings,
title = "{D}istilling {R}elation {E}mbeddings from {P}re-trained {L}anguage {M}odels",
author = "Ushio, Asahi and
Schockaert, Steven and
Camacho-Collados, Jose",
booktitle = "EMNLP 2021",
year = "2021",
address = "Online",
publisher = "Association for Computational Linguistics",
}
```
|
research-backup/roberta-large-semeval2012-average-prompt-e-triplet | research-backup | 2022-09-19T16:31:56Z | 96 | 0 | transformers | [
"transformers",
"pytorch",
"roberta",
"feature-extraction",
"dataset:semeval2012",
"model-index",
"text-embeddings-inference",
"endpoints_compatible",
"region:us"
]
| feature-extraction | 2022-07-22T18:25:55Z | ---
datasets:
- semeval2012
model-index:
- name: relbert/roberta-large-semeval2012-average-prompt-e-triplet
results:
- task:
name: Relation Mapping
type: sorting-task
dataset:
name: Relation Mapping
args: relbert/relation_mapping
type: relation-mapping
metrics:
- name: Accuracy
type: accuracy
value: 0.9255952380952381
- task:
name: Analogy Questions (SAT full)
type: multiple-choice-qa
dataset:
name: SAT full
args: relbert/analogy_questions
type: analogy-questions
metrics:
- name: Accuracy
type: accuracy
value: 0.550802139037433
- task:
name: Analogy Questions (SAT)
type: multiple-choice-qa
dataset:
name: SAT
args: relbert/analogy_questions
type: analogy-questions
metrics:
- name: Accuracy
type: accuracy
value: 0.5548961424332344
- task:
name: Analogy Questions (BATS)
type: multiple-choice-qa
dataset:
name: BATS
args: relbert/analogy_questions
type: analogy-questions
metrics:
- name: Accuracy
type: accuracy
value: 0.754863813229572
- task:
name: Analogy Questions (Google)
type: multiple-choice-qa
dataset:
name: Google
args: relbert/analogy_questions
type: analogy-questions
metrics:
- name: Accuracy
type: accuracy
value: 0.872
- task:
name: Analogy Questions (U2)
type: multiple-choice-qa
dataset:
name: U2
args: relbert/analogy_questions
type: analogy-questions
metrics:
- name: Accuracy
type: accuracy
value: 0.631578947368421
- task:
name: Analogy Questions (U4)
type: multiple-choice-qa
dataset:
name: U4
args: relbert/analogy_questions
type: analogy-questions
metrics:
- name: Accuracy
type: accuracy
value: 0.5740740740740741
- task:
name: Lexical Relation Classification (BLESS)
type: classification
dataset:
name: BLESS
args: relbert/lexical_relation_classification
type: relation-classification
metrics:
- name: F1
type: f1
value: 0.9031188790116016
- name: F1 (macro)
type: f1_macro
value: 0.896130708234777
- task:
name: Lexical Relation Classification (CogALexV)
type: classification
dataset:
name: CogALexV
args: relbert/lexical_relation_classification
type: relation-classification
metrics:
- name: F1
type: f1
value: 0.8692488262910798
- name: F1 (macro)
type: f1_macro
value: 0.7176923678524982
- task:
name: Lexical Relation Classification (EVALution)
type: classification
dataset:
name: BLESS
args: relbert/lexical_relation_classification
type: relation-classification
metrics:
- name: F1
type: f1
value: 0.6933911159263272
- name: F1 (macro)
type: f1_macro
value: 0.6793569940444139
- task:
name: Lexical Relation Classification (K&H+N)
type: classification
dataset:
name: K&H+N
args: relbert/lexical_relation_classification
type: relation-classification
metrics:
- name: F1
type: f1
value: 0.9611184530847882
- name: F1 (macro)
type: f1_macro
value: 0.8913615954101612
- task:
name: Lexical Relation Classification (ROOT09)
type: classification
dataset:
name: ROOT09
args: relbert/lexical_relation_classification
type: relation-classification
metrics:
- name: F1
type: f1
value: 0.9088060169225948
- name: F1 (macro)
type: f1_macro
value: 0.9076649125882928
---
# relbert/roberta-large-semeval2012-average-prompt-e-triplet
RelBERT fine-tuned from [roberta-large](https://huggingface.co/roberta-large) on
[semeval2012](https://huggingface.co/datasets/semeval2012).
Fine-tuning is done via [RelBERT](https://github.com/asahi417/relbert) library (see the repository for more detail).
It achieves the following results on the relation understanding tasks:
- Analogy Question ([dataset](https://huggingface.co/datasets/relbert/analogy_questions), [full result](https://huggingface.co/relbert/roberta-large-semeval2012-average-prompt-e-triplet/raw/main/analogy.json)):
- Accuracy on SAT (full): 0.550802139037433
- Accuracy on SAT: 0.5548961424332344
- Accuracy on BATS: 0.754863813229572
- Accuracy on U2: 0.631578947368421
- Accuracy on U4: 0.5740740740740741
- Accuracy on Google: 0.872
- Lexical Relation Classification ([dataset](https://huggingface.co/datasets/relbert/lexical_relation_classification), [full result](https://huggingface.co/relbert/roberta-large-semeval2012-average-prompt-e-triplet/raw/main/classification.json)):
- Micro F1 score on BLESS: 0.9031188790116016
- Micro F1 score on CogALexV: 0.8692488262910798
- Micro F1 score on EVALution: 0.6933911159263272
- Micro F1 score on K&H+N: 0.9611184530847882
- Micro F1 score on ROOT09: 0.9088060169225948
- Relation Mapping ([dataset](https://huggingface.co/datasets/relbert/relation_mapping), [full result](https://huggingface.co/relbert/roberta-large-semeval2012-average-prompt-e-triplet/raw/main/relation_mapping.json)):
- Accuracy on Relation Mapping: 0.9255952380952381
### Usage
This model can be used through the [relbert library](https://github.com/asahi417/relbert). Install the library via pip
```shell
pip install relbert
```
and activate model as below.
```python
from relbert import RelBERT
model = RelBERT("relbert/roberta-large-semeval2012-average-prompt-e-triplet")
vector = model.get_embedding(['Tokyo', 'Japan']) # shape of (1024, )
```
### Training hyperparameters
The following hyperparameters were used during training:
- model: roberta-large
- max_length: 64
- mode: average
- data: semeval2012
- n_sample: 10
- custom_template: I wasn’t aware of this relationship, but I just read in the encyclopedia that <obj> is <subj>’s <mask>
- template: None
- softmax_loss: True
- in_batch_negative: True
- parent_contrast: True
- mse_margin: 1
- epoch: 1
- lr_warmup: 10
- batch: 64
- lr: 2e-05
- lr_decay: False
- weight_decay: 0
- optimizer: adam
- momentum: 0.9
- fp16: False
- random_seed: 0
The full configuration can be found at [fine-tuning parameter file](https://huggingface.co/relbert/roberta-large-semeval2012-average-prompt-e-triplet/raw/main/trainer_config.json).
### Reference
If you use any resource from RelBERT, please consider to cite our [paper](https://aclanthology.org/2021.eacl-demos.7/).
```
@inproceedings{ushio-etal-2021-distilling-relation-embeddings,
title = "{D}istilling {R}elation {E}mbeddings from {P}re-trained {L}anguage {M}odels",
author = "Ushio, Asahi and
Schockaert, Steven and
Camacho-Collados, Jose",
booktitle = "EMNLP 2021",
year = "2021",
address = "Online",
publisher = "Association for Computational Linguistics",
}
```
|
research-backup/roberta-large-semeval2012-average-no-mask-prompt-d-triplet | research-backup | 2022-09-19T16:23:34Z | 104 | 0 | transformers | [
"transformers",
"pytorch",
"roberta",
"feature-extraction",
"dataset:semeval2012",
"model-index",
"text-embeddings-inference",
"endpoints_compatible",
"region:us"
]
| feature-extraction | 2022-07-22T16:52:00Z | ---
datasets:
- semeval2012
model-index:
- name: relbert/roberta-large-semeval2012-average-no-mask-prompt-d-triplet
results:
- task:
name: Relation Mapping
type: sorting-task
dataset:
name: Relation Mapping
args: relbert/relation_mapping
type: relation-mapping
metrics:
- name: Accuracy
type: accuracy
value: 0.815952380952381
- task:
name: Analogy Questions (SAT full)
type: multiple-choice-qa
dataset:
name: SAT full
args: relbert/analogy_questions
type: analogy-questions
metrics:
- name: Accuracy
type: accuracy
value: 0.6951871658
- task:
name: Analogy Questions (SAT)
type: multiple-choice-qa
dataset:
name: SAT
args: relbert/analogy_questions
type: analogy-questions
metrics:
- name: Accuracy
type: accuracy
value: 0.706231454
- task:
name: Analogy Questions (BATS)
type: multiple-choice-qa
dataset:
name: BATS
args: relbert/analogy_questions
type: analogy-questions
metrics:
- name: Accuracy
type: accuracy
value: 0.7882156754
- task:
name: Analogy Questions (Google)
type: multiple-choice-qa
dataset:
name: Google
args: relbert/analogy_questions
type: analogy-questions
metrics:
- name: Accuracy
type: accuracy
value: 0.924
- task:
name: Analogy Questions (U2)
type: multiple-choice-qa
dataset:
name: U2
args: relbert/analogy_questions
type: analogy-questions
metrics:
- name: Accuracy
type: accuracy
value: 0.6622807018
- task:
name: Analogy Questions (U4)
type: multiple-choice-qa
dataset:
name: U4
args: relbert/analogy_questions
type: analogy-questions
metrics:
- name: Accuracy
type: accuracy
value: 0.6527777778
- task:
name: Lexical Relation Classification (BLESS)
type: classification
dataset:
name: BLESS
args: relbert/lexical_relation_classification
type: relation-classification
metrics:
- name: F1
type: f1
value: 0.9153231881874341
- name: F1 (macro)
type: f1_macro
value: 0.9098445625290479
- task:
name: Lexical Relation Classification (CogALexV)
type: classification
dataset:
name: CogALexV
args: relbert/lexical_relation_classification
type: relation-classification
metrics:
- name: F1
type: f1
value: 0.8725352112676056
- name: F1 (macro)
type: f1_macro
value: 0.7174660438773314
- task:
name: Lexical Relation Classification (EVALution)
type: classification
dataset:
name: BLESS
args: relbert/lexical_relation_classification
type: relation-classification
metrics:
- name: F1
type: f1
value: 0.6944745395449621
- name: F1 (macro)
type: f1_macro
value: 0.688951875758847
- task:
name: Lexical Relation Classification (K&H+N)
type: classification
dataset:
name: K&H+N
args: relbert/lexical_relation_classification
type: relation-classification
metrics:
- name: F1
type: f1
value: 0.9689782291159491
- name: F1 (macro)
type: f1_macro
value: 0.90395779327521
- task:
name: Lexical Relation Classification (ROOT09)
type: classification
dataset:
name: ROOT09
args: relbert/lexical_relation_classification
type: relation-classification
metrics:
- name: F1
type: f1
value: 0.9050454403008461
- name: F1 (macro)
type: f1_macro
value: 0.9062415320017446
---
# relbert/roberta-large-semeval2012-average-no-mask-prompt-d-triplet
RelBERT fine-tuned from [roberta-large](https://huggingface.co/roberta-large) on
[semeval2012](https://huggingface.co/datasets/semeval2012).
Fine-tuning is done via [RelBERT](https://github.com/asahi417/relbert) library (see the repository for more detail).
It achieves the following results on the relation understanding tasks:
- Analogy Question ([dataset](https://huggingface.co/datasets/relbert/analogy_questions), [full result](https://huggingface.co/relbert/roberta-large-semeval2012-average-no-mask-prompt-d-triplet/raw/main/analogy.json)):
- Accuracy on SAT (full): 0.6951871658
- Accuracy on SAT: 0.706231454
- Accuracy on BATS: 0.7882156754
- Accuracy on U2: 0.6622807018
- Accuracy on U4: 0.6527777778
- Accuracy on Google: 0.924
- Lexical Relation Classification ([dataset](https://huggingface.co/datasets/relbert/lexical_relation_classification), [full result](https://huggingface.co/relbert/roberta-large-semeval2012-average-no-mask-prompt-d-triplet/raw/main/classification.json)):
- Micro F1 score on BLESS: 0.9153231881874341
- Micro F1 score on CogALexV: 0.8725352112676056
- Micro F1 score on EVALution: 0.6944745395449621
- Micro F1 score on K&H+N: 0.9689782291159491
- Micro F1 score on ROOT09: 0.9050454403008461
- Relation Mapping ([dataset](https://huggingface.co/datasets/relbert/relation_mapping), [full result](https://huggingface.co/relbert/roberta-large-semeval2012-average-no-mask-prompt-d-triplet/raw/main/relation_mapping.json)):
- Accuracy on Relation Mapping: 0.815952380952381
### Usage
This model can be used through the [relbert library](https://github.com/asahi417/relbert). Install the library via pip
```shell
pip install relbert
```
and activate model as below.
```python
from relbert import RelBERT
model = RelBERT("relbert/roberta-large-semeval2012-average-no-mask-prompt-d-triplet")
vector = model.get_embedding(['Tokyo', 'Japan']) # shape of (1024, )
```
### Training hyperparameters
The following hyperparameters were used during training:
- model: roberta-large
- max_length: 64
- mode: average_no_mask
- data: semeval2012
- n_sample: 10
- custom_template: I wasn’t aware of this relationship, but I just read in the encyclopedia that <subj> is the <mask> of <obj>
- template: None
- softmax_loss: True
- in_batch_negative: True
- parent_contrast: True
- mse_margin: 1
- epoch: 1
- lr_warmup: 10
- batch: 64
- lr: 2e-05
- lr_decay: False
- weight_decay: 0
- optimizer: adam
- momentum: 0.9
- fp16: False
- random_seed: 0
The full configuration can be found at [fine-tuning parameter file](https://huggingface.co/relbert/roberta-large-semeval2012-average-no-mask-prompt-d-triplet/raw/main/trainer_config.json).
### Reference
If you use any resource from RelBERT, please consider to cite our [paper](https://aclanthology.org/2021.eacl-demos.7/).
```
@inproceedings{ushio-etal-2021-distilling-relation-embeddings,
title = "{D}istilling {R}elation {E}mbeddings from {P}re-trained {L}anguage {M}odels",
author = "Ushio, Asahi and
Schockaert, Steven and
Camacho-Collados, Jose",
booktitle = "EMNLP 2021",
year = "2021",
address = "Online",
publisher = "Association for Computational Linguistics",
}
```
|
research-backup/roberta-large-semeval2012-average-no-mask-prompt-c-triplet | research-backup | 2022-09-19T16:19:14Z | 104 | 0 | transformers | [
"transformers",
"pytorch",
"roberta",
"feature-extraction",
"dataset:semeval2012",
"model-index",
"text-embeddings-inference",
"endpoints_compatible",
"region:us"
]
| feature-extraction | 2022-07-22T16:49:45Z | ---
datasets:
- semeval2012
model-index:
- name: relbert/roberta-large-semeval2012-average-no-mask-prompt-c-triplet
results:
- task:
name: Relation Mapping
type: sorting-task
dataset:
name: Relation Mapping
args: relbert/relation_mapping
type: relation-mapping
metrics:
- name: Accuracy
type: accuracy
value: 0.8449404761904762
- task:
name: Analogy Questions (SAT full)
type: multiple-choice-qa
dataset:
name: SAT full
args: relbert/analogy_questions
type: analogy-questions
metrics:
- name: Accuracy
type: accuracy
value: 0.5401069519
- task:
name: Analogy Questions (SAT)
type: multiple-choice-qa
dataset:
name: SAT
args: relbert/analogy_questions
type: analogy-questions
metrics:
- name: Accuracy
type: accuracy
value: 0.5400593472
- task:
name: Analogy Questions (BATS)
type: multiple-choice-qa
dataset:
name: BATS
args: relbert/analogy_questions
type: analogy-questions
metrics:
- name: Accuracy
type: accuracy
value: 0.7954419122
- task:
name: Analogy Questions (Google)
type: multiple-choice-qa
dataset:
name: Google
args: relbert/analogy_questions
type: analogy-questions
metrics:
- name: Accuracy
type: accuracy
value: 0.912
- task:
name: Analogy Questions (U2)
type: multiple-choice-qa
dataset:
name: U2
args: relbert/analogy_questions
type: analogy-questions
metrics:
- name: Accuracy
type: accuracy
value: 0.6228070175
- task:
name: Analogy Questions (U4)
type: multiple-choice-qa
dataset:
name: U4
args: relbert/analogy_questions
type: analogy-questions
metrics:
- name: Accuracy
type: accuracy
value: 0.625
- task:
name: Lexical Relation Classification (BLESS)
type: classification
dataset:
name: BLESS
args: relbert/lexical_relation_classification
type: relation-classification
metrics:
- name: F1
type: f1
value: 0.9285821907488323
- name: F1 (macro)
type: f1_macro
value: 0.9238353183237691
- task:
name: Lexical Relation Classification (CogALexV)
type: classification
dataset:
name: CogALexV
args: relbert/lexical_relation_classification
type: relation-classification
metrics:
- name: F1
type: f1
value: 0.8795774647887324
- name: F1 (macro)
type: f1_macro
value: 0.7380564609392272
- task:
name: Lexical Relation Classification (EVALution)
type: classification
dataset:
name: BLESS
args: relbert/lexical_relation_classification
type: relation-classification
metrics:
- name: F1
type: f1
value: 0.6776814734561214
- name: F1 (macro)
type: f1_macro
value: 0.6542229601159605
- task:
name: Lexical Relation Classification (K&H+N)
type: classification
dataset:
name: K&H+N
args: relbert/lexical_relation_classification
type: relation-classification
metrics:
- name: F1
type: f1
value: 0.9683522292550601
- name: F1 (macro)
type: f1_macro
value: 0.897601966442876
- task:
name: Lexical Relation Classification (ROOT09)
type: classification
dataset:
name: ROOT09
args: relbert/lexical_relation_classification
type: relation-classification
metrics:
- name: F1
type: f1
value: 0.9088060169225948
- name: F1 (macro)
type: f1_macro
value: 0.909285139662564
---
# relbert/roberta-large-semeval2012-average-no-mask-prompt-c-triplet
RelBERT fine-tuned from [roberta-large](https://huggingface.co/roberta-large) on
[semeval2012](https://huggingface.co/datasets/semeval2012).
Fine-tuning is done via [RelBERT](https://github.com/asahi417/relbert) library (see the repository for more detail).
It achieves the following results on the relation understanding tasks:
- Analogy Question ([dataset](https://huggingface.co/datasets/relbert/analogy_questions), [full result](https://huggingface.co/relbert/roberta-large-semeval2012-average-no-mask-prompt-c-triplet/raw/main/analogy.json)):
- Accuracy on SAT (full): 0.5401069519
- Accuracy on SAT: 0.5400593472
- Accuracy on BATS: 0.7954419122
- Accuracy on U2: 0.6228070175
- Accuracy on U4: 0.625
- Accuracy on Google: 0.912
- Lexical Relation Classification ([dataset](https://huggingface.co/datasets/relbert/lexical_relation_classification), [full result](https://huggingface.co/relbert/roberta-large-semeval2012-average-no-mask-prompt-c-triplet/raw/main/classification.json)):
- Micro F1 score on BLESS: 0.9285821907488323
- Micro F1 score on CogALexV: 0.8795774647887324
- Micro F1 score on EVALution: 0.6776814734561214
- Micro F1 score on K&H+N: 0.9683522292550601
- Micro F1 score on ROOT09: 0.9088060169225948
- Relation Mapping ([dataset](https://huggingface.co/datasets/relbert/relation_mapping), [full result](https://huggingface.co/relbert/roberta-large-semeval2012-average-no-mask-prompt-c-triplet/raw/main/relation_mapping.json)):
- Accuracy on Relation Mapping: 0.8449404761904762
### Usage
This model can be used through the [relbert library](https://github.com/asahi417/relbert). Install the library via pip
```shell
pip install relbert
```
and activate model as below.
```python
from relbert import RelBERT
model = RelBERT("relbert/roberta-large-semeval2012-average-no-mask-prompt-c-triplet")
vector = model.get_embedding(['Tokyo', 'Japan']) # shape of (1024, )
```
### Training hyperparameters
The following hyperparameters were used during training:
- model: roberta-large
- max_length: 64
- mode: average_no_mask
- data: semeval2012
- n_sample: 10
- custom_template: Today, I finally discovered the relation between <subj> and <obj> : <mask>
- template: None
- softmax_loss: True
- in_batch_negative: True
- parent_contrast: True
- mse_margin: 1
- epoch: 1
- lr_warmup: 10
- batch: 64
- lr: 2e-05
- lr_decay: False
- weight_decay: 0
- optimizer: adam
- momentum: 0.9
- fp16: False
- random_seed: 0
The full configuration can be found at [fine-tuning parameter file](https://huggingface.co/relbert/roberta-large-semeval2012-average-no-mask-prompt-c-triplet/raw/main/trainer_config.json).
### Reference
If you use any resource from RelBERT, please consider to cite our [paper](https://aclanthology.org/2021.eacl-demos.7/).
```
@inproceedings{ushio-etal-2021-distilling-relation-embeddings,
title = "{D}istilling {R}elation {E}mbeddings from {P}re-trained {L}anguage {M}odels",
author = "Ushio, Asahi and
Schockaert, Steven and
Camacho-Collados, Jose",
booktitle = "EMNLP 2021",
year = "2021",
address = "Online",
publisher = "Association for Computational Linguistics",
}
```
|
clboetticher-school/xlm-roberta-base-finetuned-panx-de-fr | clboetticher-school | 2022-09-19T16:15:42Z | 134 | 0 | transformers | [
"transformers",
"pytorch",
"xlm-roberta",
"token-classification",
"generated_from_trainer",
"license:mit",
"autotrain_compatible",
"endpoints_compatible",
"region:us"
]
| token-classification | 2022-09-19T15:49:33Z | ---
license: mit
tags:
- generated_from_trainer
metrics:
- f1
model-index:
- name: xlm-roberta-base-finetuned-panx-de-fr
results: []
---
<!-- This model card has been generated automatically according to the information the Trainer had access to. You
should probably proofread and complete it, then remove this comment. -->
# xlm-roberta-base-finetuned-panx-de-fr
This model is a fine-tuned version of [xlm-roberta-base](https://huggingface.co/xlm-roberta-base) on the None dataset.
It achieves the following results on the evaluation set:
- Loss: 0.1608
- F1: 0.8593
## Model description
More information needed
## Intended uses & limitations
More information needed
## Training and evaluation data
More information needed
## Training procedure
### Training hyperparameters
The following hyperparameters were used during training:
- learning_rate: 5e-05
- train_batch_size: 24
- eval_batch_size: 24
- seed: 42
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
- lr_scheduler_type: linear
- num_epochs: 3
### Training results
| Training Loss | Epoch | Step | Validation Loss | F1 |
|:-------------:|:-----:|:----:|:---------------:|:------:|
| 0.2888 | 1.0 | 715 | 0.1779 | 0.8233 |
| 0.1437 | 2.0 | 1430 | 0.1570 | 0.8497 |
| 0.0931 | 3.0 | 2145 | 0.1608 | 0.8593 |
### Framework versions
- Transformers 4.11.3
- Pytorch 1.12.1+cu113
- Datasets 1.16.1
- Tokenizers 0.10.3
|
sd-concepts-library/singsing-doll | sd-concepts-library | 2022-09-19T16:14:12Z | 0 | 0 | null | [
"license:mit",
"region:us"
]
| null | 2022-09-19T16:14:06Z | ---
license: mit
---
### Singsing doll on Stable Diffusion
This is the `<singsing>` concept taught to Stable Diffusion via Textual Inversion. You can load this concept into the [Stable Conceptualizer](https://colab.research.google.com/github/huggingface/notebooks/blob/main/diffusers/stable_conceptualizer_inference.ipynb) notebook. You can also train your own concepts and load them into the concept libraries using [this notebook](https://colab.research.google.com/github/huggingface/notebooks/blob/main/diffusers/sd_textual_inversion_training.ipynb).
Here is the new concept you will be able to use as an `object`:
|
surya07/swin-tiny-patch4-window7-224-finetuned-eurosat | surya07 | 2022-09-19T16:11:19Z | 217 | 0 | transformers | [
"transformers",
"pytorch",
"tensorboard",
"swin",
"image-classification",
"generated_from_trainer",
"dataset:imagefolder",
"license:apache-2.0",
"model-index",
"autotrain_compatible",
"endpoints_compatible",
"region:us"
]
| image-classification | 2022-09-19T14:33:01Z | ---
license: apache-2.0
tags:
- generated_from_trainer
datasets:
- imagefolder
metrics:
- accuracy
model-index:
- name: swin-tiny-patch4-window7-224-finetuned-eurosat
results:
- task:
name: Image Classification
type: image-classification
dataset:
name: imagefolder
type: imagefolder
config: default
split: train
args: default
metrics:
- name: Accuracy
type: accuracy
value: 0.875
---
<!-- This model card has been generated automatically according to the information the Trainer had access to. You
should probably proofread and complete it, then remove this comment. -->
# swin-tiny-patch4-window7-224-finetuned-eurosat
This model is a fine-tuned version of [microsoft/swin-tiny-patch4-window7-224](https://huggingface.co/microsoft/swin-tiny-patch4-window7-224) on the imagefolder dataset.
It achieves the following results on the evaluation set:
- Loss: 0.4066
- Accuracy: 0.875
## Model description
More information needed
## Intended uses & limitations
More information needed
## Training and evaluation data
More information needed
## Training procedure
### Training hyperparameters
The following hyperparameters were used during training:
- learning_rate: 5e-05
- train_batch_size: 32
- eval_batch_size: 32
- seed: 42
- gradient_accumulation_steps: 4
- total_train_batch_size: 128
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
- lr_scheduler_type: linear
- lr_scheduler_warmup_ratio: 0.1
- num_epochs: 3
### Training results
| Training Loss | Epoch | Step | Validation Loss | Accuracy |
|:-------------:|:-----:|:----:|:---------------:|:--------:|
| No log | 0.57 | 1 | 0.7569 | 0.5417 |
| No log | 1.57 | 2 | 0.5000 | 0.8333 |
| No log | 2.57 | 3 | 0.4066 | 0.875 |
### Framework versions
- Transformers 4.22.1
- Pytorch 1.12.1+cu113
- Datasets 2.4.0
- Tokenizers 0.12.1
|
research-backup/roberta-large-semeval2012-average-prompt-e-nce | research-backup | 2022-09-19T16:02:37Z | 105 | 0 | transformers | [
"transformers",
"pytorch",
"roberta",
"feature-extraction",
"dataset:relbert/semeval2012_relational_similarity",
"model-index",
"text-embeddings-inference",
"endpoints_compatible",
"region:us"
]
| feature-extraction | 2022-07-24T20:44:09Z | ---
datasets:
- relbert/semeval2012_relational_similarity
model-index:
- name: relbert/roberta-large-semeval2012-average-prompt-e-nce
results:
- task:
name: Relation Mapping
type: sorting-task
dataset:
name: Relation Mapping
args: relbert/relation_mapping
type: relation-mapping
metrics:
- name: Accuracy
type: accuracy
value: 0.848452380952381
- task:
name: Analogy Questions (SAT full)
type: multiple-choice-qa
dataset:
name: SAT full
args: relbert/analogy_questions
type: analogy-questions
metrics:
- name: Accuracy
type: accuracy
value: 0.6016042780748663
- task:
name: Analogy Questions (SAT)
type: multiple-choice-qa
dataset:
name: SAT
args: relbert/analogy_questions
type: analogy-questions
metrics:
- name: Accuracy
type: accuracy
value: 0.6023738872403561
- task:
name: Analogy Questions (BATS)
type: multiple-choice-qa
dataset:
name: BATS
args: relbert/analogy_questions
type: analogy-questions
metrics:
- name: Accuracy
type: accuracy
value: 0.7476375764313508
- task:
name: Analogy Questions (Google)
type: multiple-choice-qa
dataset:
name: Google
args: relbert/analogy_questions
type: analogy-questions
metrics:
- name: Accuracy
type: accuracy
value: 0.86
- task:
name: Analogy Questions (U2)
type: multiple-choice-qa
dataset:
name: U2
args: relbert/analogy_questions
type: analogy-questions
metrics:
- name: Accuracy
type: accuracy
value: 0.5482456140350878
- task:
name: Analogy Questions (U4)
type: multiple-choice-qa
dataset:
name: U4
args: relbert/analogy_questions
type: analogy-questions
metrics:
- name: Accuracy
type: accuracy
value: 0.6111111111111112
- task:
name: Lexical Relation Classification (BLESS)
type: classification
dataset:
name: BLESS
args: relbert/lexical_relation_classification
type: relation-classification
metrics:
- name: F1
type: f1
value: 0.9193912912460449
- name: F1 (macro)
type: f1_macro
value: 0.9171163163754675
- task:
name: Lexical Relation Classification (CogALexV)
type: classification
dataset:
name: CogALexV
args: relbert/lexical_relation_classification
type: relation-classification
metrics:
- name: F1
type: f1
value: 0.8481220657276995
- name: F1 (macro)
type: f1_macro
value: 0.6734502135237685
- task:
name: Lexical Relation Classification (EVALution)
type: classification
dataset:
name: BLESS
args: relbert/lexical_relation_classification
type: relation-classification
metrics:
- name: F1
type: f1
value: 0.685807150595883
- name: F1 (macro)
type: f1_macro
value: 0.679750083279063
- task:
name: Lexical Relation Classification (K&H+N)
type: classification
dataset:
name: K&H+N
args: relbert/lexical_relation_classification
type: relation-classification
metrics:
- name: F1
type: f1
value: 0.962092230646171
- name: F1 (macro)
type: f1_macro
value: 0.8868721386428041
- task:
name: Lexical Relation Classification (ROOT09)
type: classification
dataset:
name: ROOT09
args: relbert/lexical_relation_classification
type: relation-classification
metrics:
- name: F1
type: f1
value: 0.898464431212786
- name: F1 (macro)
type: f1_macro
value: 0.8953388906170653
---
# relbert/roberta-large-semeval2012-average-prompt-e-nce
RelBERT fine-tuned from [roberta-large](https://huggingface.co/roberta-large) on
[relbert/semeval2012_relational_similarity](https://huggingface.co/datasets/relbert/semeval2012_relational_similarity).
Fine-tuning is done via [RelBERT](https://github.com/asahi417/relbert) library (see the repository for more detail).
It achieves the following results on the relation understanding tasks:
- Analogy Question ([dataset](https://huggingface.co/datasets/relbert/analogy_questions), [full result](https://huggingface.co/relbert/roberta-large-semeval2012-average-prompt-e-nce/raw/main/analogy.json)):
- Accuracy on SAT (full): 0.6016042780748663
- Accuracy on SAT: 0.6023738872403561
- Accuracy on BATS: 0.7476375764313508
- Accuracy on U2: 0.5482456140350878
- Accuracy on U4: 0.6111111111111112
- Accuracy on Google: 0.86
- Lexical Relation Classification ([dataset](https://huggingface.co/datasets/relbert/lexical_relation_classification), [full result](https://huggingface.co/relbert/roberta-large-semeval2012-average-prompt-e-nce/raw/main/classification.json)):
- Micro F1 score on BLESS: 0.9193912912460449
- Micro F1 score on CogALexV: 0.8481220657276995
- Micro F1 score on EVALution: 0.685807150595883
- Micro F1 score on K&H+N: 0.962092230646171
- Micro F1 score on ROOT09: 0.898464431212786
- Relation Mapping ([dataset](https://huggingface.co/datasets/relbert/relation_mapping), [full result](https://huggingface.co/relbert/roberta-large-semeval2012-average-prompt-e-nce/raw/main/relation_mapping.json)):
- Accuracy on Relation Mapping: 0.848452380952381
### Usage
This model can be used through the [relbert library](https://github.com/asahi417/relbert). Install the library via pip
```shell
pip install relbert
```
and activate model as below.
```python
from relbert import RelBERT
model = RelBERT("relbert/roberta-large-semeval2012-average-prompt-e-nce")
vector = model.get_embedding(['Tokyo', 'Japan']) # shape of (1024, )
```
### Training hyperparameters
The following hyperparameters were used during training:
- model: roberta-large
- max_length: 64
- mode: average
- data: relbert/semeval2012_relational_similarity
- template_mode: manual
- template: I wasn’t aware of this relationship, but I just read in the encyclopedia that <obj> is <subj>’s <mask>
- loss_function: nce_logout
- temperature_nce_constant: 0.05
- temperature_nce_rank: {'min': 0.01, 'max': 0.05, 'type': 'linear'}
- epoch: 21
- batch: 128
- lr: 5e-06
- lr_decay: False
- lr_warmup: 1
- weight_decay: 0
- random_seed: 0
- exclude_relation: None
- n_sample: 640
- gradient_accumulation: 8
The full configuration can be found at [fine-tuning parameter file](https://huggingface.co/relbert/roberta-large-semeval2012-average-prompt-e-nce/raw/main/trainer_config.json).
### Reference
If you use any resource from RelBERT, please consider to cite our [paper](https://aclanthology.org/2021.eacl-demos.7/).
```
@inproceedings{ushio-etal-2021-distilling-relation-embeddings,
title = "{D}istilling {R}elation {E}mbeddings from {P}re-trained {L}anguage {M}odels",
author = "Ushio, Asahi and
Schockaert, Steven and
Camacho-Collados, Jose",
booktitle = "EMNLP 2021",
year = "2021",
address = "Online",
publisher = "Association for Computational Linguistics",
}
```
|
research-backup/roberta-large-semeval2012-average-no-mask-prompt-b-nce | research-backup | 2022-09-19T15:54:08Z | 112 | 0 | transformers | [
"transformers",
"pytorch",
"roberta",
"feature-extraction",
"dataset:relbert/semeval2012_relational_similarity",
"model-index",
"text-embeddings-inference",
"endpoints_compatible",
"region:us"
]
| feature-extraction | 2022-07-22T10:59:39Z | ---
datasets:
- relbert/semeval2012_relational_similarity
model-index:
- name: relbert/roberta-large-semeval2012-average-no-mask-prompt-b-nce
results:
- task:
name: Relation Mapping
type: sorting-task
dataset:
name: Relation Mapping
args: relbert/relation_mapping
type: relation-mapping
metrics:
- name: Accuracy
type: accuracy
value: 0.8173412698412699
- task:
name: Analogy Questions (SAT full)
type: multiple-choice-qa
dataset:
name: SAT full
args: relbert/analogy_questions
type: analogy-questions
metrics:
- name: Accuracy
type: accuracy
value: 0.6122994652406417
- task:
name: Analogy Questions (SAT)
type: multiple-choice-qa
dataset:
name: SAT
args: relbert/analogy_questions
type: analogy-questions
metrics:
- name: Accuracy
type: accuracy
value: 0.6142433234421365
- task:
name: Analogy Questions (BATS)
type: multiple-choice-qa
dataset:
name: BATS
args: relbert/analogy_questions
type: analogy-questions
metrics:
- name: Accuracy
type: accuracy
value: 0.7865480822679266
- task:
name: Analogy Questions (Google)
type: multiple-choice-qa
dataset:
name: Google
args: relbert/analogy_questions
type: analogy-questions
metrics:
- name: Accuracy
type: accuracy
value: 0.93
- task:
name: Analogy Questions (U2)
type: multiple-choice-qa
dataset:
name: U2
args: relbert/analogy_questions
type: analogy-questions
metrics:
- name: Accuracy
type: accuracy
value: 0.5394736842105263
- task:
name: Analogy Questions (U4)
type: multiple-choice-qa
dataset:
name: U4
args: relbert/analogy_questions
type: analogy-questions
metrics:
- name: Accuracy
type: accuracy
value: 0.6018518518518519
- task:
name: Lexical Relation Classification (BLESS)
type: classification
dataset:
name: BLESS
args: relbert/lexical_relation_classification
type: relation-classification
metrics:
- name: F1
type: f1
value: 0.9174325749585657
- name: F1 (macro)
type: f1_macro
value: 0.9108478749677724
- task:
name: Lexical Relation Classification (CogALexV)
type: classification
dataset:
name: CogALexV
args: relbert/lexical_relation_classification
type: relation-classification
metrics:
- name: F1
type: f1
value: 0.855868544600939
- name: F1 (macro)
type: f1_macro
value: 0.6923047005195835
- task:
name: Lexical Relation Classification (EVALution)
type: classification
dataset:
name: BLESS
args: relbert/lexical_relation_classification
type: relation-classification
metrics:
- name: F1
type: f1
value: 0.6836403033586133
- name: F1 (macro)
type: f1_macro
value: 0.667310500013795
- task:
name: Lexical Relation Classification (K&H+N)
type: classification
dataset:
name: K&H+N
args: relbert/lexical_relation_classification
type: relation-classification
metrics:
- name: F1
type: f1
value: 0.9517284551714544
- name: F1 (macro)
type: f1_macro
value: 0.8530904199464412
- task:
name: Lexical Relation Classification (ROOT09)
type: classification
dataset:
name: ROOT09
args: relbert/lexical_relation_classification
type: relation-classification
metrics:
- name: F1
type: f1
value: 0.9019116264493889
- name: F1 (macro)
type: f1_macro
value: 0.8996556790705655
---
# relbert/roberta-large-semeval2012-average-no-mask-prompt-b-nce
RelBERT fine-tuned from [roberta-large](https://huggingface.co/roberta-large) on
[relbert/semeval2012_relational_similarity](https://huggingface.co/datasets/relbert/semeval2012_relational_similarity).
Fine-tuning is done via [RelBERT](https://github.com/asahi417/relbert) library (see the repository for more detail).
It achieves the following results on the relation understanding tasks:
- Analogy Question ([dataset](https://huggingface.co/datasets/relbert/analogy_questions), [full result](https://huggingface.co/relbert/roberta-large-semeval2012-average-no-mask-prompt-b-nce/raw/main/analogy.json)):
- Accuracy on SAT (full): 0.6122994652406417
- Accuracy on SAT: 0.6142433234421365
- Accuracy on BATS: 0.7865480822679266
- Accuracy on U2: 0.5394736842105263
- Accuracy on U4: 0.6018518518518519
- Accuracy on Google: 0.93
- Lexical Relation Classification ([dataset](https://huggingface.co/datasets/relbert/lexical_relation_classification), [full result](https://huggingface.co/relbert/roberta-large-semeval2012-average-no-mask-prompt-b-nce/raw/main/classification.json)):
- Micro F1 score on BLESS: 0.9174325749585657
- Micro F1 score on CogALexV: 0.855868544600939
- Micro F1 score on EVALution: 0.6836403033586133
- Micro F1 score on K&H+N: 0.9517284551714544
- Micro F1 score on ROOT09: 0.9019116264493889
- Relation Mapping ([dataset](https://huggingface.co/datasets/relbert/relation_mapping), [full result](https://huggingface.co/relbert/roberta-large-semeval2012-average-no-mask-prompt-b-nce/raw/main/relation_mapping.json)):
- Accuracy on Relation Mapping: 0.8173412698412699
### Usage
This model can be used through the [relbert library](https://github.com/asahi417/relbert). Install the library via pip
```shell
pip install relbert
```
and activate model as below.
```python
from relbert import RelBERT
model = RelBERT("relbert/roberta-large-semeval2012-average-no-mask-prompt-b-nce")
vector = model.get_embedding(['Tokyo', 'Japan']) # shape of (1024, )
```
### Training hyperparameters
The following hyperparameters were used during training:
- model: roberta-large
- max_length: 64
- mode: average_no_mask
- data: relbert/semeval2012_relational_similarity
- template_mode: manual
- template: Today, I finally discovered the relation between <subj> and <obj> : <obj> is <subj>'s <mask>
- loss_function: nce_logout
- temperature_nce_constant: 0.05
- temperature_nce_rank: {'min': 0.01, 'max': 0.05, 'type': 'linear'}
- epoch: 29
- batch: 128
- lr: 5e-06
- lr_decay: False
- lr_warmup: 1
- weight_decay: 0
- random_seed: 0
- exclude_relation: None
- n_sample: 640
- gradient_accumulation: 8
The full configuration can be found at [fine-tuning parameter file](https://huggingface.co/relbert/roberta-large-semeval2012-average-no-mask-prompt-b-nce/raw/main/trainer_config.json).
### Reference
If you use any resource from RelBERT, please consider to cite our [paper](https://aclanthology.org/2021.eacl-demos.7/).
```
@inproceedings{ushio-etal-2021-distilling-relation-embeddings,
title = "{D}istilling {R}elation {E}mbeddings from {P}re-trained {L}anguage {M}odels",
author = "Ushio, Asahi and
Schockaert, Steven and
Camacho-Collados, Jose",
booktitle = "EMNLP 2021",
year = "2021",
address = "Online",
publisher = "Association for Computational Linguistics",
}
```
|
research-backup/roberta-large-semeval2012-average-no-mask-prompt-a-nce | research-backup | 2022-09-19T15:49:51Z | 106 | 0 | transformers | [
"transformers",
"pytorch",
"roberta",
"feature-extraction",
"dataset:relbert/semeval2012_relational_similarity",
"model-index",
"text-embeddings-inference",
"endpoints_compatible",
"region:us"
]
| feature-extraction | 2022-07-22T10:57:35Z | ---
datasets:
- relbert/semeval2012_relational_similarity
model-index:
- name: relbert/roberta-large-semeval2012-average-no-mask-prompt-a-nce
results:
- task:
name: Relation Mapping
type: sorting-task
dataset:
name: Relation Mapping
args: relbert/relation_mapping
type: relation-mapping
metrics:
- name: Accuracy
type: accuracy
value: 0.866547619047619
- task:
name: Analogy Questions (SAT full)
type: multiple-choice-qa
dataset:
name: SAT full
args: relbert/analogy_questions
type: analogy-questions
metrics:
- name: Accuracy
type: accuracy
value: 0.7112299465240641
- task:
name: Analogy Questions (SAT)
type: multiple-choice-qa
dataset:
name: SAT
args: relbert/analogy_questions
type: analogy-questions
metrics:
- name: Accuracy
type: accuracy
value: 0.7062314540059347
- task:
name: Analogy Questions (BATS)
type: multiple-choice-qa
dataset:
name: BATS
args: relbert/analogy_questions
type: analogy-questions
metrics:
- name: Accuracy
type: accuracy
value: 0.782657031684269
- task:
name: Analogy Questions (Google)
type: multiple-choice-qa
dataset:
name: Google
args: relbert/analogy_questions
type: analogy-questions
metrics:
- name: Accuracy
type: accuracy
value: 0.936
- task:
name: Analogy Questions (U2)
type: multiple-choice-qa
dataset:
name: U2
args: relbert/analogy_questions
type: analogy-questions
metrics:
- name: Accuracy
type: accuracy
value: 0.6754385964912281
- task:
name: Analogy Questions (U4)
type: multiple-choice-qa
dataset:
name: U4
args: relbert/analogy_questions
type: analogy-questions
metrics:
- name: Accuracy
type: accuracy
value: 0.6921296296296297
- task:
name: Lexical Relation Classification (BLESS)
type: classification
dataset:
name: BLESS
args: relbert/lexical_relation_classification
type: relation-classification
metrics:
- name: F1
type: f1
value: 0.9124604489980412
- name: F1 (macro)
type: f1_macro
value: 0.9071904229357174
- task:
name: Lexical Relation Classification (CogALexV)
type: classification
dataset:
name: CogALexV
args: relbert/lexical_relation_classification
type: relation-classification
metrics:
- name: F1
type: f1
value: 0.8607981220657277
- name: F1 (macro)
type: f1_macro
value: 0.7021043673336924
- task:
name: Lexical Relation Classification (EVALution)
type: classification
dataset:
name: BLESS
args: relbert/lexical_relation_classification
type: relation-classification
metrics:
- name: F1
type: f1
value: 0.6863488624052004
- name: F1 (macro)
type: f1_macro
value: 0.6714181204599561
- task:
name: Lexical Relation Classification (K&H+N)
type: classification
dataset:
name: K&H+N
args: relbert/lexical_relation_classification
type: relation-classification
metrics:
- name: F1
type: f1
value: 0.9499895666689852
- name: F1 (macro)
type: f1_macro
value: 0.8482944164556818
- task:
name: Lexical Relation Classification (ROOT09)
type: classification
dataset:
name: ROOT09
args: relbert/lexical_relation_classification
type: relation-classification
metrics:
- name: F1
type: f1
value: 0.9075524913820119
- name: F1 (macro)
type: f1_macro
value: 0.9080337875282686
---
# relbert/roberta-large-semeval2012-average-no-mask-prompt-a-nce
RelBERT fine-tuned from [roberta-large](https://huggingface.co/roberta-large) on
[relbert/semeval2012_relational_similarity](https://huggingface.co/datasets/relbert/semeval2012_relational_similarity).
Fine-tuning is done via [RelBERT](https://github.com/asahi417/relbert) library (see the repository for more detail).
It achieves the following results on the relation understanding tasks:
- Analogy Question ([dataset](https://huggingface.co/datasets/relbert/analogy_questions), [full result](https://huggingface.co/relbert/roberta-large-semeval2012-average-no-mask-prompt-a-nce/raw/main/analogy.json)):
- Accuracy on SAT (full): 0.7112299465240641
- Accuracy on SAT: 0.7062314540059347
- Accuracy on BATS: 0.782657031684269
- Accuracy on U2: 0.6754385964912281
- Accuracy on U4: 0.6921296296296297
- Accuracy on Google: 0.936
- Lexical Relation Classification ([dataset](https://huggingface.co/datasets/relbert/lexical_relation_classification), [full result](https://huggingface.co/relbert/roberta-large-semeval2012-average-no-mask-prompt-a-nce/raw/main/classification.json)):
- Micro F1 score on BLESS: 0.9124604489980412
- Micro F1 score on CogALexV: 0.8607981220657277
- Micro F1 score on EVALution: 0.6863488624052004
- Micro F1 score on K&H+N: 0.9499895666689852
- Micro F1 score on ROOT09: 0.9075524913820119
- Relation Mapping ([dataset](https://huggingface.co/datasets/relbert/relation_mapping), [full result](https://huggingface.co/relbert/roberta-large-semeval2012-average-no-mask-prompt-a-nce/raw/main/relation_mapping.json)):
- Accuracy on Relation Mapping: 0.866547619047619
### Usage
This model can be used through the [relbert library](https://github.com/asahi417/relbert). Install the library via pip
```shell
pip install relbert
```
and activate model as below.
```python
from relbert import RelBERT
model = RelBERT("relbert/roberta-large-semeval2012-average-no-mask-prompt-a-nce")
vector = model.get_embedding(['Tokyo', 'Japan']) # shape of (1024, )
```
### Training hyperparameters
The following hyperparameters were used during training:
- model: roberta-large
- max_length: 64
- mode: average_no_mask
- data: relbert/semeval2012_relational_similarity
- template_mode: manual
- template: Today, I finally discovered the relation between <subj> and <obj> : <subj> is the <mask> of <obj>
- loss_function: nce_logout
- temperature_nce_constant: 0.05
- temperature_nce_rank: {'min': 0.01, 'max': 0.05, 'type': 'linear'}
- epoch: 29
- batch: 128
- lr: 5e-06
- lr_decay: False
- lr_warmup: 1
- weight_decay: 0
- random_seed: 0
- exclude_relation: None
- n_sample: 640
- gradient_accumulation: 8
The full configuration can be found at [fine-tuning parameter file](https://huggingface.co/relbert/roberta-large-semeval2012-average-no-mask-prompt-a-nce/raw/main/trainer_config.json).
### Reference
If you use any resource from RelBERT, please consider to cite our [paper](https://aclanthology.org/2021.eacl-demos.7/).
```
@inproceedings{ushio-etal-2021-distilling-relation-embeddings,
title = "{D}istilling {R}elation {E}mbeddings from {P}re-trained {L}anguage {M}odels",
author = "Ushio, Asahi and
Schockaert, Steven and
Camacho-Collados, Jose",
booktitle = "EMNLP 2021",
year = "2021",
address = "Online",
publisher = "Association for Computational Linguistics",
}
```
|
research-backup/roberta-large-semeval2012-average-no-mask-prompt-d-nce | research-backup | 2022-09-19T15:45:33Z | 114 | 0 | transformers | [
"transformers",
"pytorch",
"roberta",
"feature-extraction",
"dataset:relbert/semeval2012_relational_similarity",
"model-index",
"text-embeddings-inference",
"endpoints_compatible",
"region:us"
]
| feature-extraction | 2022-07-22T10:55:31Z | ---
datasets:
- relbert/semeval2012_relational_similarity
model-index:
- name: relbert/roberta-large-semeval2012-average-no-mask-prompt-d-nce
results:
- task:
name: Relation Mapping
type: sorting-task
dataset:
name: Relation Mapping
args: relbert/relation_mapping
type: relation-mapping
metrics:
- name: Accuracy
type: accuracy
value: 0.8909722222222223
- task:
name: Analogy Questions (SAT full)
type: multiple-choice-qa
dataset:
name: SAT full
args: relbert/analogy_questions
type: analogy-questions
metrics:
- name: Accuracy
type: accuracy
value: 0.6925133689839572
- task:
name: Analogy Questions (SAT)
type: multiple-choice-qa
dataset:
name: SAT
args: relbert/analogy_questions
type: analogy-questions
metrics:
- name: Accuracy
type: accuracy
value: 0.6913946587537092
- task:
name: Analogy Questions (BATS)
type: multiple-choice-qa
dataset:
name: BATS
args: relbert/analogy_questions
type: analogy-questions
metrics:
- name: Accuracy
type: accuracy
value: 0.8037798777098388
- task:
name: Analogy Questions (Google)
type: multiple-choice-qa
dataset:
name: Google
args: relbert/analogy_questions
type: analogy-questions
metrics:
- name: Accuracy
type: accuracy
value: 0.968
- task:
name: Analogy Questions (U2)
type: multiple-choice-qa
dataset:
name: U2
args: relbert/analogy_questions
type: analogy-questions
metrics:
- name: Accuracy
type: accuracy
value: 0.6885964912280702
- task:
name: Analogy Questions (U4)
type: multiple-choice-qa
dataset:
name: U4
args: relbert/analogy_questions
type: analogy-questions
metrics:
- name: Accuracy
type: accuracy
value: 0.6898148148148148
- task:
name: Lexical Relation Classification (BLESS)
type: classification
dataset:
name: BLESS
args: relbert/lexical_relation_classification
type: relation-classification
metrics:
- name: F1
type: f1
value: 0.9273768268796143
- name: F1 (macro)
type: f1_macro
value: 0.9211786019752478
- task:
name: Lexical Relation Classification (CogALexV)
type: classification
dataset:
name: CogALexV
args: relbert/lexical_relation_classification
type: relation-classification
metrics:
- name: F1
type: f1
value: 0.8615023474178404
- name: F1 (macro)
type: f1_macro
value: 0.7077498583524542
- task:
name: Lexical Relation Classification (EVALution)
type: classification
dataset:
name: BLESS
args: relbert/lexical_relation_classification
type: relation-classification
metrics:
- name: F1
type: f1
value: 0.6917659804983749
- name: F1 (macro)
type: f1_macro
value: 0.6746361055952557
- task:
name: Lexical Relation Classification (K&H+N)
type: classification
dataset:
name: K&H+N
args: relbert/lexical_relation_classification
type: relation-classification
metrics:
- name: F1
type: f1
value: 0.9573624539194547
- name: F1 (macro)
type: f1_macro
value: 0.8730312566461178
- task:
name: Lexical Relation Classification (ROOT09)
type: classification
dataset:
name: ROOT09
args: relbert/lexical_relation_classification
type: relation-classification
metrics:
- name: F1
type: f1
value: 0.9031651519899718
- name: F1 (macro)
type: f1_macro
value: 0.9025725245537483
---
# relbert/roberta-large-semeval2012-average-no-mask-prompt-d-nce
RelBERT fine-tuned from [roberta-large](https://huggingface.co/roberta-large) on
[relbert/semeval2012_relational_similarity](https://huggingface.co/datasets/relbert/semeval2012_relational_similarity).
Fine-tuning is done via [RelBERT](https://github.com/asahi417/relbert) library (see the repository for more detail).
It achieves the following results on the relation understanding tasks:
- Analogy Question ([dataset](https://huggingface.co/datasets/relbert/analogy_questions), [full result](https://huggingface.co/relbert/roberta-large-semeval2012-average-no-mask-prompt-d-nce/raw/main/analogy.json)):
- Accuracy on SAT (full): 0.6925133689839572
- Accuracy on SAT: 0.6913946587537092
- Accuracy on BATS: 0.8037798777098388
- Accuracy on U2: 0.6885964912280702
- Accuracy on U4: 0.6898148148148148
- Accuracy on Google: 0.968
- Lexical Relation Classification ([dataset](https://huggingface.co/datasets/relbert/lexical_relation_classification), [full result](https://huggingface.co/relbert/roberta-large-semeval2012-average-no-mask-prompt-d-nce/raw/main/classification.json)):
- Micro F1 score on BLESS: 0.9273768268796143
- Micro F1 score on CogALexV: 0.8615023474178404
- Micro F1 score on EVALution: 0.6917659804983749
- Micro F1 score on K&H+N: 0.9573624539194547
- Micro F1 score on ROOT09: 0.9031651519899718
- Relation Mapping ([dataset](https://huggingface.co/datasets/relbert/relation_mapping), [full result](https://huggingface.co/relbert/roberta-large-semeval2012-average-no-mask-prompt-d-nce/raw/main/relation_mapping.json)):
- Accuracy on Relation Mapping: 0.8909722222222223
### Usage
This model can be used through the [relbert library](https://github.com/asahi417/relbert). Install the library via pip
```shell
pip install relbert
```
and activate model as below.
```python
from relbert import RelBERT
model = RelBERT("relbert/roberta-large-semeval2012-average-no-mask-prompt-d-nce")
vector = model.get_embedding(['Tokyo', 'Japan']) # shape of (1024, )
```
### Training hyperparameters
The following hyperparameters were used during training:
- model: roberta-large
- max_length: 64
- mode: average_no_mask
- data: relbert/semeval2012_relational_similarity
- template_mode: manual
- template: I wasn’t aware of this relationship, but I just read in the encyclopedia that <subj> is the <mask> of <obj>
- loss_function: nce_logout
- temperature_nce_constant: 0.05
- temperature_nce_rank: {'min': 0.01, 'max': 0.05, 'type': 'linear'}
- epoch: 29
- batch: 128
- lr: 5e-06
- lr_decay: False
- lr_warmup: 1
- weight_decay: 0
- random_seed: 0
- exclude_relation: None
- n_sample: 640
- gradient_accumulation: 8
The full configuration can be found at [fine-tuning parameter file](https://huggingface.co/relbert/roberta-large-semeval2012-average-no-mask-prompt-d-nce/raw/main/trainer_config.json).
### Reference
If you use any resource from RelBERT, please consider to cite our [paper](https://aclanthology.org/2021.eacl-demos.7/).
```
@inproceedings{ushio-etal-2021-distilling-relation-embeddings,
title = "{D}istilling {R}elation {E}mbeddings from {P}re-trained {L}anguage {M}odels",
author = "Ushio, Asahi and
Schockaert, Steven and
Camacho-Collados, Jose",
booktitle = "EMNLP 2021",
year = "2021",
address = "Online",
publisher = "Association for Computational Linguistics",
}
```
|
research-backup/roberta-large-semeval2012-average-prompt-c-nce | research-backup | 2022-09-19T15:41:13Z | 106 | 0 | transformers | [
"transformers",
"pytorch",
"roberta",
"feature-extraction",
"dataset:relbert/semeval2012_relational_similarity",
"model-index",
"text-embeddings-inference",
"endpoints_compatible",
"region:us"
]
| feature-extraction | 2022-07-22T10:53:26Z | ---
datasets:
- relbert/semeval2012_relational_similarity
model-index:
- name: relbert/roberta-large-semeval2012-average-prompt-c-nce
results:
- task:
name: Relation Mapping
type: sorting-task
dataset:
name: Relation Mapping
args: relbert/relation_mapping
type: relation-mapping
metrics:
- name: Accuracy
type: accuracy
value: 0.866547619047619
- task:
name: Analogy Questions (SAT full)
type: multiple-choice-qa
dataset:
name: SAT full
args: relbert/analogy_questions
type: analogy-questions
metrics:
- name: Accuracy
type: accuracy
value: 0.6577540106951871
- task:
name: Analogy Questions (SAT)
type: multiple-choice-qa
dataset:
name: SAT
args: relbert/analogy_questions
type: analogy-questions
metrics:
- name: Accuracy
type: accuracy
value: 0.655786350148368
- task:
name: Analogy Questions (BATS)
type: multiple-choice-qa
dataset:
name: BATS
args: relbert/analogy_questions
type: analogy-questions
metrics:
- name: Accuracy
type: accuracy
value: 0.8132295719844358
- task:
name: Analogy Questions (Google)
type: multiple-choice-qa
dataset:
name: Google
args: relbert/analogy_questions
type: analogy-questions
metrics:
- name: Accuracy
type: accuracy
value: 0.956
- task:
name: Analogy Questions (U2)
type: multiple-choice-qa
dataset:
name: U2
args: relbert/analogy_questions
type: analogy-questions
metrics:
- name: Accuracy
type: accuracy
value: 0.618421052631579
- task:
name: Analogy Questions (U4)
type: multiple-choice-qa
dataset:
name: U4
args: relbert/analogy_questions
type: analogy-questions
metrics:
- name: Accuracy
type: accuracy
value: 0.6782407407407407
- task:
name: Lexical Relation Classification (BLESS)
type: classification
dataset:
name: BLESS
args: relbert/lexical_relation_classification
type: relation-classification
metrics:
- name: F1
type: f1
value: 0.9202953141479584
- name: F1 (macro)
type: f1_macro
value: 0.9141904094450748
- task:
name: Lexical Relation Classification (CogALexV)
type: classification
dataset:
name: CogALexV
args: relbert/lexical_relation_classification
type: relation-classification
metrics:
- name: F1
type: f1
value: 0.8737089201877934
- name: F1 (macro)
type: f1_macro
value: 0.7234367902362755
- task:
name: Lexical Relation Classification (EVALution)
type: classification
dataset:
name: BLESS
args: relbert/lexical_relation_classification
type: relation-classification
metrics:
- name: F1
type: f1
value: 0.6895991332611051
- name: F1 (macro)
type: f1_macro
value: 0.6754056888161498
- task:
name: Lexical Relation Classification (K&H+N)
type: classification
dataset:
name: K&H+N
args: relbert/lexical_relation_classification
type: relation-classification
metrics:
- name: F1
type: f1
value: 0.960283786603603
- name: F1 (macro)
type: f1_macro
value: 0.8772846038877066
- task:
name: Lexical Relation Classification (ROOT09)
type: classification
dataset:
name: ROOT09
args: relbert/lexical_relation_classification
type: relation-classification
metrics:
- name: F1
type: f1
value: 0.9204011281729865
- name: F1 (macro)
type: f1_macro
value: 0.9178047382633787
---
# relbert/roberta-large-semeval2012-average-prompt-c-nce
RelBERT fine-tuned from [roberta-large](https://huggingface.co/roberta-large) on
[relbert/semeval2012_relational_similarity](https://huggingface.co/datasets/relbert/semeval2012_relational_similarity).
Fine-tuning is done via [RelBERT](https://github.com/asahi417/relbert) library (see the repository for more detail).
It achieves the following results on the relation understanding tasks:
- Analogy Question ([dataset](https://huggingface.co/datasets/relbert/analogy_questions), [full result](https://huggingface.co/relbert/roberta-large-semeval2012-average-prompt-c-nce/raw/main/analogy.json)):
- Accuracy on SAT (full): 0.6577540106951871
- Accuracy on SAT: 0.655786350148368
- Accuracy on BATS: 0.8132295719844358
- Accuracy on U2: 0.618421052631579
- Accuracy on U4: 0.6782407407407407
- Accuracy on Google: 0.956
- Lexical Relation Classification ([dataset](https://huggingface.co/datasets/relbert/lexical_relation_classification), [full result](https://huggingface.co/relbert/roberta-large-semeval2012-average-prompt-c-nce/raw/main/classification.json)):
- Micro F1 score on BLESS: 0.9202953141479584
- Micro F1 score on CogALexV: 0.8737089201877934
- Micro F1 score on EVALution: 0.6895991332611051
- Micro F1 score on K&H+N: 0.960283786603603
- Micro F1 score on ROOT09: 0.9204011281729865
- Relation Mapping ([dataset](https://huggingface.co/datasets/relbert/relation_mapping), [full result](https://huggingface.co/relbert/roberta-large-semeval2012-average-prompt-c-nce/raw/main/relation_mapping.json)):
- Accuracy on Relation Mapping: 0.866547619047619
### Usage
This model can be used through the [relbert library](https://github.com/asahi417/relbert). Install the library via pip
```shell
pip install relbert
```
and activate model as below.
```python
from relbert import RelBERT
model = RelBERT("relbert/roberta-large-semeval2012-average-prompt-c-nce")
vector = model.get_embedding(['Tokyo', 'Japan']) # shape of (1024, )
```
### Training hyperparameters
The following hyperparameters were used during training:
- model: roberta-large
- max_length: 64
- mode: average
- data: relbert/semeval2012_relational_similarity
- template_mode: manual
- template: Today, I finally discovered the relation between <subj> and <obj> : <mask>
- loss_function: nce_logout
- temperature_nce_constant: 0.05
- temperature_nce_rank: {'min': 0.01, 'max': 0.05, 'type': 'linear'}
- epoch: 29
- batch: 128
- lr: 5e-06
- lr_decay: False
- lr_warmup: 1
- weight_decay: 0
- random_seed: 0
- exclude_relation: None
- n_sample: 640
- gradient_accumulation: 8
The full configuration can be found at [fine-tuning parameter file](https://huggingface.co/relbert/roberta-large-semeval2012-average-prompt-c-nce/raw/main/trainer_config.json).
### Reference
If you use any resource from RelBERT, please consider to cite our [paper](https://aclanthology.org/2021.eacl-demos.7/).
```
@inproceedings{ushio-etal-2021-distilling-relation-embeddings,
title = "{D}istilling {R}elation {E}mbeddings from {P}re-trained {L}anguage {M}odels",
author = "Ushio, Asahi and
Schockaert, Steven and
Camacho-Collados, Jose",
booktitle = "EMNLP 2021",
year = "2021",
address = "Online",
publisher = "Association for Computational Linguistics",
}
```
|
research-backup/roberta-large-semeval2012-average-prompt-b-nce | research-backup | 2022-09-19T15:36:58Z | 104 | 0 | transformers | [
"transformers",
"pytorch",
"roberta",
"feature-extraction",
"dataset:relbert/semeval2012_relational_similarity",
"model-index",
"text-embeddings-inference",
"endpoints_compatible",
"region:us"
]
| feature-extraction | 2022-07-22T10:51:01Z | ---
datasets:
- relbert/semeval2012_relational_similarity
model-index:
- name: relbert/roberta-large-semeval2012-average-prompt-b-nce
results:
- task:
name: Relation Mapping
type: sorting-task
dataset:
name: Relation Mapping
args: relbert/relation_mapping
type: relation-mapping
metrics:
- name: Accuracy
type: accuracy
value: 0.9023809523809524
- task:
name: Analogy Questions (SAT full)
type: multiple-choice-qa
dataset:
name: SAT full
args: relbert/analogy_questions
type: analogy-questions
metrics:
- name: Accuracy
type: accuracy
value: 0.6096256684491979
- task:
name: Analogy Questions (SAT)
type: multiple-choice-qa
dataset:
name: SAT
args: relbert/analogy_questions
type: analogy-questions
metrics:
- name: Accuracy
type: accuracy
value: 0.6083086053412463
- task:
name: Analogy Questions (BATS)
type: multiple-choice-qa
dataset:
name: BATS
args: relbert/analogy_questions
type: analogy-questions
metrics:
- name: Accuracy
type: accuracy
value: 0.7854363535297387
- task:
name: Analogy Questions (Google)
type: multiple-choice-qa
dataset:
name: Google
args: relbert/analogy_questions
type: analogy-questions
metrics:
- name: Accuracy
type: accuracy
value: 0.93
- task:
name: Analogy Questions (U2)
type: multiple-choice-qa
dataset:
name: U2
args: relbert/analogy_questions
type: analogy-questions
metrics:
- name: Accuracy
type: accuracy
value: 0.5833333333333334
- task:
name: Analogy Questions (U4)
type: multiple-choice-qa
dataset:
name: U4
args: relbert/analogy_questions
type: analogy-questions
metrics:
- name: Accuracy
type: accuracy
value: 0.5995370370370371
- task:
name: Lexical Relation Classification (BLESS)
type: classification
dataset:
name: BLESS
args: relbert/lexical_relation_classification
type: relation-classification
metrics:
- name: F1
type: f1
value: 0.9174325749585657
- name: F1 (macro)
type: f1_macro
value: 0.9129994415974204
- task:
name: Lexical Relation Classification (CogALexV)
type: classification
dataset:
name: CogALexV
args: relbert/lexical_relation_classification
type: relation-classification
metrics:
- name: F1
type: f1
value: 0.8603286384976526
- name: F1 (macro)
type: f1_macro
value: 0.698172861434558
- task:
name: Lexical Relation Classification (EVALution)
type: classification
dataset:
name: BLESS
args: relbert/lexical_relation_classification
type: relation-classification
metrics:
- name: F1
type: f1
value: 0.6679306608884074
- name: F1 (macro)
type: f1_macro
value: 0.6495733078766703
- task:
name: Lexical Relation Classification (K&H+N)
type: classification
dataset:
name: K&H+N
args: relbert/lexical_relation_classification
type: relation-classification
metrics:
- name: F1
type: f1
value: 0.9611184530847882
- name: F1 (macro)
type: f1_macro
value: 0.8867329071712199
- task:
name: Lexical Relation Classification (ROOT09)
type: classification
dataset:
name: ROOT09
args: relbert/lexical_relation_classification
type: relation-classification
metrics:
- name: F1
type: f1
value: 0.9012848636790974
- name: F1 (macro)
type: f1_macro
value: 0.9017314335034342
---
# relbert/roberta-large-semeval2012-average-prompt-b-nce
RelBERT fine-tuned from [roberta-large](https://huggingface.co/roberta-large) on
[relbert/semeval2012_relational_similarity](https://huggingface.co/datasets/relbert/semeval2012_relational_similarity).
Fine-tuning is done via [RelBERT](https://github.com/asahi417/relbert) library (see the repository for more detail).
It achieves the following results on the relation understanding tasks:
- Analogy Question ([dataset](https://huggingface.co/datasets/relbert/analogy_questions), [full result](https://huggingface.co/relbert/roberta-large-semeval2012-average-prompt-b-nce/raw/main/analogy.json)):
- Accuracy on SAT (full): 0.6096256684491979
- Accuracy on SAT: 0.6083086053412463
- Accuracy on BATS: 0.7854363535297387
- Accuracy on U2: 0.5833333333333334
- Accuracy on U4: 0.5995370370370371
- Accuracy on Google: 0.93
- Lexical Relation Classification ([dataset](https://huggingface.co/datasets/relbert/lexical_relation_classification), [full result](https://huggingface.co/relbert/roberta-large-semeval2012-average-prompt-b-nce/raw/main/classification.json)):
- Micro F1 score on BLESS: 0.9174325749585657
- Micro F1 score on CogALexV: 0.8603286384976526
- Micro F1 score on EVALution: 0.6679306608884074
- Micro F1 score on K&H+N: 0.9611184530847882
- Micro F1 score on ROOT09: 0.9012848636790974
- Relation Mapping ([dataset](https://huggingface.co/datasets/relbert/relation_mapping), [full result](https://huggingface.co/relbert/roberta-large-semeval2012-average-prompt-b-nce/raw/main/relation_mapping.json)):
- Accuracy on Relation Mapping: 0.9023809523809524
### Usage
This model can be used through the [relbert library](https://github.com/asahi417/relbert). Install the library via pip
```shell
pip install relbert
```
and activate model as below.
```python
from relbert import RelBERT
model = RelBERT("relbert/roberta-large-semeval2012-average-prompt-b-nce")
vector = model.get_embedding(['Tokyo', 'Japan']) # shape of (1024, )
```
### Training hyperparameters
The following hyperparameters were used during training:
- model: roberta-large
- max_length: 64
- mode: average
- data: relbert/semeval2012_relational_similarity
- template_mode: manual
- template: Today, I finally discovered the relation between <subj> and <obj> : <obj> is <subj>'s <mask>
- loss_function: nce_logout
- temperature_nce_constant: 0.05
- temperature_nce_rank: {'min': 0.01, 'max': 0.05, 'type': 'linear'}
- epoch: 23
- batch: 128
- lr: 5e-06
- lr_decay: False
- lr_warmup: 1
- weight_decay: 0
- random_seed: 0
- exclude_relation: None
- n_sample: 640
- gradient_accumulation: 8
The full configuration can be found at [fine-tuning parameter file](https://huggingface.co/relbert/roberta-large-semeval2012-average-prompt-b-nce/raw/main/trainer_config.json).
### Reference
If you use any resource from RelBERT, please consider to cite our [paper](https://aclanthology.org/2021.eacl-demos.7/).
```
@inproceedings{ushio-etal-2021-distilling-relation-embeddings,
title = "{D}istilling {R}elation {E}mbeddings from {P}re-trained {L}anguage {M}odels",
author = "Ushio, Asahi and
Schockaert, Steven and
Camacho-Collados, Jose",
booktitle = "EMNLP 2021",
year = "2021",
address = "Online",
publisher = "Association for Computational Linguistics",
}
```
|
research-backup/roberta-large-semeval2012-average-prompt-a-nce | research-backup | 2022-09-19T15:32:33Z | 107 | 0 | transformers | [
"transformers",
"pytorch",
"roberta",
"feature-extraction",
"dataset:relbert/semeval2012_relational_similarity",
"model-index",
"text-embeddings-inference",
"endpoints_compatible",
"region:us"
]
| feature-extraction | 2022-07-22T10:47:57Z | ---
datasets:
- relbert/semeval2012_relational_similarity
model-index:
- name: relbert/roberta-large-semeval2012-average-prompt-a-nce
results:
- task:
name: Relation Mapping
type: sorting-task
dataset:
name: Relation Mapping
args: relbert/relation_mapping
type: relation-mapping
metrics:
- name: Accuracy
type: accuracy
value: 0.8719047619047618
- task:
name: Analogy Questions (SAT full)
type: multiple-choice-qa
dataset:
name: SAT full
args: relbert/analogy_questions
type: analogy-questions
metrics:
- name: Accuracy
type: accuracy
value: 0.6925133689839572
- task:
name: Analogy Questions (SAT)
type: multiple-choice-qa
dataset:
name: SAT
args: relbert/analogy_questions
type: analogy-questions
metrics:
- name: Accuracy
type: accuracy
value: 0.6913946587537092
- task:
name: Analogy Questions (BATS)
type: multiple-choice-qa
dataset:
name: BATS
args: relbert/analogy_questions
type: analogy-questions
metrics:
- name: Accuracy
type: accuracy
value: 0.7776542523624236
- task:
name: Analogy Questions (Google)
type: multiple-choice-qa
dataset:
name: Google
args: relbert/analogy_questions
type: analogy-questions
metrics:
- name: Accuracy
type: accuracy
value: 0.936
- task:
name: Analogy Questions (U2)
type: multiple-choice-qa
dataset:
name: U2
args: relbert/analogy_questions
type: analogy-questions
metrics:
- name: Accuracy
type: accuracy
value: 0.6535087719298246
- task:
name: Analogy Questions (U4)
type: multiple-choice-qa
dataset:
name: U4
args: relbert/analogy_questions
type: analogy-questions
metrics:
- name: Accuracy
type: accuracy
value: 0.6666666666666666
- task:
name: Lexical Relation Classification (BLESS)
type: classification
dataset:
name: BLESS
args: relbert/lexical_relation_classification
type: relation-classification
metrics:
- name: F1
type: f1
value: 0.9153231881874341
- name: F1 (macro)
type: f1_macro
value: 0.9116885342042641
- task:
name: Lexical Relation Classification (CogALexV)
type: classification
dataset:
name: CogALexV
args: relbert/lexical_relation_classification
type: relation-classification
metrics:
- name: F1
type: f1
value: 0.8509389671361502
- name: F1 (macro)
type: f1_macro
value: 0.6788929400995221
- task:
name: Lexical Relation Classification (EVALution)
type: classification
dataset:
name: BLESS
args: relbert/lexical_relation_classification
type: relation-classification
metrics:
- name: F1
type: f1
value: 0.6771397616468039
- name: F1 (macro)
type: f1_macro
value: 0.6568153884216413
- task:
name: Lexical Relation Classification (K&H+N)
type: classification
dataset:
name: K&H+N
args: relbert/lexical_relation_classification
type: relation-classification
metrics:
- name: F1
type: f1
value: 0.9575015649996522
- name: F1 (macro)
type: f1_macro
value: 0.8657248723477102
- task:
name: Lexical Relation Classification (ROOT09)
type: classification
dataset:
name: ROOT09
args: relbert/lexical_relation_classification
type: relation-classification
metrics:
- name: F1
type: f1
value: 0.9025383892196804
- name: F1 (macro)
type: f1_macro
value: 0.899796204657101
---
# relbert/roberta-large-semeval2012-average-prompt-a-nce
RelBERT fine-tuned from [roberta-large](https://huggingface.co/roberta-large) on
[relbert/semeval2012_relational_similarity](https://huggingface.co/datasets/relbert/semeval2012_relational_similarity).
Fine-tuning is done via [RelBERT](https://github.com/asahi417/relbert) library (see the repository for more detail).
It achieves the following results on the relation understanding tasks:
- Analogy Question ([dataset](https://huggingface.co/datasets/relbert/analogy_questions), [full result](https://huggingface.co/relbert/roberta-large-semeval2012-average-prompt-a-nce/raw/main/analogy.json)):
- Accuracy on SAT (full): 0.6925133689839572
- Accuracy on SAT: 0.6913946587537092
- Accuracy on BATS: 0.7776542523624236
- Accuracy on U2: 0.6535087719298246
- Accuracy on U4: 0.6666666666666666
- Accuracy on Google: 0.936
- Lexical Relation Classification ([dataset](https://huggingface.co/datasets/relbert/lexical_relation_classification), [full result](https://huggingface.co/relbert/roberta-large-semeval2012-average-prompt-a-nce/raw/main/classification.json)):
- Micro F1 score on BLESS: 0.9153231881874341
- Micro F1 score on CogALexV: 0.8509389671361502
- Micro F1 score on EVALution: 0.6771397616468039
- Micro F1 score on K&H+N: 0.9575015649996522
- Micro F1 score on ROOT09: 0.9025383892196804
- Relation Mapping ([dataset](https://huggingface.co/datasets/relbert/relation_mapping), [full result](https://huggingface.co/relbert/roberta-large-semeval2012-average-prompt-a-nce/raw/main/relation_mapping.json)):
- Accuracy on Relation Mapping: 0.8719047619047618
### Usage
This model can be used through the [relbert library](https://github.com/asahi417/relbert). Install the library via pip
```shell
pip install relbert
```
and activate model as below.
```python
from relbert import RelBERT
model = RelBERT("relbert/roberta-large-semeval2012-average-prompt-a-nce")
vector = model.get_embedding(['Tokyo', 'Japan']) # shape of (1024, )
```
### Training hyperparameters
The following hyperparameters were used during training:
- model: roberta-large
- max_length: 64
- mode: average
- data: relbert/semeval2012_relational_similarity
- template_mode: manual
- template: Today, I finally discovered the relation between <subj> and <obj> : <subj> is the <mask> of <obj>
- loss_function: nce_logout
- temperature_nce_constant: 0.05
- temperature_nce_rank: {'min': 0.01, 'max': 0.05, 'type': 'linear'}
- epoch: 29
- batch: 128
- lr: 5e-06
- lr_decay: False
- lr_warmup: 1
- weight_decay: 0
- random_seed: 0
- exclude_relation: None
- n_sample: 640
- gradient_accumulation: 8
The full configuration can be found at [fine-tuning parameter file](https://huggingface.co/relbert/roberta-large-semeval2012-average-prompt-a-nce/raw/main/trainer_config.json).
### Reference
If you use any resource from RelBERT, please consider to cite our [paper](https://aclanthology.org/2021.eacl-demos.7/).
```
@inproceedings{ushio-etal-2021-distilling-relation-embeddings,
title = "{D}istilling {R}elation {E}mbeddings from {P}re-trained {L}anguage {M}odels",
author = "Ushio, Asahi and
Schockaert, Steven and
Camacho-Collados, Jose",
booktitle = "EMNLP 2021",
year = "2021",
address = "Online",
publisher = "Association for Computational Linguistics",
}
```
|
research-backup/roberta-large-semeval2012-mask-prompt-e-nce | research-backup | 2022-09-19T15:28:19Z | 109 | 0 | transformers | [
"transformers",
"pytorch",
"roberta",
"feature-extraction",
"dataset:relbert/semeval2012_relational_similarity",
"model-index",
"text-embeddings-inference",
"endpoints_compatible",
"region:us"
]
| feature-extraction | 2022-07-22T10:43:51Z | ---
datasets:
- relbert/semeval2012_relational_similarity
model-index:
- name: relbert/roberta-large-semeval2012-mask-prompt-e-nce
results:
- task:
name: Relation Mapping
type: sorting-task
dataset:
name: Relation Mapping
args: relbert/relation_mapping
type: relation-mapping
metrics:
- name: Accuracy
type: accuracy
value: 0.8230555555555555
- task:
name: Analogy Questions (SAT full)
type: multiple-choice-qa
dataset:
name: SAT full
args: relbert/analogy_questions
type: analogy-questions
metrics:
- name: Accuracy
type: accuracy
value: 0.6122994652406417
- task:
name: Analogy Questions (SAT)
type: multiple-choice-qa
dataset:
name: SAT
args: relbert/analogy_questions
type: analogy-questions
metrics:
- name: Accuracy
type: accuracy
value: 0.6172106824925816
- task:
name: Analogy Questions (BATS)
type: multiple-choice-qa
dataset:
name: BATS
args: relbert/analogy_questions
type: analogy-questions
metrics:
- name: Accuracy
type: accuracy
value: 0.7787659811006115
- task:
name: Analogy Questions (Google)
type: multiple-choice-qa
dataset:
name: Google
args: relbert/analogy_questions
type: analogy-questions
metrics:
- name: Accuracy
type: accuracy
value: 0.924
- task:
name: Analogy Questions (U2)
type: multiple-choice-qa
dataset:
name: U2
args: relbert/analogy_questions
type: analogy-questions
metrics:
- name: Accuracy
type: accuracy
value: 0.6096491228070176
- task:
name: Analogy Questions (U4)
type: multiple-choice-qa
dataset:
name: U4
args: relbert/analogy_questions
type: analogy-questions
metrics:
- name: Accuracy
type: accuracy
value: 0.6203703703703703
- task:
name: Lexical Relation Classification (BLESS)
type: classification
dataset:
name: BLESS
args: relbert/lexical_relation_classification
type: relation-classification
metrics:
- name: F1
type: f1
value: 0.9328009642910954
- name: F1 (macro)
type: f1_macro
value: 0.9275146998553098
- task:
name: Lexical Relation Classification (CogALexV)
type: classification
dataset:
name: CogALexV
args: relbert/lexical_relation_classification
type: relation-classification
metrics:
- name: F1
type: f1
value: 0.8734741784037559
- name: F1 (macro)
type: f1_macro
value: 0.7242617350118036
- task:
name: Lexical Relation Classification (EVALution)
type: classification
dataset:
name: BLESS
args: relbert/lexical_relation_classification
type: relation-classification
metrics:
- name: F1
type: f1
value: 0.7004333694474539
- name: F1 (macro)
type: f1_macro
value: 0.6937902608558703
- task:
name: Lexical Relation Classification (K&H+N)
type: classification
dataset:
name: K&H+N
args: relbert/lexical_relation_classification
type: relation-classification
metrics:
- name: F1
type: f1
value: 0.9617444529456771
- name: F1 (macro)
type: f1_macro
value: 0.8874628760316924
- task:
name: Lexical Relation Classification (ROOT09)
type: classification
dataset:
name: ROOT09
args: relbert/lexical_relation_classification
type: relation-classification
metrics:
- name: F1
type: f1
value: 0.9081792541523034
- name: F1 (macro)
type: f1_macro
value: 0.9066592810364282
---
# relbert/roberta-large-semeval2012-mask-prompt-e-nce
RelBERT fine-tuned from [roberta-large](https://huggingface.co/roberta-large) on
[relbert/semeval2012_relational_similarity](https://huggingface.co/datasets/relbert/semeval2012_relational_similarity).
Fine-tuning is done via [RelBERT](https://github.com/asahi417/relbert) library (see the repository for more detail).
It achieves the following results on the relation understanding tasks:
- Analogy Question ([dataset](https://huggingface.co/datasets/relbert/analogy_questions), [full result](https://huggingface.co/relbert/roberta-large-semeval2012-mask-prompt-e-nce/raw/main/analogy.json)):
- Accuracy on SAT (full): 0.6122994652406417
- Accuracy on SAT: 0.6172106824925816
- Accuracy on BATS: 0.7787659811006115
- Accuracy on U2: 0.6096491228070176
- Accuracy on U4: 0.6203703703703703
- Accuracy on Google: 0.924
- Lexical Relation Classification ([dataset](https://huggingface.co/datasets/relbert/lexical_relation_classification), [full result](https://huggingface.co/relbert/roberta-large-semeval2012-mask-prompt-e-nce/raw/main/classification.json)):
- Micro F1 score on BLESS: 0.9328009642910954
- Micro F1 score on CogALexV: 0.8734741784037559
- Micro F1 score on EVALution: 0.7004333694474539
- Micro F1 score on K&H+N: 0.9617444529456771
- Micro F1 score on ROOT09: 0.9081792541523034
- Relation Mapping ([dataset](https://huggingface.co/datasets/relbert/relation_mapping), [full result](https://huggingface.co/relbert/roberta-large-semeval2012-mask-prompt-e-nce/raw/main/relation_mapping.json)):
- Accuracy on Relation Mapping: 0.8230555555555555
### Usage
This model can be used through the [relbert library](https://github.com/asahi417/relbert). Install the library via pip
```shell
pip install relbert
```
and activate model as below.
```python
from relbert import RelBERT
model = RelBERT("relbert/roberta-large-semeval2012-mask-prompt-e-nce")
vector = model.get_embedding(['Tokyo', 'Japan']) # shape of (1024, )
```
### Training hyperparameters
The following hyperparameters were used during training:
- model: roberta-large
- max_length: 64
- mode: mask
- data: relbert/semeval2012_relational_similarity
- template_mode: manual
- template: I wasn’t aware of this relationship, but I just read in the encyclopedia that <obj> is <subj>’s <mask>
- loss_function: nce_logout
- temperature_nce_constant: 0.05
- temperature_nce_rank: {'min': 0.01, 'max': 0.05, 'type': 'linear'}
- epoch: 23
- batch: 128
- lr: 5e-06
- lr_decay: False
- lr_warmup: 1
- weight_decay: 0
- random_seed: 0
- exclude_relation: None
- n_sample: 640
- gradient_accumulation: 8
The full configuration can be found at [fine-tuning parameter file](https://huggingface.co/relbert/roberta-large-semeval2012-mask-prompt-e-nce/raw/main/trainer_config.json).
### Reference
If you use any resource from RelBERT, please consider to cite our [paper](https://aclanthology.org/2021.eacl-demos.7/).
```
@inproceedings{ushio-etal-2021-distilling-relation-embeddings,
title = "{D}istilling {R}elation {E}mbeddings from {P}re-trained {L}anguage {M}odels",
author = "Ushio, Asahi and
Schockaert, Steven and
Camacho-Collados, Jose",
booktitle = "EMNLP 2021",
year = "2021",
address = "Online",
publisher = "Association for Computational Linguistics",
}
```
|
sd-concepts-library/jos-de-kat | sd-concepts-library | 2022-09-19T15:13:36Z | 0 | 0 | null | [
"license:mit",
"region:us"
]
| null | 2022-09-19T15:13:30Z | ---
license: mit
---
### Jos de Kat on Stable Diffusion
This is the `<kat-jos>` concept taught to Stable Diffusion via Textual Inversion. You can load this concept into the [Stable Conceptualizer](https://colab.research.google.com/github/huggingface/notebooks/blob/main/diffusers/stable_conceptualizer_inference.ipynb) notebook. You can also train your own concepts and load them into the concept libraries using [this notebook](https://colab.research.google.com/github/huggingface/notebooks/blob/main/diffusers/sd_textual_inversion_training.ipynb).
Here is the new concept you will be able to use as an `object`:
|
research-backup/roberta-large-semeval2012-mask-prompt-a-nce | research-backup | 2022-09-19T15:07:54Z | 107 | 0 | transformers | [
"transformers",
"pytorch",
"roberta",
"feature-extraction",
"dataset:relbert/semeval2012_relational_similarity",
"model-index",
"text-embeddings-inference",
"endpoints_compatible",
"region:us"
]
| feature-extraction | 2022-07-22T10:37:07Z | ---
datasets:
- relbert/semeval2012_relational_similarity
model-index:
- name: relbert/roberta-large-semeval2012-mask-prompt-a-nce
results:
- task:
name: Relation Mapping
type: sorting-task
dataset:
name: Relation Mapping
args: relbert/relation_mapping
type: relation-mapping
metrics:
- name: Accuracy
type: accuracy
value: 0.8680952380952381
- task:
name: Analogy Questions (SAT full)
type: multiple-choice-qa
dataset:
name: SAT full
args: relbert/analogy_questions
type: analogy-questions
metrics:
- name: Accuracy
type: accuracy
value: 0.7112299465240641
- task:
name: Analogy Questions (SAT)
type: multiple-choice-qa
dataset:
name: SAT
args: relbert/analogy_questions
type: analogy-questions
metrics:
- name: Accuracy
type: accuracy
value: 0.7091988130563798
- task:
name: Analogy Questions (BATS)
type: multiple-choice-qa
dataset:
name: BATS
args: relbert/analogy_questions
type: analogy-questions
metrics:
- name: Accuracy
type: accuracy
value: 0.7537520844913841
- task:
name: Analogy Questions (Google)
type: multiple-choice-qa
dataset:
name: Google
args: relbert/analogy_questions
type: analogy-questions
metrics:
- name: Accuracy
type: accuracy
value: 0.95
- task:
name: Analogy Questions (U2)
type: multiple-choice-qa
dataset:
name: U2
args: relbert/analogy_questions
type: analogy-questions
metrics:
- name: Accuracy
type: accuracy
value: 0.6622807017543859
- task:
name: Analogy Questions (U4)
type: multiple-choice-qa
dataset:
name: U4
args: relbert/analogy_questions
type: analogy-questions
metrics:
- name: Accuracy
type: accuracy
value: 0.6666666666666666
- task:
name: Lexical Relation Classification (BLESS)
type: classification
dataset:
name: BLESS
args: relbert/lexical_relation_classification
type: relation-classification
metrics:
- name: F1
type: f1
value: 0.9245140876902215
- name: F1 (macro)
type: f1_macro
value: 0.9217193105874872
- task:
name: Lexical Relation Classification (CogALexV)
type: classification
dataset:
name: CogALexV
args: relbert/lexical_relation_classification
type: relation-classification
metrics:
- name: F1
type: f1
value: 0.8809859154929578
- name: F1 (macro)
type: f1_macro
value: 0.7387527642398365
- task:
name: Lexical Relation Classification (EVALution)
type: classification
dataset:
name: BLESS
args: relbert/lexical_relation_classification
type: relation-classification
metrics:
- name: F1
type: f1
value: 0.7134344528710725
- name: F1 (macro)
type: f1_macro
value: 0.6978567457746659
- task:
name: Lexical Relation Classification (K&H+N)
type: classification
dataset:
name: K&H+N
args: relbert/lexical_relation_classification
type: relation-classification
metrics:
- name: F1
type: f1
value: 0.955484454336788
- name: F1 (macro)
type: f1_macro
value: 0.8778253752250313
- task:
name: Lexical Relation Classification (ROOT09)
type: classification
dataset:
name: ROOT09
args: relbert/lexical_relation_classification
type: relation-classification
metrics:
- name: F1
type: f1
value: 0.909746161078032
- name: F1 (macro)
type: f1_macro
value: 0.9088078445136086
---
# relbert/roberta-large-semeval2012-mask-prompt-a-nce
RelBERT fine-tuned from [roberta-large](https://huggingface.co/roberta-large) on
[relbert/semeval2012_relational_similarity](https://huggingface.co/datasets/relbert/semeval2012_relational_similarity).
Fine-tuning is done via [RelBERT](https://github.com/asahi417/relbert) library (see the repository for more detail).
It achieves the following results on the relation understanding tasks:
- Analogy Question ([dataset](https://huggingface.co/datasets/relbert/analogy_questions), [full result](https://huggingface.co/relbert/roberta-large-semeval2012-mask-prompt-a-nce/raw/main/analogy.json)):
- Accuracy on SAT (full): 0.7112299465240641
- Accuracy on SAT: 0.7091988130563798
- Accuracy on BATS: 0.7537520844913841
- Accuracy on U2: 0.6622807017543859
- Accuracy on U4: 0.6666666666666666
- Accuracy on Google: 0.95
- Lexical Relation Classification ([dataset](https://huggingface.co/datasets/relbert/lexical_relation_classification), [full result](https://huggingface.co/relbert/roberta-large-semeval2012-mask-prompt-a-nce/raw/main/classification.json)):
- Micro F1 score on BLESS: 0.9245140876902215
- Micro F1 score on CogALexV: 0.8809859154929578
- Micro F1 score on EVALution: 0.7134344528710725
- Micro F1 score on K&H+N: 0.955484454336788
- Micro F1 score on ROOT09: 0.909746161078032
- Relation Mapping ([dataset](https://huggingface.co/datasets/relbert/relation_mapping), [full result](https://huggingface.co/relbert/roberta-large-semeval2012-mask-prompt-a-nce/raw/main/relation_mapping.json)):
- Accuracy on Relation Mapping: 0.8680952380952381
### Usage
This model can be used through the [relbert library](https://github.com/asahi417/relbert). Install the library via pip
```shell
pip install relbert
```
and activate model as below.
```python
from relbert import RelBERT
model = RelBERT("relbert/roberta-large-semeval2012-mask-prompt-a-nce")
vector = model.get_embedding(['Tokyo', 'Japan']) # shape of (1024, )
```
### Training hyperparameters
The following hyperparameters were used during training:
- model: roberta-large
- max_length: 64
- mode: mask
- data: relbert/semeval2012_relational_similarity
- template_mode: manual
- template: Today, I finally discovered the relation between <subj> and <obj> : <subj> is the <mask> of <obj>
- loss_function: nce_logout
- temperature_nce_constant: 0.05
- temperature_nce_rank: {'min': 0.01, 'max': 0.05, 'type': 'linear'}
- epoch: 23
- batch: 128
- lr: 5e-06
- lr_decay: False
- lr_warmup: 1
- weight_decay: 0
- random_seed: 0
- exclude_relation: None
- n_sample: 640
- gradient_accumulation: 8
The full configuration can be found at [fine-tuning parameter file](https://huggingface.co/relbert/roberta-large-semeval2012-mask-prompt-a-nce/raw/main/trainer_config.json).
### Reference
If you use any resource from RelBERT, please consider to cite our [paper](https://aclanthology.org/2021.eacl-demos.7/).
```
@inproceedings{ushio-etal-2021-distilling-relation-embeddings,
title = "{D}istilling {R}elation {E}mbeddings from {P}re-trained {L}anguage {M}odels",
author = "Ushio, Asahi and
Schockaert, Steven and
Camacho-Collados, Jose",
booktitle = "EMNLP 2021",
year = "2021",
address = "Online",
publisher = "Association for Computational Linguistics",
}
```
|
matemato/dqn-SpaceInvadersNoFrameskip-v4 | matemato | 2022-09-19T14:54:42Z | 2 | 0 | stable-baselines3 | [
"stable-baselines3",
"SpaceInvadersNoFrameskip-v4",
"deep-reinforcement-learning",
"reinforcement-learning",
"model-index",
"region:us"
]
| reinforcement-learning | 2022-09-19T14:54:06Z | ---
library_name: stable-baselines3
tags:
- SpaceInvadersNoFrameskip-v4
- deep-reinforcement-learning
- reinforcement-learning
- stable-baselines3
model-index:
- name: DQN
results:
- metrics:
- type: mean_reward
value: 512.00 +/- 131.55
name: mean_reward
task:
type: reinforcement-learning
name: reinforcement-learning
dataset:
name: SpaceInvadersNoFrameskip-v4
type: SpaceInvadersNoFrameskip-v4
---
# **DQN** Agent playing **SpaceInvadersNoFrameskip-v4**
This is a trained model of a **DQN** agent playing **SpaceInvadersNoFrameskip-v4**
using the [stable-baselines3 library](https://github.com/DLR-RM/stable-baselines3)
and the [RL Zoo](https://github.com/DLR-RM/rl-baselines3-zoo).
The RL Zoo is a training framework for Stable Baselines3
reinforcement learning agents,
with hyperparameter optimization and pre-trained agents included.
## Usage (with SB3 RL Zoo)
RL Zoo: https://github.com/DLR-RM/rl-baselines3-zoo<br/>
SB3: https://github.com/DLR-RM/stable-baselines3<br/>
SB3 Contrib: https://github.com/Stable-Baselines-Team/stable-baselines3-contrib
```
# Download model and save it into the logs/ folder
python -m utils.load_from_hub --algo dqn --env SpaceInvadersNoFrameskip-v4 -orga matemato -f logs/
python enjoy.py --algo dqn --env SpaceInvadersNoFrameskip-v4 -f logs/
```
## Training (with the RL Zoo)
```
python train.py --algo dqn --env SpaceInvadersNoFrameskip-v4 -f logs/
# Upload the model and generate video (when possible)
python -m utils.push_to_hub --algo dqn --env SpaceInvadersNoFrameskip-v4 -f logs/ -orga matemato
```
## Hyperparameters
```python
OrderedDict([('batch_size', 32),
('buffer_size', 100000),
('env_wrapper',
['stable_baselines3.common.atari_wrappers.AtariWrapper']),
('exploration_final_eps', 0.01),
('exploration_fraction', 0.1),
('frame_stack', 4),
('gradient_steps', 1),
('learning_rate', 0.0001),
('learning_starts', 100000),
('n_timesteps', 1000000.0),
('optimize_memory_usage', False),
('policy', 'CnnPolicy'),
('target_update_interval', 1000),
('train_freq', 4),
('normalize', False)])
```
|
CoreyMorris/testpyramidsrnd | CoreyMorris | 2022-09-19T14:50:13Z | 9 | 0 | ml-agents | [
"ml-agents",
"tensorboard",
"onnx",
"unity-ml-agents",
"deep-reinforcement-learning",
"reinforcement-learning",
"ML-Agents-Pyramids",
"region:us"
]
| reinforcement-learning | 2022-09-19T14:47:27Z |
---
tags:
- unity-ml-agents
- ml-agents
- deep-reinforcement-learning
- reinforcement-learning
- ML-Agents-Pyramids
library_name: ml-agents
---
# **ppo** Agent playing **Pyramids**
This is a trained model of a **ppo** agent playing **Pyramids** using the [Unity ML-Agents Library](https://github.com/Unity-Technologies/ml-agents).
## Usage (with ML-Agents)
The Documentation: https://github.com/huggingface/ml-agents#get-started
We wrote a complete tutorial to learn to train your first agent using ML-Agents and publish it to the Hub:
### Resume the training
```
mlagents-learn <your_configuration_file_path.yaml> --run-id=<run_id> --resume
```
### Watch your Agent play
You can watch your agent **playing directly in your browser:**.
1. Go to https://huggingface.co/spaces/unity/ML-Agents-Pyramids
2. Step 1: Write your model_id: CoreyMorris/testpyramidsrnd
3. Step 2: Select your *.nn /*.onnx file
4. Click on Watch the agent play 👀
|
sd-concepts-library/black-and-white-design | sd-concepts-library | 2022-09-19T13:27:24Z | 0 | 6 | null | [
"license:mit",
"region:us"
]
| null | 2022-09-19T13:27:11Z | ---
license: mit
---
### black and white design on Stable Diffusion
This is the `<PM_style>` concept taught to Stable Diffusion via Textual Inversion. You can load this concept into the [Stable Conceptualizer](https://colab.research.google.com/github/huggingface/notebooks/blob/main/diffusers/stable_conceptualizer_inference.ipynb) notebook. You can also train your own concepts and load them into the concept libraries using [this notebook](https://colab.research.google.com/github/huggingface/notebooks/blob/main/diffusers/sd_textual_inversion_training.ipynb).
Here is the new concept you will be able to use as a `style`:
|
gokuls/bert-uncased-massive-intent-classification | gokuls | 2022-09-19T12:24:20Z | 106 | 0 | transformers | [
"transformers",
"pytorch",
"tensorboard",
"bert",
"text-classification",
"generated_from_trainer",
"dataset:massive",
"license:apache-2.0",
"model-index",
"autotrain_compatible",
"endpoints_compatible",
"region:us"
]
| text-classification | 2022-09-19T11:43:19Z | ---
license: apache-2.0
tags:
- generated_from_trainer
datasets:
- massive
metrics:
- accuracy
model-index:
- name: bert-uncased-massive-intent-classification
results:
- task:
name: Text Classification
type: text-classification
dataset:
name: massive
type: massive
config: en-US
split: train
args: en-US
metrics:
- name: Accuracy
type: accuracy
value: 0.8853910477127398
---
<!-- This model card has been generated automatically according to the information the Trainer had access to. You
should probably proofread and complete it, then remove this comment. -->
# bert-uncased-massive-intent-classification
This model is a fine-tuned version of [bert-base-uncased](https://huggingface.co/bert-base-uncased) on the massive dataset.
It achieves the following results on the evaluation set:
- Loss: 0.8396
- Accuracy: 0.8854
## Model description
More information needed
## Intended uses & limitations
More information needed
## Training and evaluation data
More information needed
## Training procedure
### Training hyperparameters
The following hyperparameters were used during training:
- learning_rate: 5e-05
- train_batch_size: 16
- eval_batch_size: 16
- seed: 33
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
- lr_scheduler_type: linear
- num_epochs: 15
- mixed_precision_training: Native AMP
### Training results
| Training Loss | Epoch | Step | Validation Loss | Accuracy |
|:-------------:|:-----:|:-----:|:---------------:|:--------:|
| 1.4984 | 1.0 | 720 | 0.6402 | 0.8495 |
| 0.4376 | 2.0 | 1440 | 0.5394 | 0.8731 |
| 0.2318 | 3.0 | 2160 | 0.5903 | 0.8760 |
| 0.1414 | 4.0 | 2880 | 0.6221 | 0.8805 |
| 0.087 | 5.0 | 3600 | 0.7072 | 0.8819 |
| 0.0622 | 6.0 | 4320 | 0.7121 | 0.8819 |
| 0.036 | 7.0 | 5040 | 0.7750 | 0.8805 |
| 0.0234 | 8.0 | 5760 | 0.7767 | 0.8834 |
| 0.0157 | 9.0 | 6480 | 0.8243 | 0.8805 |
| 0.0122 | 10.0 | 7200 | 0.8198 | 0.8839 |
| 0.0092 | 11.0 | 7920 | 0.8105 | 0.8849 |
| 0.0047 | 12.0 | 8640 | 0.8561 | 0.8844 |
| 0.0038 | 13.0 | 9360 | 0.8367 | 0.8815 |
| 0.0029 | 14.0 | 10080 | 0.8396 | 0.8854 |
| 0.0014 | 15.0 | 10800 | 0.8410 | 0.8849 |
### Framework versions
- Transformers 4.22.1
- Pytorch 1.12.1+cu113
- Datasets 2.4.0
- Tokenizers 0.12.1
|
KoboldAI/OPT-13B-Nerys-v2 | KoboldAI | 2022-09-19T11:15:55Z | 4,494 | 12 | transformers | [
"transformers",
"pytorch",
"opt",
"text-generation",
"en",
"arxiv:2205.01068",
"license:other",
"autotrain_compatible",
"text-generation-inference",
"endpoints_compatible",
"region:us"
]
| text-generation | 2022-09-19T07:52:07Z | ---
language: en
license: other
commercial: no
---
# OPT 13B - Nerys
## Model Description
OPT 13B-Nerys is a finetune created using Facebook's OPT model.
## Training data
The training data contains around 2500 ebooks in various genres (the "Pike" dataset), a CYOA dataset called "CYS" and 50 Asian "Light Novels" (the "Manga-v1" dataset).
Most parts of the dataset have been prepended using the following text: `[Genre: <genre1>, <genre2>]`
This dataset has been cleaned in the same way as fairseq-dense-13B-Nerys-v2
### 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/OPT-13B-Nerys-v2')
>>> generator("Welcome Captain Janeway, I apologize for the delay.", do_sample=True, min_length=50)
[{'generated_text': 'Welcome Captain Janeway, I apologize for the delay."\nIt's all right," Janeway said. "I'm certain that you're doing your best to keep me informed of what\'s going on."'}]
```
### Limitations and Biases
Based on known problems with NLP technology, potential relevant factors include bias (gender, profession, race and religion).
### License
OPT-6B is licensed under the OPT-175B license, Copyright (c) Meta Platforms, Inc. All Rights Reserved.
### BibTeX entry and citation info
```
@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}
}
``` |
chintagunta85/electramed-small-ADE-DRUG-DOSAGE-ner | chintagunta85 | 2022-09-19T10:48:04Z | 104 | 0 | transformers | [
"transformers",
"pytorch",
"tensorboard",
"electra",
"token-classification",
"generated_from_trainer",
"dataset:ade_drug_dosage_ner",
"model-index",
"autotrain_compatible",
"endpoints_compatible",
"region:us"
]
| token-classification | 2022-09-19T10:46:52Z | ---
tags:
- generated_from_trainer
datasets:
- ade_drug_dosage_ner
metrics:
- precision
- recall
- f1
- accuracy
model-index:
- name: electramed-small-ADE-DRUG-DOSAGE-ner
results:
- task:
name: Token Classification
type: token-classification
dataset:
name: ade_drug_dosage_ner
type: ade_drug_dosage_ner
config: ade
split: train
args: ade
metrics:
- name: Precision
type: precision
value: 0.0
- name: Recall
type: recall
value: 0.0
- name: F1
type: f1
value: 0.0
- name: Accuracy
type: accuracy
value: 0.8697318007662835
---
<!-- This model card has been generated automatically according to the information the Trainer had access to. You
should probably proofread and complete it, then remove this comment. -->
# electramed-small-ADE-DRUG-DOSAGE-ner
This model is a fine-tuned version of [giacomomiolo/electramed_small_scivocab](https://huggingface.co/giacomomiolo/electramed_small_scivocab) on the ade_drug_dosage_ner dataset.
It achieves the following results on the evaluation set:
- Loss: 0.6064
- Precision: 0.0
- Recall: 0.0
- F1: 0.0
- Accuracy: 0.8697
## Model description
More information needed
## Intended uses & limitations
More information needed
## Training and evaluation data
More information needed
## Training procedure
### Training hyperparameters
The following hyperparameters were used during training:
- learning_rate: 2e-05
- train_batch_size: 16
- eval_batch_size: 16
- seed: 42
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
- lr_scheduler_type: linear
- num_epochs: 10
### Training results
| Training Loss | Epoch | Step | Validation Loss | Precision | Recall | F1 | Accuracy |
|:-------------:|:-----:|:----:|:---------------:|:---------:|:------:|:------:|:--------:|
| 1.4165 | 1.0 | 14 | 1.3965 | 0.0255 | 0.0636 | 0.0365 | 0.7471 |
| 1.2063 | 2.0 | 28 | 1.1702 | 0.0 | 0.0 | 0.0 | 0.8697 |
| 0.9527 | 3.0 | 42 | 0.9342 | 0.0 | 0.0 | 0.0 | 0.8697 |
| 0.8238 | 4.0 | 56 | 0.7775 | 0.0 | 0.0 | 0.0 | 0.8697 |
| 0.7452 | 5.0 | 70 | 0.6945 | 0.0 | 0.0 | 0.0 | 0.8697 |
| 0.6386 | 6.0 | 84 | 0.6519 | 0.0 | 0.0 | 0.0 | 0.8697 |
| 0.6742 | 7.0 | 98 | 0.6294 | 0.0 | 0.0 | 0.0 | 0.8697 |
| 0.6669 | 8.0 | 112 | 0.6162 | 0.0 | 0.0 | 0.0 | 0.8697 |
| 0.6595 | 9.0 | 126 | 0.6090 | 0.0 | 0.0 | 0.0 | 0.8697 |
| 0.6122 | 10.0 | 140 | 0.6064 | 0.0 | 0.0 | 0.0 | 0.8697 |
### Framework versions
- Transformers 4.22.1
- Pytorch 1.12.1+cu113
- Datasets 2.4.0
- Tokenizers 0.12.1
|
balabis/layoutlmv3-finetuned-invoice | balabis | 2022-09-19T10:36:11Z | 80 | 0 | transformers | [
"transformers",
"pytorch",
"tensorboard",
"layoutlmv3",
"token-classification",
"generated_from_trainer",
"dataset:invoices",
"license:cc-by-nc-sa-4.0",
"model-index",
"autotrain_compatible",
"endpoints_compatible",
"region:us"
]
| token-classification | 2022-09-19T10:16:08Z | ---
license: cc-by-nc-sa-4.0
tags:
- generated_from_trainer
datasets:
- invoices
metrics:
- precision
- recall
- f1
- accuracy
model-index:
- name: layoutlmv3-finetuned-invoice
results:
- task:
name: Token Classification
type: token-classification
dataset:
name: invoices
type: invoices
config: sroie
split: train
args: sroie
metrics:
- name: Precision
type: precision
value: 0.975
- name: Recall
type: recall
value: 0.975
- name: F1
type: f1
value: 0.975
- name: Accuracy
type: accuracy
value: 0.975
---
<!-- This model card has been generated automatically according to the information the Trainer had access to. You
should probably proofread and complete it, then remove this comment. -->
# layoutlmv3-finetuned-invoice
This model is a fine-tuned version of [microsoft/layoutlmv3-base](https://huggingface.co/microsoft/layoutlmv3-base) on the invoices dataset.
It achieves the following results on the evaluation set:
- Loss: 0.2299
- Precision: 0.975
- Recall: 0.975
- F1: 0.975
- Accuracy: 0.975
## Model description
More information needed
## Intended uses & limitations
More information needed
## Training and evaluation data
More information needed
## Training procedure
### Training hyperparameters
The following hyperparameters were used during training:
- learning_rate: 1e-05
- train_batch_size: 2
- eval_batch_size: 2
- seed: 42
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
- lr_scheduler_type: linear
- training_steps: 2000
### Training results
| Training Loss | Epoch | Step | Validation Loss | Precision | Recall | F1 | Accuracy |
|:-------------:|:------:|:----:|:---------------:|:---------:|:------:|:-----:|:--------:|
| No log | 14.29 | 100 | 0.1616 | 0.975 | 0.975 | 0.975 | 0.975 |
| No log | 28.57 | 200 | 0.1909 | 0.975 | 0.975 | 0.975 | 0.975 |
| No log | 42.86 | 300 | 0.2046 | 0.975 | 0.975 | 0.975 | 0.975 |
| No log | 57.14 | 400 | 0.2134 | 0.975 | 0.975 | 0.975 | 0.975 |
| 0.1239 | 71.43 | 500 | 0.2299 | 0.975 | 0.975 | 0.975 | 0.975 |
| 0.1239 | 85.71 | 600 | 0.2309 | 0.975 | 0.975 | 0.975 | 0.975 |
| 0.1239 | 100.0 | 700 | 0.2342 | 0.975 | 0.975 | 0.975 | 0.975 |
| 0.1239 | 114.29 | 800 | 0.2407 | 0.975 | 0.975 | 0.975 | 0.975 |
| 0.1239 | 128.57 | 900 | 0.2428 | 0.975 | 0.975 | 0.975 | 0.975 |
| 0.0007 | 142.86 | 1000 | 0.2449 | 0.975 | 0.975 | 0.975 | 0.975 |
| 0.0007 | 157.14 | 1100 | 0.2465 | 0.975 | 0.975 | 0.975 | 0.975 |
| 0.0007 | 171.43 | 1200 | 0.2488 | 0.975 | 0.975 | 0.975 | 0.975 |
| 0.0007 | 185.71 | 1300 | 0.2515 | 0.975 | 0.975 | 0.975 | 0.975 |
| 0.0007 | 200.0 | 1400 | 0.2525 | 0.975 | 0.975 | 0.975 | 0.975 |
| 0.0004 | 214.29 | 1500 | 0.2540 | 0.975 | 0.975 | 0.975 | 0.975 |
| 0.0004 | 228.57 | 1600 | 0.2557 | 0.975 | 0.975 | 0.975 | 0.975 |
| 0.0004 | 242.86 | 1700 | 0.2564 | 0.975 | 0.975 | 0.975 | 0.975 |
| 0.0004 | 257.14 | 1800 | 0.2570 | 0.975 | 0.975 | 0.975 | 0.975 |
| 0.0004 | 271.43 | 1900 | 0.2573 | 0.975 | 0.975 | 0.975 | 0.975 |
| 0.0003 | 285.71 | 2000 | 0.2574 | 0.975 | 0.975 | 0.975 | 0.975 |
### Framework versions
- Transformers 4.23.0.dev0
- Pytorch 1.12.1+cu113
- Datasets 2.4.0
- Tokenizers 0.12.1
|
sd-concepts-library/rilakkuma | sd-concepts-library | 2022-09-19T09:50:01Z | 0 | 3 | null | [
"license:mit",
"region:us"
]
| null | 2022-09-19T09:49:57Z | ---
license: mit
---
### rilakkuma on Stable Diffusion
This is the `<rilakkuma>` concept taught to Stable Diffusion via Textual Inversion. You can load this concept into the [Stable Conceptualizer](https://colab.research.google.com/github/huggingface/notebooks/blob/main/diffusers/stable_conceptualizer_inference.ipynb) notebook. You can also train your own concepts and load them into the concept libraries using [this notebook](https://colab.research.google.com/github/huggingface/notebooks/blob/main/diffusers/sd_textual_inversion_training.ipynb).
Here is the new concept you will be able to use as an `object`:
|
IIIT-L/xlm-roberta-large-finetuned-ours-DS | IIIT-L | 2022-09-19T09:45:19Z | 103 | 0 | transformers | [
"transformers",
"pytorch",
"xlm-roberta",
"text-classification",
"generated_from_trainer",
"license:mit",
"autotrain_compatible",
"endpoints_compatible",
"region:us"
]
| text-classification | 2022-09-19T09:24:03Z | ---
license: mit
tags:
- generated_from_trainer
metrics:
- accuracy
- precision
- recall
- f1
model-index:
- name: xlm-roberta-large-finetuned-ours-DS
results: []
---
<!-- This model card has been generated automatically according to the information the Trainer had access to. You
should probably proofread and complete it, then remove this comment. -->
# xlm-roberta-large-finetuned-ours-DS
This model is a fine-tuned version of [xlm-roberta-large](https://huggingface.co/xlm-roberta-large) on an unknown dataset.
It achieves the following results on the evaluation set:
- Loss: 0.9568
- Accuracy: 0.71
- Precision: 0.6689
- Recall: 0.6607
- F1: 0.6637
## Model description
More information needed
## Intended uses & limitations
More information needed
## Training and evaluation data
More information needed
## Training procedure
### Training hyperparameters
The following hyperparameters were used during training:
- learning_rate: 1.6820964947491663e-05
- train_batch_size: 16
- eval_batch_size: 16
- seed: 43
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
- lr_scheduler_type: linear
- lr_scheduler_warmup_ratio: 0.1
- num_epochs: 6
### Training results
| Training Loss | Epoch | Step | Validation Loss | Accuracy | Precision | Recall | F1 |
|:-------------:|:-----:|:----:|:---------------:|:--------:|:---------:|:------:|:------:|
| 0.9953 | 1.99 | 199 | 0.7955 | 0.66 | 0.7533 | 0.5971 | 0.5352 |
| 0.6638 | 3.98 | 398 | 0.8043 | 0.735 | 0.7068 | 0.6782 | 0.6846 |
| 0.3457 | 5.97 | 597 | 0.9568 | 0.71 | 0.6689 | 0.6607 | 0.6637 |
### Framework versions
- Transformers 4.20.1
- Pytorch 1.10.1+cu111
- Datasets 2.3.2
- Tokenizers 0.12.1
|
CAiRE/wav2vec2-large-xlsr-53-cantonese | CAiRE | 2022-09-19T07:50:42Z | 388 | 2 | transformers | [
"transformers",
"pytorch",
"wav2vec2",
"automatic-speech-recognition",
"audio",
"speech",
"xlsr-fine-tuning-week",
"yue",
"dataset:common_voice",
"license:apache-2.0",
"model-index",
"endpoints_compatible",
"region:us"
]
| automatic-speech-recognition | 2022-04-09T07:23:48Z | ---
language:
- yue
datasets:
- common_voice
metrics:
- cer
tags:
- audio
- automatic-speech-recognition
- speech
- xlsr-fine-tuning-week
license: apache-2.0
model-index:
- name: Wav2Vec2-Large-XLSR-53-Cantonese
results:
- task:
name: Speech Recognition
type: automatic-speech-recognition
dataset:
name: Common Voice zh-HK
type: common_voice
args: zh-HK
metrics:
- name: Test CER
type: cer
value: [18.55%]
---
# Wav2Vec2-Large-XLSR-53-Cantonese
Fine-tuned [facebook/wav2vec2-large-xlsr-53](https://huggingface.co/facebook/wav2vec2-large-xlsr-53) on Cantonese using the [Common Voice Corpus 8.0](https://commonvoice.mozilla.org/en/datasets).
When using this model, make sure that your speech input is sampled at 16kHz.
The Common Voice's validated `train` and `dev` were used for training.
The script used for training can be found at [https://github.com/holylovenia/wav2vec2-pretraining](https://github.com/holylovenia/wav2vec2-pretraining).
## Usage
The model can be used directly (without a language model) as follows:
```python
import torch
import torchaudio
from datasets import load_dataset
from transformers import Wav2Vec2ForCTC, Wav2Vec2Processor
test_dataset = load_dataset("common_voice", "zh-HK", split="test[:2%]")
processor = Wav2Vec2Processor.from_pretrained("CAiRE/wav2vec2-large-xlsr-53-cantonese")
model = Wav2Vec2ForCTC.from_pretrained("CAiRE/wav2vec2-large-xlsr-53-cantonese")
# Preprocessing the datasets.
# We need to read the aduio files as arrays
def speech_file_to_array_fn(batch):
speech_array, sampling_rate = torchaudio.load(batch["path"])
resampler = torchaudio.transforms.Resample(sampling_rate, 16_000)
batch["speech"] = resampler(speech_array).squeeze().numpy()
return batch
test_dataset = test_dataset.map(speech_file_to_array_fn)
inputs = processor(test_dataset[:2]["speech"], sampling_rate=16_000, return_tensors="pt", padding=True)
with torch.no_grad():
logits = model(inputs.input_values, attention_mask=inputs.attention_mask).logits
predicted_ids = torch.argmax(logits, dim=-1)
print("Prediction:", processor.batch_decode(predicted_ids))
print("Reference:", test_dataset[:2]["sentence"])
```
## Evaluation
The model can be evaluated as follows on the zh-HK test data of Common Voice.
```python
import torch
import torchaudio
from datasets import load_dataset, load_metric
from transformers import Wav2Vec2ForCTC, Wav2Vec2Processor
import re
test_dataset = load_dataset("common_voice", "zh-HK", split="test")
wer = load_metric("cer")
processor = Wav2Vec2Processor.from_pretrained("CAiRE/wav2vec2-large-xlsr-53-cantonese")
model = Wav2Vec2ForCTC.from_pretrained("CAiRE/wav2vec2-large-xlsr-53-cantonese")
model.to("cuda")
chars_to_ignore_regex = '[\,\?\.\!\-\;\:\"\“\%\‘\'\”\�]'
# Preprocessing the datasets.
# We need to read the aduio files as arrays
def speech_file_to_array_fn(batch):
batch["sentence"] = re.sub(chars_to_ignore_regex, '', batch["sentence"]).lower()
speech_array, sampling_rate = torchaudio.load(batch["path"])
resampler = torchaudio.transforms.Resample(sampling_rate, 16_000)
batch["speech"] = resampler(speech_array).squeeze().numpy()
return batch
test_dataset = test_dataset.map(speech_file_to_array_fn)
# Preprocessing the datasets.
# We need to read the aduio files as arrays
def evaluate(batch):
inputs = processor(batch["speech"], sampling_rate=16_000, return_tensors="pt", padding=True)
with torch.no_grad():
logits = model(inputs.input_values.to("cuda"), attention_mask=inputs.attention_mask.to("cuda")).logits
pred_ids = torch.argmax(logits, dim=-1)
batch["pred_strings"] = processor.batch_decode(pred_ids)
return batch
result = test_dataset.map(evaluate, batched=True, batch_size=8)
print("CER: {:2f}".format(100 * cer.compute(predictions=result["pred_strings"], references=result["sentence"])))
```
**Test Result**: CER: 18.55 %
## Citation
If you use our code/model, please cite us:
```
@inproceedings{lovenia2022ascend,
title={ASCEND: A Spontaneous Chinese-English Dataset for Code-switching in Multi-turn Conversation},
author={Lovenia, Holy and Cahyawijaya, Samuel and Winata, Genta Indra and Xu, Peng and Yan, Xu and Liu, Zihan and Frieske, Rita and Yu, Tiezheng and Dai, Wenliang and Barezi, Elham J and others},
booktitle={Proceedings of the 13th Language Resources and Evaluation Conference (LREC)},
year={2022}
}
``` |
KoboldAI/OPT-2.7B-Erebus | KoboldAI | 2022-09-19T07:38:12Z | 5,758 | 39 | transformers | [
"transformers",
"pytorch",
"opt",
"text-generation",
"en",
"arxiv:2205.01068",
"license:other",
"autotrain_compatible",
"text-generation-inference",
"region:us"
]
| text-generation | 2022-09-19T06:41:21Z | ---
language: en
license: other
commercial: no
inference: false
---
# OPT 2.7B - Erebus
## Model description
This is the second generation of the original Shinen made by Mr. Seeker. The full dataset consists of 6 different sources, all surrounding the "Adult" theme. The name "Erebus" comes from the greek mythology, also named "darkness". This is in line with Shin'en, or "deep abyss". For inquiries, please contact the KoboldAI community. **Warning: THIS model is NOT suitable for use by minors. The model will output X-rated content.**
## Training data
The data can be divided in 6 different datasets:
- Literotica (everything with 4.5/5 or higher)
- Sexstories (everything with 90 or higher)
- Dataset-G (private dataset of X-rated stories)
- Doc's Lab (all stories)
- Pike Dataset (novels with "adult" rating)
- SoFurry (collection of various animals)
The dataset uses `[Genre: <comma-separated list of genres>]` for tagging.
### 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/OPT-2.7B-Erebus')
>>> generator("Welcome Captain Janeway, I apologize for the delay.", do_sample=True, min_length=50)
[{'generated_text': 'Welcome Captain Janeway, I apologize for the delay."\nIt's all right," Janeway said. "I'm certain that you're doing your best to keep me informed of what\'s going on."'}]
```
## Limitations and biases
Based on known problems with NLP technology, potential relevant factors include bias (gender, profession, race and religion). **Warning: This model has a very strong NSFW bias!**
### License
OPT-6.7B is licensed under the OPT-175B license, Copyright (c) Meta Platforms, Inc. All Rights Reserved.
### BibTeX entry and citation info
```
@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}
}
``` |
KoboldAI/OPT-2.7B-Nerys-v2 | KoboldAI | 2022-09-19T07:19:35Z | 1,490 | 6 | transformers | [
"transformers",
"pytorch",
"opt",
"text-generation",
"en",
"arxiv:2205.01068",
"license:other",
"autotrain_compatible",
"text-generation-inference",
"endpoints_compatible",
"region:us"
]
| text-generation | 2022-09-19T06:50:39Z | ---
language: en
license: other
commercial: no
---
# OPT 2.7B - Nerys
## Model Description
OPT 2.7B-Nerys is a finetune created using Facebook's OPT model.
## Training data
The training data contains around 2500 ebooks in various genres (the "Pike" dataset), a CYOA dataset called "CYS" and 50 Asian "Light Novels" (the "Manga-v1" dataset).
Most parts of the dataset have been prepended using the following text: `[Genre: <genre1>, <genre2>]`
This dataset has been cleaned in the same way as fairseq-dense-13B-Nerys-v2
### 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/OPT-2.7B-Nerys-v2')
>>> generator("Welcome Captain Janeway, I apologize for the delay.", do_sample=True, min_length=50)
[{'generated_text': 'Welcome Captain Janeway, I apologize for the delay."\nIt's all right," Janeway said. "I'm certain that you're doing your best to keep me informed of what\'s going on."'}]
```
### Limitations and Biases
Based on known problems with NLP technology, potential relevant factors include bias (gender, profession, race and religion).
### License
OPT-6B is licensed under the OPT-175B license, Copyright (c) Meta Platforms, Inc. All Rights Reserved.
### BibTeX entry and citation info
```
@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}
}
``` |
sd-concepts-library/tudisco | sd-concepts-library | 2022-09-19T06:46:30Z | 0 | 11 | null | [
"license:mit",
"region:us"
]
| null | 2022-09-19T06:46:23Z | ---
license: mit
---
### tudisco on Stable Diffusion
This is the `<cat-toy>` concept taught to Stable Diffusion via Textual Inversion. You can load this concept into the [Stable Conceptualizer](https://colab.research.google.com/github/huggingface/notebooks/blob/main/diffusers/stable_conceptualizer_inference.ipynb) notebook. You can also train your own concepts and load them into the concept libraries using [this notebook](https://colab.research.google.com/github/huggingface/notebooks/blob/main/diffusers/sd_textual_inversion_training.ipynb).
Here is the new concept you will be able to use as a `style`:
|
sd-concepts-library/zk | sd-concepts-library | 2022-09-19T06:40:35Z | 0 | 3 | null | [
"license:mit",
"region:us"
]
| null | 2022-09-19T06:40:28Z | ---
license: mit
---
### zk on Stable Diffusion
This is the `zk` concept taught to Stable Diffusion via Textual Inversion. You can load this concept into the [Stable Conceptualizer](https://colab.research.google.com/github/huggingface/notebooks/blob/main/diffusers/stable_conceptualizer_inference.ipynb) notebook. You can also train your own concepts and load them into the concept libraries using [this notebook](https://colab.research.google.com/github/huggingface/notebooks/blob/main/diffusers/sd_textual_inversion_training.ipynb).
Here is the new concept you will be able to use as a `style`:
|
LTP/base2 | LTP | 2022-09-19T06:36:21Z | 125 | 7 | transformers | [
"transformers",
"pytorch",
"arxiv:2009.11616",
"endpoints_compatible",
"region:us"
]
| null | 2022-08-14T04:36:19Z | 
| Language | version |
| ------------------------------------ | ------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- |
| [Python](python/interface/README.md) | [](https://pypi.org/project/ltp) [](https://pypi.org/project/ltp-core) [](https://pypi.org/project/ltp-extension) |
| [Rust](rust/ltp/README.md) | [](https://crates.io/crates/ltp) |
# LTP 4
LTP（Language Technology Platform） 提供了一系列中文自然语言处理工具，用户可以使用这些工具对于中文文本进行分词、词性标注、句法分析等等工作。
## 引用
如果您在工作中使用了 LTP，您可以引用这篇论文
```bibtex
@article{che2020n,
title={N-LTP: A Open-source Neural Chinese Language Technology Platform with Pretrained Models},
author={Che, Wanxiang and Feng, Yunlong and Qin, Libo and Liu, Ting},
journal={arXiv preprint arXiv:2009.11616},
year={2020}
}
```
**参考书：**
由哈工大社会计算与信息检索研究中心（HIT-SCIR）的多位学者共同编著的《[自然语言处理：基于预训练模型的方法](https://item.jd.com/13344628.html)
》（作者：车万翔、郭江、崔一鸣；主审：刘挺）一书现已正式出版，该书重点介绍了新的基于预训练模型的自然语言处理技术，包括基础知识、预训练词向量和预训练模型三大部分，可供广大LTP用户学习参考。
### 更新说明
- 4.2.0
- \[结构性变化\] 将 LTP 拆分成 2 个部分，维护和训练更方便，结构更清晰
- \[Legacy 模型\] 针对广大用户对于**推理速度**的需求，使用 Rust 重写了基于感知机的算法，准确率与 LTP3 版本相当，速度则是 LTP v3 的 **3.55** 倍，开启多线程更可获得 **17.17** 倍的速度提升，但目前仅支持分词、词性、命名实体三大任务
- \[深度学习模型\] 即基于 PyTorch 实现的深度学习模型，支持全部的6大任务（分词/词性/命名实体/语义角色/依存句法/语义依存）
- \[其他改进\] 改进了模型训练方法
- \[共同\] 提供了训练脚本和训练样例，使得用户能够更方便地使用私有的数据，自行训练个性化的模型
- \[深度学习模型\] 采用 hydra 对训练过程进行配置，方便广大用户修改模型训练参数以及对 LTP 进行扩展（比如使用其他包中的 Module）
- \[其他变化\] 分词、依存句法分析 (Eisner) 和 语义依存分析 (Eisner) 任务的解码算法使用 Rust 实现，速度更快
- \[新特性\] 模型上传至 [Huggingface Hub](https://huggingface.co/LTP)，支持自动下载，下载速度更快，并且支持用户自行上传自己训练的模型供LTP进行推理使用
- \[破坏性变更\] 改用 Pipeline API 进行推理，方便后续进行更深入的性能优化（如SDP和SDPG很大一部分是重叠的，重用可以加快推理速度），使用说明参见[Github快速使用部分](https://github.com/hit-scir/ltp)
- 4.1.0
- 提供了自定义分词等功能
- 修复了一些bug
- 4.0.0
- 基于Pytorch 开发，原生 Python 接口
- 可根据需要自由选择不同速度和指标的模型
- 分词、词性、命名实体、依存句法、语义角色、语义依存6大任务
## 快速使用
### [Python](python/interface/README.md)
```bash
pip install -U ltp ltp-core ltp-extension -i https://pypi.org/simple # 安装 ltp
```
**注：** 如果遇到任何错误，请尝试使用上述命令重新安装 ltp，如果依然报错，请在 Github issues 中反馈。
```python
import torch
from ltp import LTP
ltp = LTP("LTP/small") # 默认加载 Small 模型
# 将模型移动到 GPU 上
if torch.cuda.is_available():
# ltp.cuda()
ltp.to("cuda")
output = ltp.pipeline(["他叫汤姆去拿外衣。"], tasks=["cws", "pos", "ner", "srl", "dep", "sdp"])
# 使用字典格式作为返回结果
print(output.cws) # print(output[0]) / print(output['cws']) # 也可以使用下标访问
print(output.pos)
print(output.sdp)
# 使用感知机算法实现的分词、词性和命名实体识别，速度比较快，但是精度略低
ltp = LTP("LTP/legacy")
# cws, pos, ner = ltp.pipeline(["他叫汤姆去拿外衣。"], tasks=["cws", "ner"]).to_tuple() # error: NER 需要 词性标注任务的结果
cws, pos, ner = ltp.pipeline(["他叫汤姆去拿外衣。"], tasks=["cws", "pos", "ner"]).to_tuple() # to tuple 可以自动转换为元组格式
# 使用元组格式作为返回结果
print(cws, pos, ner)
```
**[详细说明](python/interface/docs/quickstart.rst)**
### [Rust](rust/ltp/README.md)
```rust
use std::fs::File;
use itertools::multizip;
use ltp::{CWSModel, POSModel, NERModel, ModelSerde, Format, Codec};
fn main() -> Result<(), Box<dyn std::error::Error>> {
let file = File::open("data/legacy-models/cws_model.bin")?;
let cws: CWSModel = ModelSerde::load(file, Format::AVRO(Codec::Deflate))?;
let file = File::open("data/legacy-models/pos_model.bin")?;
let pos: POSModel = ModelSerde::load(file, Format::AVRO(Codec::Deflate))?;
let file = File::open("data/legacy-models/ner_model.bin")?;
let ner: NERModel = ModelSerde::load(file, Format::AVRO(Codec::Deflate))?;
let words = cws.predict("他叫汤姆去拿外衣。")?;
let pos = pos.predict(&words)?;
let ner = ner.predict((&words, &pos))?;
for (w, p, n) in multizip((words, pos, ner)) {
println!("{}/{}/{}", w, p, n);
}
Ok(())
}
```
## 模型性能以及下载地址
| 深度学习模型 | 分词 | 词性 | 命名实体 | 语义角色 | 依存句法 | 语义依存 | 速度(句/S) |
| :---------------------------------------: | :---: | :---: | :---: | :---: | :---: | :---: | :-----: |
| [Base](https://huggingface.co/LTP/base) | 98.7 | 98.5 | 95.4 | 80.6 | 89.5 | 75.2 | 39.12 |
| [Base1](https://huggingface.co/LTP/base1) | 99.22 | 98.73 | 96.39 | 79.28 | 89.57 | 76.57 | --.-- |
| [Base2](https://huggingface.co/LTP/base2) | 99.18 | 98.69 | 95.97 | 79.49 | 90.19 | 76.62 | --.-- |
| [Small](https://huggingface.co/LTP/small) | 98.4 | 98.2 | 94.3 | 78.4 | 88.3 | 74.7 | 43.13 |
| [Tiny](https://huggingface.co/LTP/tiny) | 96.8 | 97.1 | 91.6 | 70.9 | 83.8 | 70.1 | 53.22 |
| 感知机算法 | 分词 | 词性 | 命名实体 | 速度(句/s) | 备注 |
| :-----------------------------------------: | :---: | :---: | :---: | :------: | :------------------------: |
| [Legacy](https://huggingface.co/LTP/legacy) | 97.93 | 98.41 | 94.28 | 21581.48 | [性能详情](rust/ltp/README.md) |
**注：感知机算法速度为开启16线程速度**
## 构建 Wheel 包
```shell script
make bdist
```
## 其他语言绑定
**感知机算法**
- [Rust](rust/ltp)
- [C/C++](rust/ltp-cffi)
**深度学习算法**
- [Rust](https://github.com/HIT-SCIR/libltp/tree/master/ltp-rs)
- [C++](https://github.com/HIT-SCIR/libltp/tree/master/ltp-cpp)
- [Java](https://github.com/HIT-SCIR/libltp/tree/master/ltp-java)
## 作者信息
- 冯云龙 \<\<[[email protected]](mailto:[email protected])>>
## 开源协议
1. 语言技术平台面向国内外大学、中科院各研究所以及个人研究者免费开放源代码，但如上述机构和个人将该平台用于商业目的（如企业合作项目等）则需要付费。
2. 除上述机构以外的企事业单位，如申请使用该平台，需付费。
3. 凡涉及付费问题，请发邮件到 [email protected] 洽商。
4. 如果您在 LTP 基础上发表论文或取得科研成果，请您在发表论文和申报成果时声明“使用了哈工大社会计算与信息检索研究中心研制的语言技术平台（LTP）”.
同时，发信给[email protected]，说明发表论文或申报成果的题目、出处等。
|
LTP/base1 | LTP | 2022-09-19T06:36:16Z | 6,604 | 5 | transformers | [
"transformers",
"pytorch",
"arxiv:2009.11616",
"endpoints_compatible",
"region:us"
]
| null | 2022-08-14T04:35:09Z | 
| Language | version |
| ------------------------------------ | ------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- |
| [Python](python/interface/README.md) | [](https://pypi.org/project/ltp) [](https://pypi.org/project/ltp-core) [](https://pypi.org/project/ltp-extension) |
| [Rust](rust/ltp/README.md) | [](https://crates.io/crates/ltp) |
# LTP 4
LTP（Language Technology Platform） 提供了一系列中文自然语言处理工具，用户可以使用这些工具对于中文文本进行分词、词性标注、句法分析等等工作。
## 引用
如果您在工作中使用了 LTP，您可以引用这篇论文
```bibtex
@article{che2020n,
title={N-LTP: A Open-source Neural Chinese Language Technology Platform with Pretrained Models},
author={Che, Wanxiang and Feng, Yunlong and Qin, Libo and Liu, Ting},
journal={arXiv preprint arXiv:2009.11616},
year={2020}
}
```
**参考书：**
由哈工大社会计算与信息检索研究中心（HIT-SCIR）的多位学者共同编著的《[自然语言处理：基于预训练模型的方法](https://item.jd.com/13344628.html)
》（作者：车万翔、郭江、崔一鸣；主审：刘挺）一书现已正式出版，该书重点介绍了新的基于预训练模型的自然语言处理技术，包括基础知识、预训练词向量和预训练模型三大部分，可供广大LTP用户学习参考。
### 更新说明
- 4.2.0
- \[结构性变化\] 将 LTP 拆分成 2 个部分，维护和训练更方便，结构更清晰
- \[Legacy 模型\] 针对广大用户对于**推理速度**的需求，使用 Rust 重写了基于感知机的算法，准确率与 LTP3 版本相当，速度则是 LTP v3 的 **3.55** 倍，开启多线程更可获得 **17.17** 倍的速度提升，但目前仅支持分词、词性、命名实体三大任务
- \[深度学习模型\] 即基于 PyTorch 实现的深度学习模型，支持全部的6大任务（分词/词性/命名实体/语义角色/依存句法/语义依存）
- \[其他改进\] 改进了模型训练方法
- \[共同\] 提供了训练脚本和训练样例，使得用户能够更方便地使用私有的数据，自行训练个性化的模型
- \[深度学习模型\] 采用 hydra 对训练过程进行配置，方便广大用户修改模型训练参数以及对 LTP 进行扩展（比如使用其他包中的 Module）
- \[其他变化\] 分词、依存句法分析 (Eisner) 和 语义依存分析 (Eisner) 任务的解码算法使用 Rust 实现，速度更快
- \[新特性\] 模型上传至 [Huggingface Hub](https://huggingface.co/LTP)，支持自动下载，下载速度更快，并且支持用户自行上传自己训练的模型供LTP进行推理使用
- \[破坏性变更\] 改用 Pipeline API 进行推理，方便后续进行更深入的性能优化（如SDP和SDPG很大一部分是重叠的，重用可以加快推理速度），使用说明参见[Github快速使用部分](https://github.com/hit-scir/ltp)
- 4.1.0
- 提供了自定义分词等功能
- 修复了一些bug
- 4.0.0
- 基于Pytorch 开发，原生 Python 接口
- 可根据需要自由选择不同速度和指标的模型
- 分词、词性、命名实体、依存句法、语义角色、语义依存6大任务
## 快速使用
### [Python](python/interface/README.md)
```bash
pip install -U ltp ltp-core ltp-extension -i https://pypi.org/simple # 安装 ltp
```
**注：** 如果遇到任何错误，请尝试使用上述命令重新安装 ltp，如果依然报错，请在 Github issues 中反馈。
```python
import torch
from ltp import LTP
ltp = LTP("LTP/small") # 默认加载 Small 模型
# 将模型移动到 GPU 上
if torch.cuda.is_available():
# ltp.cuda()
ltp.to("cuda")
output = ltp.pipeline(["他叫汤姆去拿外衣。"], tasks=["cws", "pos", "ner", "srl", "dep", "sdp"])
# 使用字典格式作为返回结果
print(output.cws) # print(output[0]) / print(output['cws']) # 也可以使用下标访问
print(output.pos)
print(output.sdp)
# 使用感知机算法实现的分词、词性和命名实体识别，速度比较快，但是精度略低
ltp = LTP("LTP/legacy")
# cws, pos, ner = ltp.pipeline(["他叫汤姆去拿外衣。"], tasks=["cws", "ner"]).to_tuple() # error: NER 需要 词性标注任务的结果
cws, pos, ner = ltp.pipeline(["他叫汤姆去拿外衣。"], tasks=["cws", "pos", "ner"]).to_tuple() # to tuple 可以自动转换为元组格式
# 使用元组格式作为返回结果
print(cws, pos, ner)
```
**[详细说明](python/interface/docs/quickstart.rst)**
### [Rust](rust/ltp/README.md)
```rust
use std::fs::File;
use itertools::multizip;
use ltp::{CWSModel, POSModel, NERModel, ModelSerde, Format, Codec};
fn main() -> Result<(), Box<dyn std::error::Error>> {
let file = File::open("data/legacy-models/cws_model.bin")?;
let cws: CWSModel = ModelSerde::load(file, Format::AVRO(Codec::Deflate))?;
let file = File::open("data/legacy-models/pos_model.bin")?;
let pos: POSModel = ModelSerde::load(file, Format::AVRO(Codec::Deflate))?;
let file = File::open("data/legacy-models/ner_model.bin")?;
let ner: NERModel = ModelSerde::load(file, Format::AVRO(Codec::Deflate))?;
let words = cws.predict("他叫汤姆去拿外衣。")?;
let pos = pos.predict(&words)?;
let ner = ner.predict((&words, &pos))?;
for (w, p, n) in multizip((words, pos, ner)) {
println!("{}/{}/{}", w, p, n);
}
Ok(())
}
```
## 模型性能以及下载地址
| 深度学习模型 | 分词 | 词性 | 命名实体 | 语义角色 | 依存句法 | 语义依存 | 速度(句/S) |
| :---------------------------------------: | :---: | :---: | :---: | :---: | :---: | :---: | :-----: |
| [Base](https://huggingface.co/LTP/base) | 98.7 | 98.5 | 95.4 | 80.6 | 89.5 | 75.2 | 39.12 |
| [Base1](https://huggingface.co/LTP/base1) | 99.22 | 98.73 | 96.39 | 79.28 | 89.57 | 76.57 | --.-- |
| [Base2](https://huggingface.co/LTP/base2) | 99.18 | 98.69 | 95.97 | 79.49 | 90.19 | 76.62 | --.-- |
| [Small](https://huggingface.co/LTP/small) | 98.4 | 98.2 | 94.3 | 78.4 | 88.3 | 74.7 | 43.13 |
| [Tiny](https://huggingface.co/LTP/tiny) | 96.8 | 97.1 | 91.6 | 70.9 | 83.8 | 70.1 | 53.22 |
| 感知机算法 | 分词 | 词性 | 命名实体 | 速度(句/s) | 备注 |
| :-----------------------------------------: | :---: | :---: | :---: | :------: | :------------------------: |
| [Legacy](https://huggingface.co/LTP/legacy) | 97.93 | 98.41 | 94.28 | 21581.48 | [性能详情](rust/ltp/README.md) |
**注：感知机算法速度为开启16线程速度**
## 构建 Wheel 包
```shell script
make bdist
```
## 其他语言绑定
**感知机算法**
- [Rust](rust/ltp)
- [C/C++](rust/ltp-cffi)
**深度学习算法**
- [Rust](https://github.com/HIT-SCIR/libltp/tree/master/ltp-rs)
- [C++](https://github.com/HIT-SCIR/libltp/tree/master/ltp-cpp)
- [Java](https://github.com/HIT-SCIR/libltp/tree/master/ltp-java)
## 作者信息
- 冯云龙 \<\<[[email protected]](mailto:[email protected])>>
## 开源协议
1. 语言技术平台面向国内外大学、中科院各研究所以及个人研究者免费开放源代码，但如上述机构和个人将该平台用于商业目的（如企业合作项目等）则需要付费。
2. 除上述机构以外的企事业单位，如申请使用该平台，需付费。
3. 凡涉及付费问题，请发邮件到 [email protected] 洽商。
4. 如果您在 LTP 基础上发表论文或取得科研成果，请您在发表论文和申报成果时声明“使用了哈工大社会计算与信息检索研究中心研制的语言技术平台（LTP）”.
同时，发信给[email protected]，说明发表论文或申报成果的题目、出处等。
|
LTP/legacy | LTP | 2022-09-19T06:35:53Z | 207 | 4 | transformers | [
"transformers",
"arxiv:2009.11616",
"endpoints_compatible",
"region:us"
]
| null | 2022-08-14T05:05:24Z | 
| Language | version |
| ------------------------------------ | ------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- |
| [Python](python/interface/README.md) | [](https://pypi.org/project/ltp) [](https://pypi.org/project/ltp-core) [](https://pypi.org/project/ltp-extension) |
| [Rust](rust/ltp/README.md) | [](https://crates.io/crates/ltp) |
# LTP 4
LTP（Language Technology Platform） 提供了一系列中文自然语言处理工具，用户可以使用这些工具对于中文文本进行分词、词性标注、句法分析等等工作。
## 引用
如果您在工作中使用了 LTP，您可以引用这篇论文
```bibtex
@article{che2020n,
title={N-LTP: A Open-source Neural Chinese Language Technology Platform with Pretrained Models},
author={Che, Wanxiang and Feng, Yunlong and Qin, Libo and Liu, Ting},
journal={arXiv preprint arXiv:2009.11616},
year={2020}
}
```
**参考书：**
由哈工大社会计算与信息检索研究中心（HIT-SCIR）的多位学者共同编著的《[自然语言处理：基于预训练模型的方法](https://item.jd.com/13344628.html)
》（作者：车万翔、郭江、崔一鸣；主审：刘挺）一书现已正式出版，该书重点介绍了新的基于预训练模型的自然语言处理技术，包括基础知识、预训练词向量和预训练模型三大部分，可供广大LTP用户学习参考。
### 更新说明
- 4.2.0
- \[结构性变化\] 将 LTP 拆分成 2 个部分，维护和训练更方便，结构更清晰
- \[Legacy 模型\] 针对广大用户对于**推理速度**的需求，使用 Rust 重写了基于感知机的算法，准确率与 LTP3 版本相当，速度则是 LTP v3 的 **3.55** 倍，开启多线程更可获得 **17.17** 倍的速度提升，但目前仅支持分词、词性、命名实体三大任务
- \[深度学习模型\] 即基于 PyTorch 实现的深度学习模型，支持全部的6大任务（分词/词性/命名实体/语义角色/依存句法/语义依存）
- \[其他改进\] 改进了模型训练方法
- \[共同\] 提供了训练脚本和训练样例，使得用户能够更方便地使用私有的数据，自行训练个性化的模型
- \[深度学习模型\] 采用 hydra 对训练过程进行配置，方便广大用户修改模型训练参数以及对 LTP 进行扩展（比如使用其他包中的 Module）
- \[其他变化\] 分词、依存句法分析 (Eisner) 和 语义依存分析 (Eisner) 任务的解码算法使用 Rust 实现，速度更快
- \[新特性\] 模型上传至 [Huggingface Hub](https://huggingface.co/LTP)，支持自动下载，下载速度更快，并且支持用户自行上传自己训练的模型供LTP进行推理使用
- \[破坏性变更\] 改用 Pipeline API 进行推理，方便后续进行更深入的性能优化（如SDP和SDPG很大一部分是重叠的，重用可以加快推理速度），使用说明参见[Github快速使用部分](https://github.com/hit-scir/ltp)
- 4.1.0
- 提供了自定义分词等功能
- 修复了一些bug
- 4.0.0
- 基于Pytorch 开发，原生 Python 接口
- 可根据需要自由选择不同速度和指标的模型
- 分词、词性、命名实体、依存句法、语义角色、语义依存6大任务
## 快速使用
### [Python](python/interface/README.md)
```bash
pip install -U ltp ltp-core ltp-extension -i https://pypi.org/simple # 安装 ltp
```
**注：** 如果遇到任何错误，请尝试使用上述命令重新安装 ltp，如果依然报错，请在 Github issues 中反馈。
```python
import torch
from ltp import LTP
ltp = LTP("LTP/small") # 默认加载 Small 模型
# 将模型移动到 GPU 上
if torch.cuda.is_available():
# ltp.cuda()
ltp.to("cuda")
output = ltp.pipeline(["他叫汤姆去拿外衣。"], tasks=["cws", "pos", "ner", "srl", "dep", "sdp"])
# 使用字典格式作为返回结果
print(output.cws) # print(output[0]) / print(output['cws']) # 也可以使用下标访问
print(output.pos)
print(output.sdp)
# 使用感知机算法实现的分词、词性和命名实体识别，速度比较快，但是精度略低
ltp = LTP("LTP/legacy")
# cws, pos, ner = ltp.pipeline(["他叫汤姆去拿外衣。"], tasks=["cws", "ner"]).to_tuple() # error: NER 需要 词性标注任务的结果
cws, pos, ner = ltp.pipeline(["他叫汤姆去拿外衣。"], tasks=["cws", "pos", "ner"]).to_tuple() # to tuple 可以自动转换为元组格式
# 使用元组格式作为返回结果
print(cws, pos, ner)
```
**[详细说明](python/interface/docs/quickstart.rst)**
### [Rust](rust/ltp/README.md)
```rust
use std::fs::File;
use itertools::multizip;
use ltp::{CWSModel, POSModel, NERModel, ModelSerde, Format, Codec};
fn main() -> Result<(), Box<dyn std::error::Error>> {
let file = File::open("data/legacy-models/cws_model.bin")?;
let cws: CWSModel = ModelSerde::load(file, Format::AVRO(Codec::Deflate))?;
let file = File::open("data/legacy-models/pos_model.bin")?;
let pos: POSModel = ModelSerde::load(file, Format::AVRO(Codec::Deflate))?;
let file = File::open("data/legacy-models/ner_model.bin")?;
let ner: NERModel = ModelSerde::load(file, Format::AVRO(Codec::Deflate))?;
let words = cws.predict("他叫汤姆去拿外衣。")?;
let pos = pos.predict(&words)?;
let ner = ner.predict((&words, &pos))?;
for (w, p, n) in multizip((words, pos, ner)) {
println!("{}/{}/{}", w, p, n);
}
Ok(())
}
```
## 模型性能以及下载地址
| 深度学习模型 | 分词 | 词性 | 命名实体 | 语义角色 | 依存句法 | 语义依存 | 速度(句/S) |
| :---------------------------------------: | :---: | :---: | :---: | :---: | :---: | :---: | :-----: |
| [Base](https://huggingface.co/LTP/base) | 98.7 | 98.5 | 95.4 | 80.6 | 89.5 | 75.2 | 39.12 |
| [Base1](https://huggingface.co/LTP/base1) | 99.22 | 98.73 | 96.39 | 79.28 | 89.57 | 76.57 | --.-- |
| [Base2](https://huggingface.co/LTP/base2) | 99.18 | 98.69 | 95.97 | 79.49 | 90.19 | 76.62 | --.-- |
| [Small](https://huggingface.co/LTP/small) | 98.4 | 98.2 | 94.3 | 78.4 | 88.3 | 74.7 | 43.13 |
| [Tiny](https://huggingface.co/LTP/tiny) | 96.8 | 97.1 | 91.6 | 70.9 | 83.8 | 70.1 | 53.22 |
| 感知机算法 | 分词 | 词性 | 命名实体 | 速度(句/s) | 备注 |
| :-----------------------------------------: | :---: | :---: | :---: | :------: | :------------------------: |
| [Legacy](https://huggingface.co/LTP/legacy) | 97.93 | 98.41 | 94.28 | 21581.48 | [性能详情](rust/ltp/README.md) |
**注：感知机算法速度为开启16线程速度**
## 构建 Wheel 包
```shell script
make bdist
```
## 其他语言绑定
**感知机算法**
- [Rust](rust/ltp)
- [C/C++](rust/ltp-cffi)
**深度学习算法**
- [Rust](https://github.com/HIT-SCIR/libltp/tree/master/ltp-rs)
- [C++](https://github.com/HIT-SCIR/libltp/tree/master/ltp-cpp)
- [Java](https://github.com/HIT-SCIR/libltp/tree/master/ltp-java)
## 作者信息
- 冯云龙 \<\<[[email protected]](mailto:[email protected])>>
## 开源协议
1. 语言技术平台面向国内外大学、中科院各研究所以及个人研究者免费开放源代码，但如上述机构和个人将该平台用于商业目的（如企业合作项目等）则需要付费。
2. 除上述机构以外的企事业单位，如申请使用该平台，需付费。
3. 凡涉及付费问题，请发邮件到 [email protected] 洽商。
4. 如果您在 LTP 基础上发表论文或取得科研成果，请您在发表论文和申报成果时声明“使用了哈工大社会计算与信息检索研究中心研制的语言技术平台（LTP）”.
同时，发信给[email protected]，说明发表论文或申报成果的题目、出处等。
|
sd-concepts-library/mu-sadr | sd-concepts-library | 2022-09-19T04:43:07Z | 0 | 0 | null | [
"license:mit",
"region:us"
]
| null | 2022-09-19T04:43:03Z | ---
license: mit
---
### mu-sadr on Stable Diffusion
This is the `<783463b>` concept taught to Stable Diffusion via Textual Inversion. You can load this concept into the [Stable Conceptualizer](https://colab.research.google.com/github/huggingface/notebooks/blob/main/diffusers/stable_conceptualizer_inference.ipynb) notebook. You can also train your own concepts and load them into the concept libraries using [this notebook](https://colab.research.google.com/github/huggingface/notebooks/blob/main/diffusers/sd_textual_inversion_training.ipynb).
Here is the new concept you will be able to use as an `object`:
|
nguyenthanhasia/BERT-LARD | nguyenthanhasia | 2022-09-19T02:51:22Z | 105 | 0 | transformers | [
"transformers",
"pytorch",
"bert",
"feature-extraction",
"text-embeddings-inference",
"endpoints_compatible",
"region:us"
]
| feature-extraction | 2022-09-19T02:42:46Z | This is BERT-LARD Pretrained Model
|
gary109/ai-light-dance_singing4_ft_wav2vec2-large-xlsr-53-5gram-v4-2 | gary109 | 2022-09-19T01:39:28Z | 78 | 0 | transformers | [
"transformers",
"pytorch",
"tensorboard",
"wav2vec2",
"automatic-speech-recognition",
"gary109/AI_Light_Dance",
"generated_from_trainer",
"endpoints_compatible",
"region:us"
]
| automatic-speech-recognition | 2022-09-16T03:02:43Z | ---
tags:
- automatic-speech-recognition
- gary109/AI_Light_Dance
- generated_from_trainer
model-index:
- name: ai-light-dance_singing4_ft_wav2vec2-large-xlsr-53-5gram-v4-2
results: []
---
<!-- This model card has been generated automatically according to the information the Trainer had access to. You
should probably proofread and complete it, then remove this comment. -->
# ai-light-dance_singing4_ft_wav2vec2-large-xlsr-53-5gram-v4-2
This model is a fine-tuned version of [gary109/ai-light-dance_singing2_ft_wav2vec2-large-xlsr-53-5gram-v4-2](https://huggingface.co/gary109/ai-light-dance_singing2_ft_wav2vec2-large-xlsr-53-5gram-v4-2) on the GARY109/AI_LIGHT_DANCE - ONSET-SINGING4 dataset.
It achieves the following results on the evaluation set:
- Loss: 0.2311
- Wer: 0.1042
## Model description
More information needed
## Intended uses & limitations
More information needed
## Training and evaluation data
More information needed
## Training procedure
### Training hyperparameters
The following hyperparameters were used during training:
- learning_rate: 4e-05
- train_batch_size: 8
- eval_batch_size: 8
- seed: 42
- gradient_accumulation_steps: 8
- total_train_batch_size: 64
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
- lr_scheduler_type: linear
- lr_scheduler_warmup_steps: 500
- num_epochs: 100.0
- mixed_precision_training: Native AMP
### Training results
| Training Loss | Epoch | Step | Validation Loss | Wer |
|:-------------:|:-----:|:----:|:---------------:|:------:|
| 0.9329 | 1.0 | 72 | 0.4334 | 0.1349 |
| 0.7631 | 2.0 | 144 | 0.3390 | 0.1318 |
| 0.7575 | 3.0 | 216 | 0.3182 | 0.1441 |
| 0.667 | 4.0 | 288 | 0.2995 | 0.1288 |
| 0.6474 | 5.0 | 360 | 0.3211 | 0.1574 |
| 0.6094 | 6.0 | 432 | 0.2944 | 0.1530 |
| 0.5586 | 7.0 | 504 | 0.3446 | 0.1809 |
| 0.5728 | 8.0 | 576 | 0.2682 | 0.1246 |
| 0.575 | 9.0 | 648 | 0.2610 | 0.1244 |
| 0.5882 | 10.0 | 720 | 0.2463 | 0.1199 |
| 0.5367 | 11.0 | 792 | 0.2542 | 0.1093 |
| 0.5261 | 12.0 | 864 | 0.2523 | 0.1131 |
| 0.5091 | 13.0 | 936 | 0.2491 | 0.1175 |
| 0.5323 | 14.0 | 1008 | 0.2535 | 0.1208 |
| 0.5478 | 15.0 | 1080 | 0.2542 | 0.1204 |
| 0.4775 | 16.0 | 1152 | 0.2476 | 0.1138 |
| 0.4922 | 17.0 | 1224 | 0.2535 | 0.1162 |
| 0.4893 | 18.0 | 1296 | 0.2501 | 0.1143 |
| 0.4672 | 19.0 | 1368 | 0.2703 | 0.1162 |
| 0.4764 | 20.0 | 1440 | 0.2632 | 0.1253 |
| 0.4716 | 21.0 | 1512 | 0.2534 | 0.1097 |
| 0.4733 | 22.0 | 1584 | 0.2496 | 0.1086 |
| 0.4577 | 23.0 | 1656 | 0.2637 | 0.1225 |
| 0.4714 | 24.0 | 1728 | 0.2489 | 0.1102 |
| 0.4615 | 25.0 | 1800 | 0.2399 | 0.1020 |
| 0.4636 | 26.0 | 1872 | 0.2311 | 0.1042 |
| 0.4564 | 27.0 | 1944 | 0.2413 | 0.1042 |
| 0.4452 | 28.0 | 2016 | 0.2436 | 0.1037 |
| 0.4336 | 29.0 | 2088 | 0.2484 | 0.1070 |
| 0.4628 | 30.0 | 2160 | 0.2385 | 0.1005 |
| 0.4475 | 31.0 | 2232 | 0.2446 | 0.1075 |
| 0.4264 | 32.0 | 2304 | 0.2548 | 0.1068 |
| 0.4417 | 33.0 | 2376 | 0.2442 | 0.1046 |
| 0.4165 | 34.0 | 2448 | 0.2458 | 0.1045 |
| 0.4398 | 35.0 | 2520 | 0.2475 | 0.0979 |
| 0.4334 | 36.0 | 2592 | 0.2375 | 0.1027 |
| 0.4279 | 37.0 | 2664 | 0.2462 | 0.1053 |
| 0.4213 | 38.0 | 2736 | 0.2402 | 0.1028 |
| 0.4394 | 39.0 | 2808 | 0.2385 | 0.1020 |
| 0.4415 | 40.0 | 2880 | 0.2428 | 0.1021 |
| 0.4173 | 41.0 | 2952 | 0.2356 | 0.1016 |
| 0.4006 | 42.0 | 3024 | 0.2461 | 0.1007 |
| 0.4055 | 43.0 | 3096 | 0.2412 | 0.0998 |
| 0.4163 | 44.0 | 3168 | 0.2378 | 0.0976 |
| 0.4143 | 45.0 | 3240 | 0.2471 | 0.0996 |
| 0.4132 | 46.0 | 3312 | 0.2457 | 0.1004 |
| 0.3991 | 47.0 | 3384 | 0.2350 | 0.1019 |
| 0.4014 | 48.0 | 3456 | 0.2400 | 0.1025 |
| 0.416 | 49.0 | 3528 | 0.2370 | 0.1010 |
| 0.4067 | 50.0 | 3600 | 0.2444 | 0.1010 |
| 0.3876 | 51.0 | 3672 | 0.2491 | 0.1057 |
| 0.3964 | 52.0 | 3744 | 0.2451 | 0.1075 |
| 0.3903 | 53.0 | 3816 | 0.2395 | 0.1003 |
| 0.4036 | 54.0 | 3888 | 0.2446 | 0.1016 |
| 0.3936 | 55.0 | 3960 | 0.2520 | 0.0997 |
| 0.4094 | 56.0 | 4032 | 0.2401 | 0.0992 |
| 0.3977 | 57.0 | 4104 | 0.2498 | 0.1019 |
| 0.3942 | 58.0 | 4176 | 0.2496 | 0.0989 |
| 0.4052 | 59.0 | 4248 | 0.2507 | 0.1021 |
| 0.3995 | 60.0 | 4320 | 0.2382 | 0.0999 |
| 0.407 | 61.0 | 4392 | 0.2517 | 0.1037 |
| 0.4067 | 62.0 | 4464 | 0.2430 | 0.1034 |
| 0.3887 | 63.0 | 4536 | 0.2415 | 0.0974 |
| 0.3837 | 64.0 | 4608 | 0.2435 | 0.0991 |
| 0.3954 | 65.0 | 4680 | 0.2384 | 0.0985 |
| 0.3726 | 66.0 | 4752 | 0.2550 | 0.1009 |
| 0.3659 | 67.0 | 4824 | 0.2523 | 0.0967 |
| 0.376 | 68.0 | 4896 | 0.2571 | 0.0973 |
| 0.3759 | 69.0 | 4968 | 0.2528 | 0.0981 |
| 0.3862 | 70.0 | 5040 | 0.2496 | 0.0976 |
| 0.367 | 71.0 | 5112 | 0.2465 | 0.0942 |
| 0.3688 | 72.0 | 5184 | 0.2505 | 0.0968 |
| 0.3817 | 73.0 | 5256 | 0.2525 | 0.0973 |
| 0.3675 | 74.0 | 5328 | 0.2441 | 0.0964 |
| 0.3727 | 75.0 | 5400 | 0.2440 | 0.0973 |
| 0.371 | 76.0 | 5472 | 0.2510 | 0.0971 |
| 0.3761 | 77.0 | 5544 | 0.2398 | 0.0961 |
| 0.358 | 78.0 | 5616 | 0.2485 | 0.0956 |
| 0.3521 | 79.0 | 5688 | 0.2438 | 0.0955 |
| 0.3722 | 80.0 | 5760 | 0.2422 | 0.0967 |
| 0.3663 | 81.0 | 5832 | 0.2463 | 0.0949 |
| 0.3716 | 82.0 | 5904 | 0.2467 | 0.0965 |
| 0.361 | 83.0 | 5976 | 0.2458 | 0.0957 |
| 0.3626 | 84.0 | 6048 | 0.2501 | 0.0942 |
| 0.3755 | 85.0 | 6120 | 0.2486 | 0.0943 |
| 0.3693 | 86.0 | 6192 | 0.2497 | 0.0947 |
| 0.3548 | 87.0 | 6264 | 0.2515 | 0.0958 |
| 0.3583 | 88.0 | 6336 | 0.2433 | 0.0937 |
| 0.3678 | 89.0 | 6408 | 0.2456 | 0.0933 |
| 0.3718 | 90.0 | 6480 | 0.2443 | 0.0937 |
| 0.3614 | 91.0 | 6552 | 0.2457 | 0.0957 |
| 0.3541 | 92.0 | 6624 | 0.2450 | 0.0953 |
| 0.3671 | 93.0 | 6696 | 0.2469 | 0.0944 |
| 0.3503 | 94.0 | 6768 | 0.2459 | 0.0956 |
| 0.3692 | 95.0 | 6840 | 0.2461 | 0.0944 |
| 0.362 | 96.0 | 6912 | 0.2430 | 0.0945 |
| 0.3431 | 97.0 | 6984 | 0.2454 | 0.0952 |
| 0.3597 | 98.0 | 7056 | 0.2454 | 0.0940 |
| 0.354 | 99.0 | 7128 | 0.2448 | 0.0939 |
| 0.3597 | 100.0 | 7200 | 0.2449 | 0.0943 |
### Framework versions
- Transformers 4.21.0.dev0
- Pytorch 1.9.1+cu102
- Datasets 2.3.3.dev0
- Tokenizers 0.12.1
|
sd-concepts-library/rcrumb-portraits-style | sd-concepts-library | 2022-09-19T00:13:10Z | 0 | 4 | null | [
"license:mit",
"region:us"
]
| null | 2022-09-19T00:12:56Z | ---
license: mit
---
### rcrumb portraits style on Stable Diffusion
This is the `<rcrumb-portraits>` concept taught to Stable Diffusion via Textual Inversion. You can load this concept into the [Stable Conceptualizer](https://colab.research.google.com/github/huggingface/notebooks/blob/main/diffusers/stable_conceptualizer_inference.ipynb) notebook. You can also train your own concepts and load them into the concept libraries using [this notebook](https://colab.research.google.com/github/huggingface/notebooks/blob/main/diffusers/sd_textual_inversion_training.ipynb).
Here is the new concept you will be able to use as a `style`:
|
Naimul/banglabert-finetuned-squad | Naimul | 2022-09-18T23:07:28Z | 103 | 0 | transformers | [
"transformers",
"pytorch",
"tensorboard",
"electra",
"question-answering",
"generated_from_trainer",
"dataset:squad_bn",
"endpoints_compatible",
"region:us"
]
| question-answering | 2022-09-18T21:07:39Z | ---
tags:
- generated_from_trainer
datasets:
- squad_bn
model-index:
- name: banglabert-finetuned-squad
results: []
---
<!-- This model card has been generated automatically according to the information the Trainer had access to. You
should probably proofread and complete it, then remove this comment. -->
# banglabert-finetuned-squad
This model is a fine-tuned version of [csebuetnlp/banglabert](https://huggingface.co/csebuetnlp/banglabert) on the squad_bn dataset.
It achieves the following results on the evaluation set:
- Loss: 1.4421
## Model description
More information needed
## Intended uses & limitations
More information needed
## Training and evaluation data
More information needed
## Training procedure
### Training hyperparameters
The following hyperparameters were used during training:
- learning_rate: 2e-05
- train_batch_size: 16
- eval_batch_size: 16
- seed: 42
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
- lr_scheduler_type: linear
- num_epochs: 1
### Training results
| Training Loss | Epoch | Step | Validation Loss |
|:-------------:|:-----:|:----:|:---------------:|
| 1.3649 | 1.0 | 7397 | 1.4421 |
### Framework versions
- Transformers 4.20.1
- Pytorch 1.11.0
- Datasets 2.1.0
- Tokenizers 0.12.1
|
theojolliffe/model2-bart-reverse | theojolliffe | 2022-09-18T23:04:39Z | 106 | 0 | transformers | [
"transformers",
"pytorch",
"tensorboard",
"bart",
"text2text-generation",
"generated_from_trainer",
"license:apache-2.0",
"autotrain_compatible",
"endpoints_compatible",
"region:us"
]
| text2text-generation | 2022-09-18T22:15:33Z | ---
license: apache-2.0
tags:
- generated_from_trainer
metrics:
- rouge
model-index:
- name: model2-bart-reverse
results: []
---
<!-- This model card has been generated automatically according to the information the Trainer had access to. You
should probably proofread and complete it, then remove this comment. -->
# model2-bart-reverse
This model is a fine-tuned version of [theojolliffe/model-2-bart-reverse-raw](https://huggingface.co/theojolliffe/model-2-bart-reverse-raw) on the None dataset.
It achieves the following results on the evaluation set:
- Loss: 0.5768
- Rouge1: 50.0024
- Rouge2: 44.5149
- Rougel: 50.408
- Rougelsum: 50.0015
- Gen Len: 20.0
## Model description
More information needed
## Intended uses & limitations
More information needed
## Training and evaluation data
More information needed
## Training procedure
### Training hyperparameters
The following hyperparameters were used during training:
- learning_rate: 2e-05
- train_batch_size: 2
- eval_batch_size: 2
- seed: 42
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
- lr_scheduler_type: linear
- num_epochs: 4
- mixed_precision_training: Native AMP
### Training results
| Training Loss | Epoch | Step | Validation Loss | Rouge1 | Rouge2 | Rougel | Rougelsum | Gen Len |
|:-------------:|:-----:|:----:|:---------------:|:-------:|:-------:|:-------:|:---------:|:-------:|
| No log | 1.0 | 160 | 0.5107 | 54.2246 | 50.7271 | 51.1954 | 54.4944 | 20.0 |
| No log | 2.0 | 320 | 0.5895 | 48.1317 | 43.4207 | 48.6594 | 48.4308 | 20.0 |
| No log | 3.0 | 480 | 0.5833 | 51.7747 | 46.8312 | 52.47 | 52.1239 | 20.0 |
| 0.4286 | 4.0 | 640 | 0.5768 | 50.0024 | 44.5149 | 50.408 | 50.0015 | 20.0 |
### Framework versions
- Transformers 4.22.1
- Pytorch 1.12.1+cu113
- Datasets 2.4.0
- Tokenizers 0.12.1
|
sd-concepts-library/vietstoneking | sd-concepts-library | 2022-09-18T22:54:46Z | 0 | 0 | null | [
"license:mit",
"region:us"
]
| null | 2022-09-18T22:54:32Z | ---
license: mit
---
### vietstoneking on Stable Diffusion
This is the `<vietstoneking>` concept taught to Stable Diffusion via Textual Inversion. You can load this concept into the [Stable Conceptualizer](https://colab.research.google.com/github/huggingface/notebooks/blob/main/diffusers/stable_conceptualizer_inference.ipynb) notebook. You can also train your own concepts and load them into the concept libraries using [this notebook](https://colab.research.google.com/github/huggingface/notebooks/blob/main/diffusers/sd_textual_inversion_training.ipynb).
Here is the new concept you will be able to use as an `object`:
|
Ravneet/ddpm-butterflies-128 | Ravneet | 2022-09-18T20:52:17Z | 4 | 0 | diffusers | [
"diffusers",
"tensorboard",
"en",
"dataset:huggan/smithsonian_butterflies_subset",
"license:apache-2.0",
"diffusers:DDPMPipeline",
"region:us"
]
| null | 2022-09-18T19:37:59Z | ---
language: en
license: apache-2.0
library_name: diffusers
tags: []
datasets: huggan/smithsonian_butterflies_subset
metrics: []
---
<!-- This model card has been generated automatically according to the information the training script had access to. You
should probably proofread and complete it, then remove this comment. -->
# ddpm-butterflies-128
## Model description
This diffusion model is trained with the [🤗 Diffusers](https://github.com/huggingface/diffusers) library
on the `huggan/smithsonian_butterflies_subset` dataset.
## Intended uses & limitations
#### How to use
```python
# TODO: add an example code snippet for running this diffusion pipeline
```
#### Limitations and bias
[TODO: provide examples of latent issues and potential remediations]
## Training data
[TODO: describe the data used to train the model]
### Training hyperparameters
The following hyperparameters were used during training:
- learning_rate: 0.0001
- train_batch_size: 16
- eval_batch_size: 16
- gradient_accumulation_steps: 1
- optimizer: AdamW with betas=(None, None), weight_decay=None and epsilon=None
- lr_scheduler: None
- lr_warmup_steps: 500
- ema_inv_gamma: None
- ema_inv_gamma: None
- ema_inv_gamma: None
- mixed_precision: fp16
### Training results
📈 [TensorBoard logs](https://huggingface.co/Ravneet/ddpm-butterflies-128/tensorboard?#scalars)
|
DrishtiSharma/LayoutLMv3-Finetuned-CORD_100 | DrishtiSharma | 2022-09-18T19:38:50Z | 83 | 0 | transformers | [
"transformers",
"pytorch",
"tensorboard",
"layoutlmv3",
"token-classification",
"generated_from_trainer",
"dataset:cord-layoutlmv3",
"license:cc-by-nc-sa-4.0",
"model-index",
"autotrain_compatible",
"endpoints_compatible",
"region:us"
]
| token-classification | 2022-09-18T18:35:30Z | ---
license: cc-by-nc-sa-4.0
tags:
- generated_from_trainer
datasets:
- cord-layoutlmv3
metrics:
- precision
- recall
- f1
- accuracy
model-index:
- name: LayoutLMv3-Finetuned-CORD_100
results:
- task:
name: Token Classification
type: token-classification
dataset:
name: cord-layoutlmv3
type: cord-layoutlmv3
config: cord
split: train
args: cord
metrics:
- name: Precision
type: precision
value: 0.9524870081662955
- name: Recall
type: recall
value: 0.9603293413173652
- name: F1
type: f1
value: 0.9563920983973164
- name: Accuracy
type: accuracy
value: 0.9647707979626485
---
<!-- This model card has been generated automatically according to the information the Trainer had access to. You
should probably proofread and complete it, then remove this comment. -->
# LayoutLMv3-Finetuned-CORD_100
This model is a fine-tuned version of [microsoft/layoutlmv3-base](https://huggingface.co/microsoft/layoutlmv3-base) on the cord-layoutlmv3 dataset.
It achieves the following results on the evaluation set:
- Loss: 0.1948
- Precision: 0.9525
- Recall: 0.9603
- F1: 0.9564
- Accuracy: 0.9648
## Model description
More information needed
## Intended uses & limitations
More information needed
## Training and evaluation data
More information needed
## Training procedure
### Training hyperparameters
The following hyperparameters were used during training:
- learning_rate: 1.1e-05
- train_batch_size: 5
- eval_batch_size: 5
- seed: 42
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
- lr_scheduler_type: linear
- training_steps: 3000
### Training results
| Training Loss | Epoch | Step | Validation Loss | Precision | Recall | F1 | Accuracy |
|:-------------:|:-----:|:----:|:---------------:|:---------:|:------:|:------:|:--------:|
| No log | 1.56 | 250 | 0.9568 | 0.7298 | 0.7844 | 0.7561 | 0.7992 |
| 1.3271 | 3.12 | 500 | 0.5239 | 0.8398 | 0.8713 | 0.8553 | 0.8858 |
| 1.3271 | 4.69 | 750 | 0.3586 | 0.8945 | 0.9207 | 0.9074 | 0.9300 |
| 0.3495 | 6.25 | 1000 | 0.2716 | 0.9298 | 0.9416 | 0.9357 | 0.9410 |
| 0.3495 | 7.81 | 1250 | 0.2331 | 0.9198 | 0.9356 | 0.9276 | 0.9474 |
| 0.1725 | 9.38 | 1500 | 0.2134 | 0.9379 | 0.9499 | 0.9438 | 0.9529 |
| 0.1725 | 10.94 | 1750 | 0.2079 | 0.9401 | 0.9513 | 0.9457 | 0.9605 |
| 0.1116 | 12.5 | 2000 | 0.1992 | 0.9554 | 0.9618 | 0.9586 | 0.9656 |
| 0.1116 | 14.06 | 2250 | 0.1941 | 0.9517 | 0.9588 | 0.9553 | 0.9631 |
| 0.0762 | 15.62 | 2500 | 0.1966 | 0.9503 | 0.9588 | 0.9545 | 0.9639 |
| 0.0762 | 17.19 | 2750 | 0.1951 | 0.9510 | 0.9588 | 0.9549 | 0.9626 |
| 0.0636 | 18.75 | 3000 | 0.1948 | 0.9525 | 0.9603 | 0.9564 | 0.9648 |
### Framework versions
- Transformers 4.22.1
- Pytorch 1.12.1+cu113
- Datasets 2.4.0
- Tokenizers 0.12.1
|
theodotus/DialoGPT-uk | theodotus | 2022-09-18T19:08:01Z | 0 | 1 | generic | [
"generic",
"text2text-generation",
"uk",
"license:mit",
"region:us"
]
| text2text-generation | 2022-09-18T13:55:21Z | ---
language:
- uk
tags:
- text2text-generation
library_name: generic
license: mit
---
# Attribution
OPT-175B is licensed under the [OPT-175B license](https://github.com/facebookresearch/metaseq/blob/main/projects/OPT/MODEL_LICENSE.md), Copyright (c) Meta Platforms, Inc. All Rights Reserved. |
Subsets and Splits