Search is not available for this dataset
pipeline_tag
stringclasses 48
values | library_name
stringclasses 205
values | text
stringlengths 0
18.3M
| metadata
stringlengths 2
1.07B
| id
stringlengths 5
122
| last_modified
null | tags
listlengths 1
1.84k
| sha
null | created_at
stringlengths 25
25
|
|---|---|---|---|---|---|---|---|---|
token-classification
|
transformers
|
{}
|
ckauth/ck-ner-subgroup
| null |
[
"transformers",
"tf",
"bert",
"token-classification",
"autotrain_compatible",
"endpoints_compatible",
"region:us"
] | null |
2022-03-02T23:29:05+00:00
|
|
fill-mask
|
transformers
|
<!-- This model card has been generated automatically according to the information Keras had access to. You should
probably proofread and complete it, then remove this comment. -->
# nlu_sherlock_model
This model is a fine-tuned version of [roberta-base](https://huggingface.co/roberta-base) on an unknown dataset.
It achieves the following results on the evaluation set:
## Model description
More information needed
## Intended uses & limitations
More information needed
## Training and evaluation data
More information needed
## Training procedure
### Training hyperparameters
The following hyperparameters were used during training:
- optimizer: {'name': 'AdamWeightDecay', 'learning_rate': {'class_name': 'WarmUp', 'config': {'initial_learning_rate': 2e-05, 'decay_schedule_fn': {'class_name': 'PolynomialDecay', 'config': {'initial_learning_rate': 2e-05, 'decay_steps': -947, 'end_learning_rate': 0.0, 'power': 1.0, 'cycle': False, 'name': None}, '__passive_serialization__': True}, 'warmup_steps': 1000, 'power': 1.0, 'name': None}}, 'decay': 0.0, 'beta_1': 0.9, 'beta_2': 0.999, 'epsilon': 1e-08, 'amsgrad': False, 'weight_decay_rate': 0.01}
- training_precision: float32
### Training results
### Framework versions
- Transformers 4.16.2
- TensorFlow 2.8.0
- Datasets 1.18.3
- Tokenizers 0.11.0
|
{"license": "mit", "tags": ["generated_from_keras_callback"], "model-index": [{"name": "nlu_sherlock_model", "results": []}]}
|
ckenlam/nlu_sherlock_model
| null |
[
"transformers",
"tf",
"roberta",
"fill-mask",
"generated_from_keras_callback",
"license:mit",
"autotrain_compatible",
"endpoints_compatible",
"region:us"
] | null |
2022-03-02T23:29:05+00:00
|
fill-mask
|
transformers
|
<!-- This model card has been generated automatically according to the information Keras had access to. You should
probably proofread and complete it, then remove this comment. -->
# nlu_sherlock_model_20220220
This model is a fine-tuned version of [roberta-base](https://huggingface.co/roberta-base) on an unknown dataset.
It achieves the following results on the evaluation set:
## Model description
More information needed
## Intended uses & limitations
More information needed
## Training and evaluation data
More information needed
## Training procedure
### Training hyperparameters
The following hyperparameters were used during training:
- optimizer: {'name': 'AdamWeightDecay', 'learning_rate': {'class_name': 'WarmUp', 'config': {'initial_learning_rate': 2e-05, 'decay_schedule_fn': {'class_name': 'PolynomialDecay', 'config': {'initial_learning_rate': 2e-05, 'decay_steps': -955, 'end_learning_rate': 0.0, 'power': 1.0, 'cycle': False, 'name': None}, '__passive_serialization__': True}, 'warmup_steps': 1000, 'power': 1.0, 'name': None}}, 'decay': 0.0, 'beta_1': 0.9, 'beta_2': 0.999, 'epsilon': 1e-08, 'amsgrad': False, 'weight_decay_rate': 0.01}
- training_precision: float32
### Training results
### Framework versions
- Transformers 4.16.2
- TensorFlow 2.8.0
- Datasets 1.18.3
- Tokenizers 0.11.0
|
{"license": "mit", "tags": ["generated_from_keras_callback"], "model-index": [{"name": "nlu_sherlock_model_20220220", "results": []}]}
|
ckenlam/nlu_sherlock_model_20220220
| null |
[
"transformers",
"tf",
"roberta",
"fill-mask",
"generated_from_keras_callback",
"license:mit",
"autotrain_compatible",
"endpoints_compatible",
"region:us"
] | null |
2022-03-02T23:29:05+00:00
|
token-classification
|
transformers
|
# CKIP ALBERT Base Chinese
This project provides traditional Chinese transformers models (including ALBERT, BERT, GPT2) and NLP tools (including word segmentation, part-of-speech tagging, named entity recognition).
這個專案提供了繁體中文的 transformers 模型（包含 ALBERT、BERT、GPT2）及自然語言處理工具（包含斷詞、詞性標記、實體辨識）。
## Homepage
- https://github.com/ckiplab/ckip-transformers
## Contributers
- [Mu Yang](https://muyang.pro) at [CKIP](https://ckip.iis.sinica.edu.tw) (Author & Maintainer)
## Usage
Please use BertTokenizerFast as tokenizer instead of AutoTokenizer.
請使用 BertTokenizerFast 而非 AutoTokenizer。
```
from transformers import (
BertTokenizerFast,
AutoModel,
)
tokenizer = BertTokenizerFast.from_pretrained('bert-base-chinese')
model = AutoModel.from_pretrained('ckiplab/albert-base-chinese-ner')
```
For full usage and more information, please refer to https://github.com/ckiplab/ckip-transformers.
有關完整使用方法及其他資訊，請參見 https://github.com/ckiplab/ckip-transformers 。
|
{"language": ["zh"], "license": "gpl-3.0", "tags": ["pytorch", "token-classification", "albert", "zh"], "thumbnail": "https://ckip.iis.sinica.edu.tw/files/ckip_logo.png"}
|
ckiplab/albert-base-chinese-ner
| null |
[
"transformers",
"pytorch",
"albert",
"token-classification",
"zh",
"license:gpl-3.0",
"autotrain_compatible",
"endpoints_compatible",
"region:us"
] | null |
2022-03-02T23:29:05+00:00
|
token-classification
|
transformers
|
# CKIP ALBERT Base Chinese
This project provides traditional Chinese transformers models (including ALBERT, BERT, GPT2) and NLP tools (including word segmentation, part-of-speech tagging, named entity recognition).
這個專案提供了繁體中文的 transformers 模型（包含 ALBERT、BERT、GPT2）及自然語言處理工具（包含斷詞、詞性標記、實體辨識）。
## Homepage
- https://github.com/ckiplab/ckip-transformers
## Contributers
- [Mu Yang](https://muyang.pro) at [CKIP](https://ckip.iis.sinica.edu.tw) (Author & Maintainer)
## Usage
Please use BertTokenizerFast as tokenizer instead of AutoTokenizer.
請使用 BertTokenizerFast 而非 AutoTokenizer。
```
from transformers import (
BertTokenizerFast,
AutoModel,
)
tokenizer = BertTokenizerFast.from_pretrained('bert-base-chinese')
model = AutoModel.from_pretrained('ckiplab/albert-base-chinese-pos')
```
For full usage and more information, please refer to https://github.com/ckiplab/ckip-transformers.
有關完整使用方法及其他資訊，請參見 https://github.com/ckiplab/ckip-transformers 。
|
{"language": ["zh"], "license": "gpl-3.0", "tags": ["pytorch", "token-classification", "albert", "zh"], "thumbnail": "https://ckip.iis.sinica.edu.tw/files/ckip_logo.png"}
|
ckiplab/albert-base-chinese-pos
| null |
[
"transformers",
"pytorch",
"albert",
"token-classification",
"zh",
"license:gpl-3.0",
"autotrain_compatible",
"endpoints_compatible",
"region:us"
] | null |
2022-03-02T23:29:05+00:00
|
token-classification
|
transformers
|
# CKIP ALBERT Base Chinese
This project provides traditional Chinese transformers models (including ALBERT, BERT, GPT2) and NLP tools (including word segmentation, part-of-speech tagging, named entity recognition).
這個專案提供了繁體中文的 transformers 模型（包含 ALBERT、BERT、GPT2）及自然語言處理工具（包含斷詞、詞性標記、實體辨識）。
## Homepage
- https://github.com/ckiplab/ckip-transformers
## Contributers
- [Mu Yang](https://muyang.pro) at [CKIP](https://ckip.iis.sinica.edu.tw) (Author & Maintainer)
## Usage
Please use BertTokenizerFast as tokenizer instead of AutoTokenizer.
請使用 BertTokenizerFast 而非 AutoTokenizer。
```
from transformers import (
BertTokenizerFast,
AutoModel,
)
tokenizer = BertTokenizerFast.from_pretrained('bert-base-chinese')
model = AutoModel.from_pretrained('ckiplab/albert-base-chinese-ws')
```
For full usage and more information, please refer to https://github.com/ckiplab/ckip-transformers.
有關完整使用方法及其他資訊，請參見 https://github.com/ckiplab/ckip-transformers 。
|
{"language": ["zh"], "license": "gpl-3.0", "tags": ["pytorch", "token-classification", "albert", "zh"], "thumbnail": "https://ckip.iis.sinica.edu.tw/files/ckip_logo.png"}
|
ckiplab/albert-base-chinese-ws
| null |
[
"transformers",
"pytorch",
"albert",
"token-classification",
"zh",
"license:gpl-3.0",
"autotrain_compatible",
"endpoints_compatible",
"region:us"
] | null |
2022-03-02T23:29:05+00:00
|
fill-mask
|
transformers
|
# CKIP ALBERT Base Chinese
This project provides traditional Chinese transformers models (including ALBERT, BERT, GPT2) and NLP tools (including word segmentation, part-of-speech tagging, named entity recognition).
這個專案提供了繁體中文的 transformers 模型（包含 ALBERT、BERT、GPT2）及自然語言處理工具（包含斷詞、詞性標記、實體辨識）。
## Homepage
- https://github.com/ckiplab/ckip-transformers
## Contributers
- [Mu Yang](https://muyang.pro) at [CKIP](https://ckip.iis.sinica.edu.tw) (Author & Maintainer)
## Usage
Please use BertTokenizerFast as tokenizer instead of AutoTokenizer.
請使用 BertTokenizerFast 而非 AutoTokenizer。
```
from transformers import (
BertTokenizerFast,
AutoModel,
)
tokenizer = BertTokenizerFast.from_pretrained('bert-base-chinese')
model = AutoModel.from_pretrained('ckiplab/albert-base-chinese')
```
For full usage and more information, please refer to https://github.com/ckiplab/ckip-transformers.
有關完整使用方法及其他資訊，請參見 https://github.com/ckiplab/ckip-transformers 。
|
{"language": ["zh"], "license": "gpl-3.0", "tags": ["pytorch", "lm-head", "albert", "zh"], "thumbnail": "https://ckip.iis.sinica.edu.tw/files/ckip_logo.png"}
|
ckiplab/albert-base-chinese
| null |
[
"transformers",
"pytorch",
"albert",
"fill-mask",
"lm-head",
"zh",
"license:gpl-3.0",
"autotrain_compatible",
"endpoints_compatible",
"region:us"
] | null |
2022-03-02T23:29:05+00:00
|
token-classification
|
transformers
|
# CKIP ALBERT Tiny Chinese
This project provides traditional Chinese transformers models (including ALBERT, BERT, GPT2) and NLP tools (including word segmentation, part-of-speech tagging, named entity recognition).
這個專案提供了繁體中文的 transformers 模型（包含 ALBERT、BERT、GPT2）及自然語言處理工具（包含斷詞、詞性標記、實體辨識）。
## Homepage
- https://github.com/ckiplab/ckip-transformers
## Contributers
- [Mu Yang](https://muyang.pro) at [CKIP](https://ckip.iis.sinica.edu.tw) (Author & Maintainer)
## Usage
Please use BertTokenizerFast as tokenizer instead of AutoTokenizer.
請使用 BertTokenizerFast 而非 AutoTokenizer。
```
from transformers import (
BertTokenizerFast,
AutoModel,
)
tokenizer = BertTokenizerFast.from_pretrained('bert-base-chinese')
model = AutoModel.from_pretrained('ckiplab/albert-tiny-chinese-ner')
```
For full usage and more information, please refer to https://github.com/ckiplab/ckip-transformers.
有關完整使用方法及其他資訊，請參見 https://github.com/ckiplab/ckip-transformers 。
|
{"language": ["zh"], "license": "gpl-3.0", "tags": ["pytorch", "token-classification", "albert", "zh"], "thumbnail": "https://ckip.iis.sinica.edu.tw/files/ckip_logo.png"}
|
ckiplab/albert-tiny-chinese-ner
| null |
[
"transformers",
"pytorch",
"albert",
"token-classification",
"zh",
"license:gpl-3.0",
"autotrain_compatible",
"endpoints_compatible",
"region:us"
] | null |
2022-03-02T23:29:05+00:00
|
token-classification
|
transformers
|
# CKIP ALBERT Tiny Chinese
This project provides traditional Chinese transformers models (including ALBERT, BERT, GPT2) and NLP tools (including word segmentation, part-of-speech tagging, named entity recognition).
這個專案提供了繁體中文的 transformers 模型（包含 ALBERT、BERT、GPT2）及自然語言處理工具（包含斷詞、詞性標記、實體辨識）。
## Homepage
- https://github.com/ckiplab/ckip-transformers
## Contributers
- [Mu Yang](https://muyang.pro) at [CKIP](https://ckip.iis.sinica.edu.tw) (Author & Maintainer)
## Usage
Please use BertTokenizerFast as tokenizer instead of AutoTokenizer.
請使用 BertTokenizerFast 而非 AutoTokenizer。
```
from transformers import (
BertTokenizerFast,
AutoModel,
)
tokenizer = BertTokenizerFast.from_pretrained('bert-base-chinese')
model = AutoModel.from_pretrained('ckiplab/albert-tiny-chinese-pos')
```
For full usage and more information, please refer to https://github.com/ckiplab/ckip-transformers.
有關完整使用方法及其他資訊，請參見 https://github.com/ckiplab/ckip-transformers 。
|
{"language": ["zh"], "license": "gpl-3.0", "tags": ["pytorch", "token-classification", "albert", "zh"], "thumbnail": "https://ckip.iis.sinica.edu.tw/files/ckip_logo.png"}
|
ckiplab/albert-tiny-chinese-pos
| null |
[
"transformers",
"pytorch",
"albert",
"token-classification",
"zh",
"license:gpl-3.0",
"autotrain_compatible",
"endpoints_compatible",
"region:us"
] | null |
2022-03-02T23:29:05+00:00
|
token-classification
|
transformers
|
# CKIP ALBERT Tiny Chinese
This project provides traditional Chinese transformers models (including ALBERT, BERT, GPT2) and NLP tools (including word segmentation, part-of-speech tagging, named entity recognition).
這個專案提供了繁體中文的 transformers 模型（包含 ALBERT、BERT、GPT2）及自然語言處理工具（包含斷詞、詞性標記、實體辨識）。
## Homepage
- https://github.com/ckiplab/ckip-transformers
## Contributers
- [Mu Yang](https://muyang.pro) at [CKIP](https://ckip.iis.sinica.edu.tw) (Author & Maintainer)
## Usage
Please use BertTokenizerFast as tokenizer instead of AutoTokenizer.
請使用 BertTokenizerFast 而非 AutoTokenizer。
```
from transformers import (
BertTokenizerFast,
AutoModel,
)
tokenizer = BertTokenizerFast.from_pretrained('bert-base-chinese')
model = AutoModel.from_pretrained('ckiplab/albert-tiny-chinese-ws')
```
For full usage and more information, please refer to https://github.com/ckiplab/ckip-transformers.
有關完整使用方法及其他資訊，請參見 https://github.com/ckiplab/ckip-transformers 。
|
{"language": ["zh"], "license": "gpl-3.0", "tags": ["pytorch", "token-classification", "albert", "zh"], "thumbnail": "https://ckip.iis.sinica.edu.tw/files/ckip_logo.png"}
|
ckiplab/albert-tiny-chinese-ws
| null |
[
"transformers",
"pytorch",
"albert",
"token-classification",
"zh",
"license:gpl-3.0",
"autotrain_compatible",
"endpoints_compatible",
"has_space",
"region:us"
] | null |
2022-03-02T23:29:05+00:00
|
fill-mask
|
transformers
|
# CKIP ALBERT Tiny Chinese
This project provides traditional Chinese transformers models (including ALBERT, BERT, GPT2) and NLP tools (including word segmentation, part-of-speech tagging, named entity recognition).
這個專案提供了繁體中文的 transformers 模型（包含 ALBERT、BERT、GPT2）及自然語言處理工具（包含斷詞、詞性標記、實體辨識）。
## Homepage
- https://github.com/ckiplab/ckip-transformers
## Contributers
- [Mu Yang](https://muyang.pro) at [CKIP](https://ckip.iis.sinica.edu.tw) (Author & Maintainer)
## Usage
Please use BertTokenizerFast as tokenizer instead of AutoTokenizer.
請使用 BertTokenizerFast 而非 AutoTokenizer。
```
from transformers import (
BertTokenizerFast,
AutoModel,
)
tokenizer = BertTokenizerFast.from_pretrained('bert-base-chinese')
model = AutoModel.from_pretrained('ckiplab/albert-tiny-chinese')
```
For full usage and more information, please refer to https://github.com/ckiplab/ckip-transformers.
有關完整使用方法及其他資訊，請參見 https://github.com/ckiplab/ckip-transformers 。
|
{"language": ["zh"], "license": "gpl-3.0", "tags": ["pytorch", "lm-head", "albert", "zh"], "thumbnail": "https://ckip.iis.sinica.edu.tw/files/ckip_logo.png"}
|
ckiplab/albert-tiny-chinese
| null |
[
"transformers",
"pytorch",
"albert",
"fill-mask",
"lm-head",
"zh",
"license:gpl-3.0",
"autotrain_compatible",
"endpoints_compatible",
"has_space",
"region:us"
] | null |
2022-03-02T23:29:05+00:00
|
token-classification
|
transformers
|
# CKIP BERT Base Chinese
This project provides traditional Chinese transformers models (including ALBERT, BERT, GPT2) and NLP tools (including word segmentation, part-of-speech tagging, named entity recognition).
這個專案提供了繁體中文的 transformers 模型（包含 ALBERT、BERT、GPT2）及自然語言處理工具（包含斷詞、詞性標記、實體辨識）。
## Homepage
- https://github.com/ckiplab/ckip-transformers
## Contributers
- [Mu Yang](https://muyang.pro) at [CKIP](https://ckip.iis.sinica.edu.tw) (Author & Maintainer)
## Usage
Please use BertTokenizerFast as tokenizer instead of AutoTokenizer.
請使用 BertTokenizerFast 而非 AutoTokenizer。
```
from transformers import (
BertTokenizerFast,
AutoModel,
)
tokenizer = BertTokenizerFast.from_pretrained('bert-base-chinese')
model = AutoModel.from_pretrained('ckiplab/bert-base-chinese-ner')
```
For full usage and more information, please refer to https://github.com/ckiplab/ckip-transformers.
有關完整使用方法及其他資訊，請參見 https://github.com/ckiplab/ckip-transformers 。
|
{"language": ["zh"], "license": "gpl-3.0", "tags": ["pytorch", "token-classification", "bert", "zh"], "thumbnail": "https://ckip.iis.sinica.edu.tw/files/ckip_logo.png"}
|
ckiplab/bert-base-chinese-ner
| null |
[
"transformers",
"pytorch",
"jax",
"bert",
"token-classification",
"zh",
"license:gpl-3.0",
"autotrain_compatible",
"endpoints_compatible",
"has_space",
"region:us"
] | null |
2022-03-02T23:29:05+00:00
|
token-classification
|
transformers
|
# CKIP BERT Base Chinese
This project provides traditional Chinese transformers models (including ALBERT, BERT, GPT2) and NLP tools (including word segmentation, part-of-speech tagging, named entity recognition).
這個專案提供了繁體中文的 transformers 模型（包含 ALBERT、BERT、GPT2）及自然語言處理工具（包含斷詞、詞性標記、實體辨識）。
## Homepage
- https://github.com/ckiplab/ckip-transformers
## Contributers
- [Mu Yang](https://muyang.pro) at [CKIP](https://ckip.iis.sinica.edu.tw) (Author & Maintainer)
## Usage
Please use BertTokenizerFast as tokenizer instead of AutoTokenizer.
請使用 BertTokenizerFast 而非 AutoTokenizer。
```
from transformers import (
BertTokenizerFast,
AutoModel,
)
tokenizer = BertTokenizerFast.from_pretrained('bert-base-chinese')
model = AutoModel.from_pretrained('ckiplab/bert-base-chinese-pos')
```
For full usage and more information, please refer to https://github.com/ckiplab/ckip-transformers.
有關完整使用方法及其他資訊，請參見 https://github.com/ckiplab/ckip-transformers 。
|
{"language": ["zh"], "license": "gpl-3.0", "tags": ["pytorch", "token-classification", "bert", "zh"], "thumbnail": "https://ckip.iis.sinica.edu.tw/files/ckip_logo.png"}
|
ckiplab/bert-base-chinese-pos
| null |
[
"transformers",
"pytorch",
"jax",
"bert",
"token-classification",
"zh",
"license:gpl-3.0",
"autotrain_compatible",
"endpoints_compatible",
"region:us"
] | null |
2022-03-02T23:29:05+00:00
|
token-classification
|
transformers
|
# CKIP BERT Base Chinese
This project provides traditional Chinese transformers models (including ALBERT, BERT, GPT2) and NLP tools (including word segmentation, part-of-speech tagging, named entity recognition).
這個專案提供了繁體中文的 transformers 模型（包含 ALBERT、BERT、GPT2）及自然語言處理工具（包含斷詞、詞性標記、實體辨識）。
## Homepage
- https://github.com/ckiplab/ckip-transformers
## Contributers
- [Mu Yang](https://muyang.pro) at [CKIP](https://ckip.iis.sinica.edu.tw) (Author & Maintainer)
## Usage
Please use BertTokenizerFast as tokenizer instead of AutoTokenizer.
請使用 BertTokenizerFast 而非 AutoTokenizer。
```
from transformers import (
BertTokenizerFast,
AutoModel,
)
tokenizer = BertTokenizerFast.from_pretrained('bert-base-chinese')
model = AutoModel.from_pretrained('ckiplab/bert-base-chinese-ws')
```
For full usage and more information, please refer to https://github.com/ckiplab/ckip-transformers.
有關完整使用方法及其他資訊，請參見 https://github.com/ckiplab/ckip-transformers 。
|
{"language": ["zh"], "license": "gpl-3.0", "tags": ["pytorch", "token-classification", "bert", "zh"], "thumbnail": "https://ckip.iis.sinica.edu.tw/files/ckip_logo.png"}
|
ckiplab/bert-base-chinese-ws
| null |
[
"transformers",
"pytorch",
"jax",
"bert",
"token-classification",
"zh",
"license:gpl-3.0",
"autotrain_compatible",
"endpoints_compatible",
"region:us"
] | null |
2022-03-02T23:29:05+00:00
|
fill-mask
|
transformers
|
# CKIP BERT Base Chinese
This project provides traditional Chinese transformers models (including ALBERT, BERT, GPT2) and NLP tools (including word segmentation, part-of-speech tagging, named entity recognition).
這個專案提供了繁體中文的 transformers 模型（包含 ALBERT、BERT、GPT2）及自然語言處理工具（包含斷詞、詞性標記、實體辨識）。
## Homepage
- https://github.com/ckiplab/ckip-transformers
## Contributers
- [Mu Yang](https://muyang.pro) at [CKIP](https://ckip.iis.sinica.edu.tw) (Author & Maintainer)
## Usage
Please use BertTokenizerFast as tokenizer instead of AutoTokenizer.
請使用 BertTokenizerFast 而非 AutoTokenizer。
```
from transformers import (
BertTokenizerFast,
AutoModel,
)
tokenizer = BertTokenizerFast.from_pretrained('bert-base-chinese')
model = AutoModel.from_pretrained('ckiplab/bert-base-chinese')
```
For full usage and more information, please refer to https://github.com/ckiplab/ckip-transformers.
有關完整使用方法及其他資訊，請參見 https://github.com/ckiplab/ckip-transformers 。
|
{"language": ["zh"], "license": "gpl-3.0", "tags": ["pytorch", "lm-head", "bert", "zh"], "thumbnail": "https://ckip.iis.sinica.edu.tw/files/ckip_logo.png"}
|
ckiplab/bert-base-chinese
| null |
[
"transformers",
"pytorch",
"jax",
"bert",
"fill-mask",
"lm-head",
"zh",
"license:gpl-3.0",
"autotrain_compatible",
"endpoints_compatible",
"has_space",
"region:us"
] | null |
2022-03-02T23:29:05+00:00
|
text-generation
|
transformers
|
# CKIP GPT2 Base Chinese
This project provides traditional Chinese transformers models (including ALBERT, BERT, GPT2) and NLP tools (including word segmentation, part-of-speech tagging, named entity recognition).
這個專案提供了繁體中文的 transformers 模型（包含 ALBERT、BERT、GPT2）及自然語言處理工具（包含斷詞、詞性標記、實體辨識）。
## Homepage
- https://github.com/ckiplab/ckip-transformers
## Contributers
- [Mu Yang](https://muyang.pro) at [CKIP](https://ckip.iis.sinica.edu.tw) (Author & Maintainer)
## Usage
Please use BertTokenizerFast as tokenizer instead of AutoTokenizer.
請使用 BertTokenizerFast 而非 AutoTokenizer。
```
from transformers import (
BertTokenizerFast,
AutoModel,
)
tokenizer = BertTokenizerFast.from_pretrained('bert-base-chinese')
model = AutoModel.from_pretrained('ckiplab/gpt2-base-chinese')
```
For full usage and more information, please refer to https://github.com/ckiplab/ckip-transformers.
有關完整使用方法及其他資訊，請參見 https://github.com/ckiplab/ckip-transformers 。
|
{"language": ["zh"], "license": "gpl-3.0", "tags": ["pytorch", "lm-head", "gpt2", "zh"], "thumbnail": "https://ckip.iis.sinica.edu.tw/files/ckip_logo.png"}
|
ckiplab/gpt2-base-chinese
| null |
[
"transformers",
"pytorch",
"jax",
"gpt2",
"text-generation",
"lm-head",
"zh",
"license:gpl-3.0",
"autotrain_compatible",
"endpoints_compatible",
"has_space",
"text-generation-inference",
"region:us"
] | null |
2022-03-02T23:29:05+00:00
|
fill-mask
|
transformers
|
{}
|
cl-nagoya/defsent-bert-base-uncased-cls
| null |
[
"transformers",
"pytorch",
"bert",
"fill-mask",
"autotrain_compatible",
"endpoints_compatible",
"region:us"
] | null |
2022-03-02T23:29:05+00:00
|
|
fill-mask
|
transformers
|
{}
|
cl-nagoya/defsent-bert-base-uncased-max
| null |
[
"transformers",
"pytorch",
"bert",
"fill-mask",
"autotrain_compatible",
"endpoints_compatible",
"region:us"
] | null |
2022-03-02T23:29:05+00:00
|
|
fill-mask
|
transformers
|
{}
|
cl-nagoya/defsent-bert-base-uncased-mean
| null |
[
"transformers",
"pytorch",
"bert",
"fill-mask",
"autotrain_compatible",
"endpoints_compatible",
"region:us"
] | null |
2022-03-02T23:29:05+00:00
|
|
fill-mask
|
transformers
|
{}
|
cl-nagoya/defsent-bert-large-uncased-cls
| null |
[
"transformers",
"pytorch",
"bert",
"fill-mask",
"autotrain_compatible",
"endpoints_compatible",
"region:us"
] | null |
2022-03-02T23:29:05+00:00
|
|
fill-mask
|
transformers
|
{}
|
cl-nagoya/defsent-bert-large-uncased-max
| null |
[
"transformers",
"pytorch",
"bert",
"fill-mask",
"autotrain_compatible",
"endpoints_compatible",
"region:us"
] | null |
2022-03-02T23:29:05+00:00
|
|
fill-mask
|
transformers
|
{}
|
cl-nagoya/defsent-bert-large-uncased-mean
| null |
[
"transformers",
"pytorch",
"bert",
"fill-mask",
"autotrain_compatible",
"endpoints_compatible",
"region:us"
] | null |
2022-03-02T23:29:05+00:00
|
|
fill-mask
|
transformers
|
{}
|
cl-nagoya/defsent-roberta-base-cls
| null |
[
"transformers",
"pytorch",
"roberta",
"fill-mask",
"autotrain_compatible",
"endpoints_compatible",
"region:us"
] | null |
2022-03-02T23:29:05+00:00
|
|
fill-mask
|
transformers
|
{}
|
cl-nagoya/defsent-roberta-base-max
| null |
[
"transformers",
"pytorch",
"roberta",
"fill-mask",
"autotrain_compatible",
"endpoints_compatible",
"region:us"
] | null |
2022-03-02T23:29:05+00:00
|
|
fill-mask
|
transformers
|
{}
|
cl-nagoya/defsent-roberta-base-mean
| null |
[
"transformers",
"pytorch",
"roberta",
"fill-mask",
"autotrain_compatible",
"endpoints_compatible",
"region:us"
] | null |
2022-03-02T23:29:05+00:00
|
|
fill-mask
|
transformers
|
{}
|
cl-nagoya/defsent-roberta-large-cls
| null |
[
"transformers",
"pytorch",
"roberta",
"fill-mask",
"autotrain_compatible",
"endpoints_compatible",
"region:us"
] | null |
2022-03-02T23:29:05+00:00
|
|
fill-mask
|
transformers
|
{}
|
cl-nagoya/defsent-roberta-large-max
| null |
[
"transformers",
"pytorch",
"roberta",
"fill-mask",
"autotrain_compatible",
"endpoints_compatible",
"region:us"
] | null |
2022-03-02T23:29:05+00:00
|
|
fill-mask
|
transformers
|
{}
|
cl-nagoya/defsent-roberta-large-mean
| null |
[
"transformers",
"pytorch",
"roberta",
"fill-mask",
"autotrain_compatible",
"endpoints_compatible",
"region:us"
] | null |
2022-03-02T23:29:05+00:00
|
|
fill-mask
|
transformers
|
# BERT base Japanese (character-level tokenization with whole word masking, jawiki-20200831)
This is a [BERT](https://github.com/google-research/bert) model pretrained on texts in the Japanese language.
This version of the model processes input texts with word-level tokenization based on the Unidic 2.1.2 dictionary (available in [unidic-lite](https://pypi.org/project/unidic-lite/) package), followed by character-level tokenization.
Additionally, the model is trained with the whole word masking enabled for the masked language modeling (MLM) objective.
The codes for the pretraining are available at [cl-tohoku/bert-japanese](https://github.com/cl-tohoku/bert-japanese/tree/v2.0).
## Model architecture
The model architecture is the same as the original BERT base model; 12 layers, 768 dimensions of hidden states, and 12 attention heads.
## Training Data
The models are trained on the Japanese version of Wikipedia.
The training corpus is generated from the Wikipedia Cirrussearch dump file as of August 31, 2020.
The generated corpus files are 4.0GB in total, containing approximately 30M sentences.
We used the [MeCab](https://taku910.github.io/mecab/) morphological parser with [mecab-ipadic-NEologd](https://github.com/neologd/mecab-ipadic-neologd) dictionary to split texts into sentences.
## Tokenization
The texts are first tokenized by MeCab with the Unidic 2.1.2 dictionary and then split into characters.
The vocabulary size is 6144.
We used [`fugashi`](https://github.com/polm/fugashi) and [`unidic-lite`](https://github.com/polm/unidic-lite) packages for the tokenization.
## Training
The models are trained with the same configuration as the original BERT; 512 tokens per instance, 256 instances per batch, and 1M training steps.
For training of the MLM (masked language modeling) objective, we introduced whole word masking in which all of the subword tokens corresponding to a single word (tokenized by MeCab) are masked at once.
For training of each model, we used a v3-8 instance of Cloud TPUs provided by [TensorFlow Research Cloud program](https://www.tensorflow.org/tfrc/).
The training took about 5 days to finish.
## Licenses
The pretrained models are distributed under the terms of the [Creative Commons Attribution-ShareAlike 3.0](https://creativecommons.org/licenses/by-sa/3.0/).
## Acknowledgments
This model is trained with Cloud TPUs provided by [TensorFlow Research Cloud](https://www.tensorflow.org/tfrc/) program.
|
{"language": "ja", "license": "cc-by-sa-4.0", "datasets": ["wikipedia"], "widget": [{"text": "\u6771\u5317\u5927\u5b66\u3067[MASK]\u306e\u7814\u7a76\u3092\u3057\u3066\u3044\u307e\u3059\u3002"}]}
|
tohoku-nlp/bert-base-japanese-char-v2
| null |
[
"transformers",
"pytorch",
"tf",
"jax",
"bert",
"fill-mask",
"ja",
"dataset:wikipedia",
"license:cc-by-sa-4.0",
"autotrain_compatible",
"endpoints_compatible",
"region:us"
] | null |
2022-03-02T23:29:05+00:00
|
fill-mask
|
transformers
|
# BERT base Japanese (character tokenization, whole word masking enabled)
This is a [BERT](https://github.com/google-research/bert) model pretrained on texts in the Japanese language.
This version of the model processes input texts with word-level tokenization based on the IPA dictionary, followed by character-level tokenization.
Additionally, the model is trained with the whole word masking enabled for the masked language modeling (MLM) objective.
The codes for the pretraining are available at [cl-tohoku/bert-japanese](https://github.com/cl-tohoku/bert-japanese/tree/v1.0).
## Model architecture
The model architecture is the same as the original BERT base model; 12 layers, 768 dimensions of hidden states, and 12 attention heads.
## Training Data
The model is trained on Japanese Wikipedia as of September 1, 2019.
To generate the training corpus, [WikiExtractor](https://github.com/attardi/wikiextractor) is used to extract plain texts from a dump file of Wikipedia articles.
The text files used for the training are 2.6GB in size, consisting of approximately 17M sentences.
## Tokenization
The texts are first tokenized by [MeCab](https://taku910.github.io/mecab/) morphological parser with the IPA dictionary and then split into characters.
The vocabulary size is 4000.
## Training
The model is trained with the same configuration as the original BERT; 512 tokens per instance, 256 instances per batch, and 1M training steps.
For the training of the MLM (masked language modeling) objective, we introduced the **Whole Word Masking** in which all of the subword tokens corresponding to a single word (tokenized by MeCab) are masked at once.
## Licenses
The pretrained models are distributed under the terms of the [Creative Commons Attribution-ShareAlike 3.0](https://creativecommons.org/licenses/by-sa/3.0/).
## Acknowledgments
For training models, we used Cloud TPUs provided by [TensorFlow Research Cloud](https://www.tensorflow.org/tfrc/) program.
|
{"language": "ja", "license": "cc-by-sa-4.0", "datasets": ["wikipedia"], "widget": [{"text": "\u4ed9\u53f0\u306f\u300c[MASK]\u306e\u90fd\u300d\u3068\u547c\u3070\u308c\u3066\u3044\u308b\u3002"}]}
|
tohoku-nlp/bert-base-japanese-char-whole-word-masking
| null |
[
"transformers",
"pytorch",
"tf",
"jax",
"bert",
"fill-mask",
"ja",
"dataset:wikipedia",
"license:cc-by-sa-4.0",
"autotrain_compatible",
"endpoints_compatible",
"region:us"
] | null |
2022-03-02T23:29:05+00:00
|
fill-mask
|
transformers
|
# BERT base Japanese (character tokenization)
This is a [BERT](https://github.com/google-research/bert) model pretrained on texts in the Japanese language.
This version of the model processes input texts with word-level tokenization based on the IPA dictionary, followed by character-level tokenization.
The codes for the pretraining are available at [cl-tohoku/bert-japanese](https://github.com/cl-tohoku/bert-japanese/tree/v1.0).
## Model architecture
The model architecture is the same as the original BERT base model; 12 layers, 768 dimensions of hidden states, and 12 attention heads.
## Training Data
The model is trained on Japanese Wikipedia as of September 1, 2019.
To generate the training corpus, [WikiExtractor](https://github.com/attardi/wikiextractor) is used to extract plain texts from a dump file of Wikipedia articles.
The text files used for the training are 2.6GB in size, consisting of approximately 17M sentences.
## Tokenization
The texts are first tokenized by [MeCab](https://taku910.github.io/mecab/) morphological parser with the IPA dictionary and then split into characters.
The vocabulary size is 4000.
## Training
The model is trained with the same configuration as the original BERT; 512 tokens per instance, 256 instances per batch, and 1M training steps.
## Licenses
The pretrained models are distributed under the terms of the [Creative Commons Attribution-ShareAlike 3.0](https://creativecommons.org/licenses/by-sa/3.0/).
## Acknowledgments
For training models, we used Cloud TPUs provided by [TensorFlow Research Cloud](https://www.tensorflow.org/tfrc/) program.
|
{"language": "ja", "license": "cc-by-sa-4.0", "datasets": ["wikipedia"], "widget": [{"text": "\u4ed9\u53f0\u306f\u300c[MASK]\u306e\u90fd\u300d\u3068\u547c\u3070\u308c\u3066\u3044\u308b\u3002"}]}
|
tohoku-nlp/bert-base-japanese-char
| null |
[
"transformers",
"pytorch",
"tf",
"jax",
"bert",
"fill-mask",
"ja",
"dataset:wikipedia",
"license:cc-by-sa-4.0",
"autotrain_compatible",
"endpoints_compatible",
"region:us"
] | null |
2022-03-02T23:29:05+00:00
|
fill-mask
|
transformers
|
# BERT base Japanese (unidic-lite with whole word masking, jawiki-20200831)
This is a [BERT](https://github.com/google-research/bert) model pretrained on texts in the Japanese language.
This version of the model processes input texts with word-level tokenization based on the Unidic 2.1.2 dictionary (available in [unidic-lite](https://pypi.org/project/unidic-lite/) package), followed by the WordPiece subword tokenization.
Additionally, the model is trained with the whole word masking enabled for the masked language modeling (MLM) objective.
The codes for the pretraining are available at [cl-tohoku/bert-japanese](https://github.com/cl-tohoku/bert-japanese/tree/v2.0).
## Model architecture
The model architecture is the same as the original BERT base model; 12 layers, 768 dimensions of hidden states, and 12 attention heads.
## Training Data
The models are trained on the Japanese version of Wikipedia.
The training corpus is generated from the Wikipedia Cirrussearch dump file as of August 31, 2020.
The generated corpus files are 4.0GB in total, containing approximately 30M sentences.
We used the [MeCab](https://taku910.github.io/mecab/) morphological parser with [mecab-ipadic-NEologd](https://github.com/neologd/mecab-ipadic-neologd) dictionary to split texts into sentences.
## Tokenization
The texts are first tokenized by MeCab with the Unidic 2.1.2 dictionary and then split into subwords by the WordPiece algorithm.
The vocabulary size is 32768.
We used [`fugashi`](https://github.com/polm/fugashi) and [`unidic-lite`](https://github.com/polm/unidic-lite) packages for the tokenization.
## Training
The models are trained with the same configuration as the original BERT; 512 tokens per instance, 256 instances per batch, and 1M training steps.
For training of the MLM (masked language modeling) objective, we introduced whole word masking in which all of the subword tokens corresponding to a single word (tokenized by MeCab) are masked at once.
For training of each model, we used a v3-8 instance of Cloud TPUs provided by [TensorFlow Research Cloud program](https://www.tensorflow.org/tfrc/).
The training took about 5 days to finish.
## Licenses
The pretrained models are distributed under the terms of the [Creative Commons Attribution-ShareAlike 3.0](https://creativecommons.org/licenses/by-sa/3.0/).
## Acknowledgments
This model is trained with Cloud TPUs provided by [TensorFlow Research Cloud](https://www.tensorflow.org/tfrc/) program.
|
{"language": "ja", "license": "cc-by-sa-4.0", "datasets": ["wikipedia"], "widget": [{"text": "\u6771\u5317\u5927\u5b66\u3067[MASK]\u306e\u7814\u7a76\u3092\u3057\u3066\u3044\u307e\u3059\u3002"}]}
|
tohoku-nlp/bert-base-japanese-v2
| null |
[
"transformers",
"pytorch",
"tf",
"jax",
"bert",
"fill-mask",
"ja",
"dataset:wikipedia",
"license:cc-by-sa-4.0",
"autotrain_compatible",
"endpoints_compatible",
"has_space",
"region:us"
] | null |
2022-03-02T23:29:05+00:00
|
fill-mask
|
transformers
|
# BERT base Japanese (IPA dictionary, whole word masking enabled)
This is a [BERT](https://github.com/google-research/bert) model pretrained on texts in the Japanese language.
This version of the model processes input texts with word-level tokenization based on the IPA dictionary, followed by the WordPiece subword tokenization.
Additionally, the model is trained with the whole word masking enabled for the masked language modeling (MLM) objective.
The codes for the pretraining are available at [cl-tohoku/bert-japanese](https://github.com/cl-tohoku/bert-japanese/tree/v1.0).
## Model architecture
The model architecture is the same as the original BERT base model; 12 layers, 768 dimensions of hidden states, and 12 attention heads.
## Training Data
The model is trained on Japanese Wikipedia as of September 1, 2019.
To generate the training corpus, [WikiExtractor](https://github.com/attardi/wikiextractor) is used to extract plain texts from a dump file of Wikipedia articles.
The text files used for the training are 2.6GB in size, consisting of approximately 17M sentences.
## Tokenization
The texts are first tokenized by [MeCab](https://taku910.github.io/mecab/) morphological parser with the IPA dictionary and then split into subwords by the WordPiece algorithm.
The vocabulary size is 32000.
## Training
The model is trained with the same configuration as the original BERT; 512 tokens per instance, 256 instances per batch, and 1M training steps.
For the training of the MLM (masked language modeling) objective, we introduced the **Whole Word Masking** in which all of the subword tokens corresponding to a single word (tokenized by MeCab) are masked at once.
## Licenses
The pretrained models are distributed under the terms of the [Creative Commons Attribution-ShareAlike 3.0](https://creativecommons.org/licenses/by-sa/3.0/).
## Acknowledgments
For training models, we used Cloud TPUs provided by [TensorFlow Research Cloud](https://www.tensorflow.org/tfrc/) program.
|
{"language": "ja", "license": "cc-by-sa-4.0", "datasets": ["wikipedia"], "widget": [{"text": "\u6771\u5317\u5927\u5b66\u3067[MASK]\u306e\u7814\u7a76\u3092\u3057\u3066\u3044\u307e\u3059\u3002"}]}
|
tohoku-nlp/bert-base-japanese-whole-word-masking
| null |
[
"transformers",
"pytorch",
"tf",
"jax",
"bert",
"fill-mask",
"ja",
"dataset:wikipedia",
"license:cc-by-sa-4.0",
"autotrain_compatible",
"endpoints_compatible",
"region:us"
] | null |
2022-03-02T23:29:05+00:00
|
fill-mask
|
transformers
|
# BERT base Japanese (IPA dictionary)
This is a [BERT](https://github.com/google-research/bert) model pretrained on texts in the Japanese language.
This version of the model processes input texts with word-level tokenization based on the IPA dictionary, followed by the WordPiece subword tokenization.
The codes for the pretraining are available at [cl-tohoku/bert-japanese](https://github.com/cl-tohoku/bert-japanese/tree/v1.0).
## Model architecture
The model architecture is the same as the original BERT base model; 12 layers, 768 dimensions of hidden states, and 12 attention heads.
## Training Data
The model is trained on Japanese Wikipedia as of September 1, 2019.
To generate the training corpus, [WikiExtractor](https://github.com/attardi/wikiextractor) is used to extract plain texts from a dump file of Wikipedia articles.
The text files used for the training are 2.6GB in size, consisting of approximately 17M sentences.
## Tokenization
The texts are first tokenized by [MeCab](https://taku910.github.io/mecab/) morphological parser with the IPA dictionary and then split into subwords by the WordPiece algorithm.
The vocabulary size is 32000.
## Training
The model is trained with the same configuration as the original BERT; 512 tokens per instance, 256 instances per batch, and 1M training steps.
## Licenses
The pretrained models are distributed under the terms of the [Creative Commons Attribution-ShareAlike 3.0](https://creativecommons.org/licenses/by-sa/3.0/).
## Acknowledgments
For training models, we used Cloud TPUs provided by [TensorFlow Research Cloud](https://www.tensorflow.org/tfrc/) program.
|
{"language": "ja", "license": "cc-by-sa-4.0", "datasets": ["wikipedia"], "widget": [{"text": "\u6771\u5317\u5927\u5b66\u3067[MASK]\u306e\u7814\u7a76\u3092\u3057\u3066\u3044\u307e\u3059\u3002"}]}
|
tohoku-nlp/bert-base-japanese
| null |
[
"transformers",
"pytorch",
"tf",
"jax",
"bert",
"fill-mask",
"ja",
"dataset:wikipedia",
"license:cc-by-sa-4.0",
"autotrain_compatible",
"endpoints_compatible",
"region:us"
] | null |
2022-03-02T23:29:05+00:00
|
fill-mask
|
transformers
|
# BERT large Japanese (character-level tokenization with whole word masking, jawiki-20200831)
This is a [BERT](https://github.com/google-research/bert) model pretrained on texts in the Japanese language.
This version of the model processes input texts with word-level tokenization based on the Unidic 2.1.2 dictionary (available in [unidic-lite](https://pypi.org/project/unidic-lite/) package), followed by character-level tokenization.
Additionally, the model is trained with the whole word masking enabled for the masked language modeling (MLM) objective.
The codes for the pretraining are available at [cl-tohoku/bert-japanese](https://github.com/cl-tohoku/bert-japanese/tree/v2.0).
## Model architecture
The model architecture is the same as the original BERT large model; 24 layers, 1024 dimensions of hidden states, and 16 attention heads.
## Training Data
The models are trained on the Japanese version of Wikipedia.
The training corpus is generated from the Wikipedia Cirrussearch dump file as of August 31, 2020.
The generated corpus files are 4.0GB in total, containing approximately 30M sentences.
We used the [MeCab](https://taku910.github.io/mecab/) morphological parser with [mecab-ipadic-NEologd](https://github.com/neologd/mecab-ipadic-neologd) dictionary to split texts into sentences.
## Tokenization
The texts are first tokenized by MeCab with the Unidic 2.1.2 dictionary and then split into characters.
The vocabulary size is 6144.
We used [`fugashi`](https://github.com/polm/fugashi) and [`unidic-lite`](https://github.com/polm/unidic-lite) packages for the tokenization.
## Training
The models are trained with the same configuration as the original BERT; 512 tokens per instance, 256 instances per batch, and 1M training steps.
For training of the MLM (masked language modeling) objective, we introduced whole word masking in which all of the subword tokens corresponding to a single word (tokenized by MeCab) are masked at once.
For training of each model, we used a v3-8 instance of Cloud TPUs provided by [TensorFlow Research Cloud program](https://www.tensorflow.org/tfrc/).
The training took about 5 days to finish.
## Licenses
The pretrained models are distributed under the terms of the [Creative Commons Attribution-ShareAlike 3.0](https://creativecommons.org/licenses/by-sa/3.0/).
## Acknowledgments
This model is trained with Cloud TPUs provided by [TensorFlow Research Cloud](https://www.tensorflow.org/tfrc/) program.
|
{"language": "ja", "license": "cc-by-sa-4.0", "datasets": ["wikipedia"], "widget": [{"text": "\u6771\u5317\u5927\u5b66\u3067[MASK]\u306e\u7814\u7a76\u3092\u3057\u3066\u3044\u307e\u3059\u3002"}]}
|
tohoku-nlp/bert-large-japanese-char
| null |
[
"transformers",
"pytorch",
"tf",
"jax",
"bert",
"fill-mask",
"ja",
"dataset:wikipedia",
"license:cc-by-sa-4.0",
"autotrain_compatible",
"endpoints_compatible",
"region:us"
] | null |
2022-03-02T23:29:05+00:00
|
fill-mask
|
transformers
|
# BERT large Japanese (unidic-lite with whole word masking, jawiki-20200831)
This is a [BERT](https://github.com/google-research/bert) model pretrained on texts in the Japanese language.
This version of the model processes input texts with word-level tokenization based on the Unidic 2.1.2 dictionary (available in [unidic-lite](https://pypi.org/project/unidic-lite/) package), followed by the WordPiece subword tokenization.
Additionally, the model is trained with the whole word masking enabled for the masked language modeling (MLM) objective.
The codes for the pretraining are available at [cl-tohoku/bert-japanese](https://github.com/cl-tohoku/bert-japanese/tree/v2.0).
## Model architecture
The model architecture is the same as the original BERT large model; 24 layers, 1024 dimensions of hidden states, and 16 attention heads.
## Training Data
The models are trained on the Japanese version of Wikipedia.
The training corpus is generated from the Wikipedia Cirrussearch dump file as of August 31, 2020.
The generated corpus files are 4.0GB in total, containing approximately 30M sentences.
We used the [MeCab](https://taku910.github.io/mecab/) morphological parser with [mecab-ipadic-NEologd](https://github.com/neologd/mecab-ipadic-neologd) dictionary to split texts into sentences.
## Tokenization
The texts are first tokenized by MeCab with the Unidic 2.1.2 dictionary and then split into subwords by the WordPiece algorithm.
The vocabulary size is 32768.
We used [`fugashi`](https://github.com/polm/fugashi) and [`unidic-lite`](https://github.com/polm/unidic-lite) packages for the tokenization.
## Training
The models are trained with the same configuration as the original BERT; 512 tokens per instance, 256 instances per batch, and 1M training steps.
For training of the MLM (masked language modeling) objective, we introduced whole word masking in which all of the subword tokens corresponding to a single word (tokenized by MeCab) are masked at once.
For training of each model, we used a v3-8 instance of Cloud TPUs provided by [TensorFlow Research Cloud program](https://www.tensorflow.org/tfrc/).
The training took about 5 days to finish.
## Licenses
The pretrained models are distributed under the terms of the [Creative Commons Attribution-ShareAlike 3.0](https://creativecommons.org/licenses/by-sa/3.0/).
## Acknowledgments
This model is trained with Cloud TPUs provided by [TensorFlow Research Cloud](https://www.tensorflow.org/tfrc/) program.
|
{"language": "ja", "license": "cc-by-sa-4.0", "datasets": ["wikipedia"], "widget": [{"text": "\u6771\u5317\u5927\u5b66\u3067[MASK]\u306e\u7814\u7a76\u3092\u3057\u3066\u3044\u307e\u3059\u3002"}]}
|
tohoku-nlp/bert-large-japanese
| null |
[
"transformers",
"pytorch",
"tf",
"jax",
"bert",
"fill-mask",
"ja",
"dataset:wikipedia",
"license:cc-by-sa-4.0",
"autotrain_compatible",
"endpoints_compatible",
"region:us"
] | null |
2022-03-02T23:29:05+00:00
|
fill-mask
|
transformers
|
{}
|
tohoku-nlp/roberta-base-japanese
| null |
[
"transformers",
"pytorch",
"roberta",
"fill-mask",
"autotrain_compatible",
"endpoints_compatible",
"region:us"
] | null |
2022-03-02T23:29:05+00:00
|
|
null | null |
{}
|
cla/sr
| null |
[
"region:us"
] | null |
2022-03-02T23:29:05+00:00
|
|
question-answering
|
transformers
|
{}
|
clagator/biobert_squad2_cased
| null |
[
"transformers",
"pytorch",
"jax",
"bert",
"question-answering",
"endpoints_compatible",
"region:us"
] | null |
2022-03-02T23:29:05+00:00
|
|
feature-extraction
|
transformers
|
{}
|
clagator/biobert_v1.1_pubmed_nli_sts
| null |
[
"transformers",
"pytorch",
"jax",
"bert",
"feature-extraction",
"endpoints_compatible",
"region:us"
] | null |
2022-03-02T23:29:05+00:00
|
|
text-generation
|
transformers
|
# A somewhat positive chatbot
|
{"tags": ["conversational"]}
|
clairesb/kindness_bot
| null |
[
"transformers",
"pytorch",
"gpt2",
"text-generation",
"conversational",
"autotrain_compatible",
"endpoints_compatible",
"text-generation-inference",
"region:us"
] | null |
2022-03-02T23:29:05+00:00
|
text-generation
|
transformers
|
# Affirmation Bot
|
{"tags": ["conversational"]}
|
clairesb/kindness_bot_repo
| null |
[
"transformers",
"pytorch",
"gpt2",
"text-generation",
"conversational",
"autotrain_compatible",
"endpoints_compatible",
"text-generation-inference",
"region:us"
] | null |
2022-03-02T23:29:05+00:00
|
text-classification
|
transformers
|
# Multi-lingual sentiment prediction trained from COVID19-related tweets
Repository: [https://github.com/clampert/multilingual-sentiment-analysis/](https://github.com/clampert/multilingual-sentiment-analysis/)
Model trained on a large-scale (18437530 examples) dataset of
multi-lingual tweets that was collected between March 2020
and November 2021 using Twitter’s Streaming API with varying
COVID19-related keywords. Labels were auto-general based on
the presence of positive and negative emoticons. For details
on the dataset, see our IEEE BigData 2021 publication.
Base model is [sentence-transformers/stsb-xlm-r-multilingual](https://huggingface.co/sentence-transformers/stsb-xlm-r-multilingual).
It was finetuned for sequence classification with `positive`
and `negative` labels for two epochs (48 hours on 8xP100 GPUs).
## Citation
If you use our model your work, please cite:
```
@inproceedings{lampert2021overcoming,
title={Overcoming Rare-Language Discrimination in Multi-Lingual Sentiment Analysis},
author={Jasmin Lampert and Christoph H. Lampert},
booktitle={IEEE International Conference on Big Data (BigData)},
year={2021},
note={Special Session: Machine Learning on Big Data},
}
```
Enjoy!
|
{"language": "multilingual", "license": "apache-2.0", "tags": ["sentiment-analysis", "multilingual"], "pipeline_tag": "text-classification", "widget": [{"text": "I am very happy.", "example_title": "English"}, {"text": "Heute bin ich schlecht drauf.", "example_title": "Deutsch"}, {"text": "Quel cauchemard!", "example_title": "Francais"}, {"text": "\u0e09\u0e31\u0e19\u0e23\u0e31\u0e01\u0e24\u0e14\u0e39\u0e43\u0e1a\u0e44\u0e21\u0e49\u0e1c\u0e25\u0e34", "example_title": "\u0e20\u0e32\u0e29\u0e32\u0e44\u0e17\u0e22"}]}
|
clampert/multilingual-sentiment-covid19
| null |
[
"transformers",
"pytorch",
"xlm-roberta",
"text-classification",
"sentiment-analysis",
"multilingual",
"license:apache-2.0",
"autotrain_compatible",
"endpoints_compatible",
"region:us"
] | null |
2022-03-02T23:29:05+00:00
|
text-generation
|
transformers
|
{"tags": ["conversational"]}
|
clancystudios/DialoGPT-medium-Morty
| null |
[
"transformers",
"pytorch",
"gpt2",
"text-generation",
"conversational",
"autotrain_compatible",
"endpoints_compatible",
"text-generation-inference",
"region:us"
] | null |
2022-03-02T23:29:05+00:00
|
|
null | null |
# KGR10 FastText Polish word embeddings
Distributional language model (both textual and binary) for Polish (word embeddings) trained on KGR10 corpus (over 4 billion of words) using Fasttext with the following variants (all possible combinations):
- dimension: 100, 300
- method: skipgram, cbow
- tool: FastText, Magnitude
- source text: plain, plain.lower, plain.lemma, plain.lemma.lower
## Models
In the repository you can find 4 selected models, that were examined in the paper (see Citation).
A model that performed the best is the default model/config (see `default_config.json`).
## Usage
To use these embedding models easily, it is required to install [embeddings](https://github.com/CLARIN-PL/embeddings).
```bash
pip install clarinpl-embeddings
```
### Utilising the default model (the easiest way)
Word embedding:
```python
from embeddings.embedding.auto_flair import AutoFlairWordEmbedding
from flair.data import Sentence
sentence = Sentence("Myśl z duszy leci bystro, Nim się w słowach złamie.")
embedding = AutoFlairWordEmbedding.from_hub("clarin-pl/fastText-kgr10")
embedding.embed([sentence])
for token in sentence:
print(token)
print(token.embedding)
```
Document embedding (averaged over words):
```python
from embeddings.embedding.auto_flair import AutoFlairDocumentEmbedding
from flair.data import Sentence
sentence = Sentence("Myśl z duszy leci bystro, Nim się w słowach złamie.")
embedding = AutoFlairDocumentEmbedding.from_hub("clarin-pl/fastText-kgr10")
embedding.embed([sentence])
print(sentence.embedding)
```
### Customisable way
Word embedding:
```python
from embeddings.embedding.static.embedding import AutoStaticWordEmbedding
from embeddings.embedding.static.fasttext import KGR10FastTextConfig
from flair.data import Sentence
config = KGR10FastTextConfig(method='cbow', dimension=100)
embedding = AutoStaticWordEmbedding.from_config(config)
sentence = Sentence("Myśl z duszy leci bystro, Nim się w słowach złamie.")
embedding.embed([sentence])
for token in sentence:
print(token)
print(token.embedding)
```
Document embedding (averaged over words):
```python
from embeddings.embedding.static.embedding import AutoStaticDocumentEmbedding
from embeddings.embedding.static.fasttext import KGR10FastTextConfig
from flair.data import Sentence
config = KGR10FastTextConfig(method='cbow', dimension=100)
embedding = AutoStaticDocumentEmbedding.from_config(config)
sentence = Sentence("Myśl z duszy leci bystro, Nim się w słowach złamie.")
embedding.embed([sentence])
print(sentence.embedding)
```
## Citation
The link below leads to the NextCloud directory with all variants of embeddings. If you use it, please cite the following article:
```
@article{kocon2018embeddings,
author = {Koco\'{n}, Jan and Gawor, Micha{\l}},
title = {Evaluating {KGR10} {P}olish word embeddings in the recognition of temporal
expressions using {BiLSTM-CRF}},
journal = {Schedae Informaticae},
volume = {27},
year = {2018},
url = {http://www.ejournals.eu/Schedae-Informaticae/2018/Volume-27/art/13931/},
doi = {10.4467/20838476SI.18.008.10413}
}
```
|
{"language": "pl", "tags": ["fastText"], "datasets": ["kgr10"]}
|
clarin-pl/fastText-kgr10
| null |
[
"fastText",
"pl",
"dataset:kgr10",
"region:us"
] | null |
2022-03-02T23:29:05+00:00
|
fill-mask
|
transformers
|
{}
|
clarin-pl/herbert-kgr10
| null |
[
"transformers",
"pytorch",
"bert",
"fill-mask",
"autotrain_compatible",
"endpoints_compatible",
"region:us"
] | null |
2022-03-02T23:29:05+00:00
|
|
fill-mask
|
transformers
|
# Work in Progress Polish RoBERTa
The model has been trained for about 5% time of the target. We will publish new increments as they will be trained.
The model pre-trained on KGR10 corpora.
More about model at [CLARIN-dspace](https://huggingface.co/clarin/roberta-polish-v1)
## Usage
## Huggingface model hub
## Acknowledgments
[CLARIN-PL and CLARIN-BIZ project](https://clarin-pl.eu/)
|
{}
|
clarin-pl/roberta-polish-kgr10
| null |
[
"transformers",
"pytorch",
"jax",
"roberta",
"fill-mask",
"autotrain_compatible",
"endpoints_compatible",
"has_space",
"region:us"
] | null |
2022-03-02T23:29:05+00:00
|
null | null |
# KGR10 word2vec Polish word embeddings
Distributional language models for Polish trained on the KGR10 corpora.
## Models
In the repository you can find two selected models, that were selected after evaluation (see table below).
A model that performed the best is the default model/config (see `default_config.json`).
|method|dimension|hs|mwe||
|---|---|---|---| --- |
|cbow|300|false|true| <-- default |
|skipgram|300|true|true|
## Usage
To use these embedding models easily, it is required to install [embeddings](https://github.com/CLARIN-PL/embeddings).
```bash
pip install clarinpl-embeddings
```
### Utilising the default model (the easiest way)
Word embedding:
```python
from embeddings.embedding.auto_flair import AutoFlairWordEmbedding
from flair.data import Sentence
sentence = Sentence("Myśl z duszy leci bystro, Nim się w słowach złamie.")
embedding = AutoFlairWordEmbedding.from_hub("clarin-pl/word2vec-kgr10")
embedding.embed([sentence])
for token in sentence:
print(token)
print(token.embedding)
```
Document embedding (averaged over words):
```python
from embeddings.embedding.auto_flair import AutoFlairDocumentEmbedding
from flair.data import Sentence
sentence = Sentence("Myśl z duszy leci bystro, Nim się w słowach złamie.")
embedding = AutoFlairDocumentEmbedding.from_hub("clarin-pl/word2vec-kgr10")
embedding.embed([sentence])
print(sentence.embedding)
```
### Customisable way
Word embedding:
```python
from embeddings.embedding.static.embedding import AutoStaticWordEmbedding
from embeddings.embedding.static.word2vec import KGR10Word2VecConfig
from flair.data import Sentence
config = KGR10Word2VecConfig(method='skipgram', hs=False)
embedding = AutoStaticWordEmbedding.from_config(config)
sentence = Sentence("Myśl z duszy leci bystro, Nim się w słowach złamie.")
embedding.embed([sentence])
for token in sentence:
print(token)
print(token.embedding)
```
Document embedding (averaged over words):
```python
from embeddings.embedding.static.embedding import AutoStaticDocumentEmbedding
from embeddings.embedding.static.word2vec import KGR10Word2VecConfig
from flair.data import Sentence
config = KGR10Word2VecConfig(method='skipgram', hs=False)
embedding = AutoStaticDocumentEmbedding.from_config(config)
sentence = Sentence("Myśl z duszy leci bystro, Nim się w słowach złamie.")
embedding.embed([sentence])
print(sentence.embedding)
```
## Citation
```
Piasecki, Maciej; Janz, Arkadiusz; Kaszewski, Dominik; et al., 2017, Word Embeddings for Polish, CLARIN-PL digital repository, http://hdl.handle.net/11321/442.
```
or
```
@misc{11321/442,
title = {Word Embeddings for Polish},
author = {Piasecki, Maciej and Janz, Arkadiusz and Kaszewski, Dominik and Czachor, Gabriela},
url = {http://hdl.handle.net/11321/442},
note = {{CLARIN}-{PL} digital repository},
copyright = {{GNU} {GPL3}},
year = {2017}
}
```
|
{"language": "pl", "tags": ["word2vec"], "datasets": ["KGR10"]}
|
clarin-pl/word2vec-kgr10
| null |
[
"word2vec",
"pl",
"dataset:KGR10",
"has_space",
"region:us"
] | null |
2022-03-02T23:29:05+00:00
|
text-classification
|
transformers
|
# bcms-bertic-frenk-hate
Text classification model based on [`classla/bcms-bertic`](https://huggingface.co/classla/bcms-bertic) and fine-tuned on the [FRENK dataset](https://www.clarin.si/repository/xmlui/handle/11356/1433) comprising of LGBT and migrant hatespeech. Only the Croatian subset of the data was used for fine-tuning and the dataset has been relabeled for binary classification (offensive or acceptable).
## Fine-tuning hyperparameters
Fine-tuning was performed with `simpletransformers`. Beforehand a brief hyperparameter optimisation was performed and the presumed optimal hyperparameters are:
```python
model_args = {
"num_train_epochs": 12,
"learning_rate": 1e-5,
"train_batch_size": 74}
```
## Performance
The same pipeline was run with two other transformer models and `fasttext` for comparison. Accuracy and macro F1 score were recorded for each of the 6 fine-tuning sessions and post festum analyzed.
| model | average accuracy | average macro F1 |
|----------------------------|------------------|------------------|
| bcms-bertic-frenk-hate | 0.8313 | 0.8219 |
| EMBEDDIA/crosloengual-bert | 0.8054 | 0.796 |
| xlm-roberta-base | 0.7175 | 0.7049 |
| fasttext | 0.771 | 0.754 |
From recorded accuracies and macro F1 scores p-values were also calculated:
Comparison with `crosloengual-bert`:
| test | accuracy p-value | macro F1 p-value |
|----------------|------------------|------------------|
| Wilcoxon | 0.00781 | 0.00781 |
| Mann Whithney | 0.00108 | 0.00108 |
| Student t-test | 2.43e-10 | 1.27e-10 |
Comparison with `xlm-roberta-base`:
| test | accuracy p-value | macro F1 p-value |
|----------------|------------------|------------------|
| Wilcoxon | 0.00781 | 0.00781 |
| Mann Whithney | 0.00107 | 0.00108 |
| Student t-test | 4.83e-11 | 5.61e-11 |
## Use examples
```python
from simpletransformers.classification import ClassificationModel
model = ClassificationModel(
"bert", "5roop/bcms-bertic-frenk-hate", use_cuda=True,
)
predictions, logit_output = model.predict(['Ne odbacujem da će RH primiti još migranata iz Afganistana, no neće biti novog vala',
"Potpredsjednik Vlade i ministar branitelja Tomo Medved komentirao je Vladine planove za zakonsku zabranu pozdrava 'za dom spremni' "])
predictions
### Output:
### array([0, 0])
```
## Citation
If you use the model, please cite the following paper on which the original model is based:
```
@inproceedings{ljubesic-lauc-2021-bertic,
title = "{BERT}i{\'c} - The Transformer Language Model for {B}osnian, {C}roatian, {M}ontenegrin and {S}erbian",
author = "Ljube{\v{s}}i{\'c}, Nikola and Lauc, Davor",
booktitle = "Proceedings of the 8th Workshop on Balto-Slavic Natural Language Processing",
month = apr,
year = "2021",
address = "Kiyv, Ukraine",
publisher = "Association for Computational Linguistics",
url = "https://www.aclweb.org/anthology/2021.bsnlp-1.5",
pages = "37--42",
}
```
and the dataset used for fine-tuning:
```
@misc{ljubešić2019frenk,
title={The FRENK Datasets of Socially Unacceptable Discourse in Slovene and English},
author={Nikola Ljubešić and Darja Fišer and Tomaž Erjavec},
year={2019},
eprint={1906.02045},
archivePrefix={arXiv},
primaryClass={cs.CL},
url={https://arxiv.org/abs/1906.02045}
}
```
|
{"language": "hr", "license": "cc-by-sa-4.0", "tags": ["text-classification", "hate-speech"], "widget": [{"text": "Potpredsjednik Vlade i ministar branitelja Tomo Medved komentirao je Vladine planove za zakonsku zabranu pozdrava 'za dom spremni'."}]}
|
classla/bcms-bertic-frenk-hate
| null |
[
"transformers",
"pytorch",
"safetensors",
"bert",
"text-classification",
"hate-speech",
"hr",
"arxiv:1906.02045",
"license:cc-by-sa-4.0",
"autotrain_compatible",
"endpoints_compatible",
"region:us"
] | null |
2022-03-02T23:29:05+00:00
|
null |
transformers
|
# BERTić* [bert-ich] /bɜrtitʃ/ - A transformer language model for Bosnian, Croatian, Montenegrin and Serbian
* The name should resemble the facts (1) that the model was trained in Zagreb, Croatia, where diminutives ending in -ić (as in fotić, smajlić, hengić etc.) are very popular, and (2) that most surnames in the countries where these languages are spoken end in -ić (with diminutive etymology as well).
This is the smaller generator of the main [discriminator model](https://huggingface.co/classla/bcms-bertic), useful if you want to continue pre-training the discriminator model.
If you use the model, please cite the following paper:
```
@inproceedings{ljubesic-lauc-2021-bertic,
title = "{BERT}i{\'c} - The Transformer Language Model for {B}osnian, {C}roatian, {M}ontenegrin and {S}erbian",
author = "Ljube{\v{s}}i{\'c}, Nikola and Lauc, Davor",
booktitle = "Proceedings of the 8th Workshop on Balto-Slavic Natural Language Processing",
month = apr,
year = "2021",
address = "Kiyv, Ukraine",
publisher = "Association for Computational Linguistics",
url = "https://www.aclweb.org/anthology/2021.bsnlp-1.5",
pages = "37--42",
}
```
|
{"language": ["hr", "bs", "sr", "cnr", "hbs"], "license": "apache-2.0", "tags": ["masked-lm"], "widget": [{"text": "Zovem se Marko i radim u [MASK]."}]}
|
classla/bcms-bertic-generator
| null |
[
"transformers",
"pytorch",
"electra",
"pretraining",
"masked-lm",
"hr",
"bs",
"sr",
"cnr",
"hbs",
"license:apache-2.0",
"endpoints_compatible",
"region:us"
] | null |
2022-03-02T23:29:05+00:00
|
null |
transformers
|
{}
|
classla/bcms-bertic-geo
| null |
[
"transformers",
"pytorch",
"electra",
"endpoints_compatible",
"region:us"
] | null |
2022-03-02T23:29:05+00:00
|
|
token-classification
|
transformers
|
# The [BERTić](https://huggingface.co/classla/bcms-bertic)* [bert-ich] /bɜrtitʃ/ model fine-tuned for the task of named entity recognition in Bosnian, Croatian, Montenegrin and Serbian (BCMS)
* The name should resemble the facts (1) that the model was trained in Zagreb, Croatia, where diminutives ending in -ić (as in fotić, smajlić, hengić etc.) are very popular, and (2) that most surnames in the countries where these languages are spoken end in -ić (with diminutive etymology as well).
This is a fine-tuned version of the [BERTić](https://huggingface.co/classla/bcms-bertic) model for the task of named entity recognition (PER, LOC, ORG, MISC). The fine-tuning was performed on the following datasets:
- the [hr500k](http://hdl.handle.net/11356/1183) dataset, 500 thousand tokens in size, standard Croatian
- the [SETimes.SR](http://hdl.handle.net/11356/1200) dataset, 87 thousand tokens in size, standard Serbian
- the [ReLDI-hr](http://hdl.handle.net/11356/1241) dataset, 89 thousand tokens in size, Internet (Twitter) Croatian
- the [ReLDI-sr](http://hdl.handle.net/11356/1240) dataset, 92 thousand tokens in size, Internet (Twitter) Serbian
The data was augmented with missing diacritics and standard data was additionally over-represented. The F1 obtained on dev data (train and test was merged into train) is 91.38. For a more detailed per-dataset evaluation of the BERTić model on the NER task have a look at the [main model page](https://huggingface.co/classla/bcms-bertic).
If you use this fine-tuned model, please cite the following paper:
```
@inproceedings{ljubesic-lauc-2021-bertic,
title = "{BERT}i{\'c} - The Transformer Language Model for {B}osnian, {C}roatian, {M}ontenegrin and {S}erbian",
author = "Ljube{\v{s}}i{\'c}, Nikola and Lauc, Davor",
booktitle = "Proceedings of the 8th Workshop on Balto-Slavic Natural Language Processing",
month = apr,
year = "2021",
address = "Kiyv, Ukraine",
publisher = "Association for Computational Linguistics",
url = "https://www.aclweb.org/anthology/2021.bsnlp-1.5",
pages = "37--42",
}
```
When running the model in `simpletransformers`, the order of labels has to be set as well.
```
from simpletransformers.ner import NERModel, NERArgs
model_args = NERArgs()
model_args.labels_list = ['B-LOC','B-MISC','B-ORG','B-PER','I-LOC','I-MISC','I-ORG','I-PER','O']
model = NERModel('electra', 'classla/bcms-bertic-ner', args=model_args)
```
|
{"language": ["hr", "bs", "sr", "cnr", "hbs"], "license": "apache-2.0", "widget": [{"text": "Zovem se Marko i \u017eivim u Zagrebu. Studirao sam u Beogradu na Filozofskom fakultetu. Obo\u017eavam album Moanin."}]}
|
classla/bcms-bertic-ner
| null |
[
"transformers",
"pytorch",
"safetensors",
"electra",
"token-classification",
"hr",
"bs",
"sr",
"cnr",
"hbs",
"license:apache-2.0",
"autotrain_compatible",
"endpoints_compatible",
"has_space",
"region:us"
] | null |
2022-03-02T23:29:05+00:00
|
null |
transformers
|
# BERTić* [bert-ich] /bɜrtitʃ/ - A transformer language model for Bosnian, Croatian, Montenegrin and Serbian
* The name should resemble the facts (1) that the model was trained in Zagreb, Croatia, where diminutives ending in -ić (as in fotić, smajlić, hengić etc.) are very popular, and (2) that most surnames in the countries where these languages are spoken end in -ić (with diminutive etymology as well).
This Electra model was trained on more than 8 billion tokens of Bosnian, Croatian, Montenegrin and Serbian text.
***new*** We have published a version of this model fine-tuned on the named entity recognition task ([bcms-bertic-ner](https://huggingface.co/classla/bcms-bertic-ner)) and on the hate speech detection task ([bcms-bertic-frenk-hate](https://huggingface.co/classla/bcms-bertic-frenk-hate)).
If you use the model, please cite the following paper:
```
@inproceedings{ljubesic-lauc-2021-bertic,
title = "{BERT}i{\'c} - The Transformer Language Model for {B}osnian, {C}roatian, {M}ontenegrin and {S}erbian",
author = "Ljube{\v{s}}i{\'c}, Nikola and Lauc, Davor",
booktitle = "Proceedings of the 8th Workshop on Balto-Slavic Natural Language Processing",
month = apr,
year = "2021",
address = "Kiyv, Ukraine",
publisher = "Association for Computational Linguistics",
url = "https://www.aclweb.org/anthology/2021.bsnlp-1.5",
pages = "37--42",
}
```
## Benchmarking
Comparing this model to [multilingual BERT](https://huggingface.co/bert-base-multilingual-cased) and [CroSloEngual BERT](https://huggingface.co/EMBEDDIA/crosloengual-bert) on the tasks of (1) part-of-speech tagging, (2) named entity recognition, (3) geolocation prediction, and (4) commonsense causal reasoning, shows the BERTić model to be superior to the other two.
### Part-of-speech tagging
Evaluation metric is (seqeval) microF1. Reported are means of five runs. Best results are presented in bold. Statistical significance is calculated between two best-performing systems via a two-tailed t-test (* p<=0.05, ** p<=0.01, *** p<=0.001, ***** p<=0.0001).
Dataset | Language | Variety | CLASSLA | mBERT | cseBERT | BERTić
---|---|---|---|---|---|---
hr500k | Croatian | standard | 93.87 | 94.60 | 95.74 | **95.81*****
reldi-hr | Croatian | internet non-standard | - | 88.87 | 91.63 | **92.28*****
SETimes.SR | Serbian | standard | 95.00 | 95.50 | **96.41** | 96.31
reldi-sr | Serbian | internet non-standard | - | 91.26 | 93.54 | **93.90*****
### Named entity recognition
Evaluation metric is (seqeval) microF1. Reported are means of five runs. Best results are presented in bold. Statistical significance is calculated between two best-performing systems via a two-tailed t-test (* p<=0.05, ** p<=0.01, *** p<=0.001, ***** p<=0.0001).
Dataset | Language | Variety | CLASSLA | mBERT | cseBERT | BERTić
---|---|---|---|---|---|---
hr500k | Croatian | standard | 80.13 | 85.67 | 88.98 | **89.21******
reldi-hr | Croatian | internet non-standard | - | 76.06 | 81.38 | **83.05******
SETimes.SR | Serbian | standard | 84.64 | **92.41** | 92.28 | 92.02
reldi-sr | Serbian | internet non-standard | - | 81.29 | 82.76 | **87.92******
### Geolocation prediction
The dataset comes from the VarDial 2020 evaluation campaign's shared task on [Social Media variety Geolocation prediction](https://sites.google.com/view/vardial2020/evaluation-campaign). The task is to predict the latitude and longitude of a tweet given its text.
Evaluation metrics are median and mean of distance between gold and predicted geolocations (lower is better). No statistical significance is computed due to large test set (39,723 instances). Centroid baseline predicts each text to be created in the centroid of the training dataset.
System | Median | Mean
---|---|---
centroid | 107.10 | 145.72
mBERT | 42.25 | 82.05
cseBERT | 40.76 | 81.88
BERTić | **37.96** | **79.30**
### Choice Of Plausible Alternatives
The dataset is a translation of the [COPA dataset](https://people.ict.usc.edu/~gordon/copa.html) into Croatian ([link to the dataset](http://hdl.handle.net/11356/1404)).
Evaluation metric is accuracy. Reported are means of five runs. Best results are presented in bold. Statistical significance is calculated between two best-performing systems via a two-tailed t-test (* p<=0.05, ** p<=0.01, *** p<=0.001, ***** p<=0.0001).
System | Accuracy
---|---
random | 50.00
mBERT | 54.12
cseBERT | 61.80
BERTić | **65.76****
|
{"language": ["hr", "bs", "sr", "cnr", "hbs"], "license": "apache-2.0"}
|
classla/bcms-bertic
| null |
[
"transformers",
"pytorch",
"electra",
"pretraining",
"hr",
"bs",
"sr",
"cnr",
"hbs",
"license:apache-2.0",
"endpoints_compatible",
"has_space",
"region:us"
] | null |
2022-03-02T23:29:05+00:00
|
null |
transformers
|
{}
|
classla/bert-base-german-dbmdz-uncased-geo
| null |
[
"transformers",
"pytorch",
"bert",
"endpoints_compatible",
"region:us"
] | null |
2022-03-02T23:29:05+00:00
|
|
text-classification
|
transformers
|
# roberta-base-frenk-hate
Text classification model based on [`roberta-base`](https://huggingface.co/roberta-base) and fine-tuned on the [FRENK dataset](https://www.clarin.si/repository/xmlui/handle/11356/1433) comprising of LGBT and migrant hatespeech. Only the English subset of the data was used for fine-tuning and the dataset has been relabeled for binary classification (offensive or acceptable).
## Fine-tuning hyperparameters
Fine-tuning was performed with `simpletransformers`. Beforehand a brief hyperparameter optimisation was performed and the presumed optimal hyperparameters are:
```python
model_args = {
"num_train_epochs": 6,
"learning_rate": 3e-6,
"train_batch_size": 69}
```
## Performance
The same pipeline was run with two other transformer models and `fasttext` for comparison. Accuracy and macro F1 score were recorded for each of the 6 fine-tuning sessions and post festum analyzed.
| model | average accuracy | average macro F1|
|---|---|---|
|roberta-base-frenk-hate|0.7915|0.7785|
|xlm-roberta-large |0.7904|0.77876|
|xlm-roberta-base |0.7577|0.7402|
|fasttext|0.725 |0.707 |
From recorded accuracies and macro F1 scores p-values were also calculated:
Comparison with `xlm-roberta-base`:
| test | accuracy p-value | macro F1 p-value|
| --- | --- | --- |
|Wilcoxon|0.00781|0.00781|
|Mann Whithney U-test|0.00108|0.00108|
|Student t-test | 1.35e-08 | 1.05e-07|
Comparison with `xlm-roberta-large` yielded inconclusive results. `roberta-base` has average accuracy 0.7915, while `xlm-roberta-large` has average accuracy of 0.7904. If macro F1 scores were to be compared, `roberta-base` actually has lower average than `xlm-roberta-large`: 0.77852 vs 0.77876 respectively. The same statistical tests were performed with the premise that `roberta-base` has greater metrics, and the results are given below.
| test | accuracy p-value | macro F1 p-value|
| --- | --- | --- |
|Wilcoxon|0.188|0.406|
|Mann Whithey|0.375|0.649|
|Student t-test | 0.681| 0.934|
With reversed premise (i.e., that `xlm-roberta-large` has greater statistics) the Wilcoxon p-value for macro F1 scores for this case reaches 0.656, Mann-Whithey p-value is 0.399, and of course the Student p-value stays the same. It was therefore concluded that performance of the two models are not statistically significantly different from one another.
## Use examples
```python
from simpletransformers.classification import ClassificationModel
model_args = {
"num_train_epochs": 6,
"learning_rate": 3e-6,
"train_batch_size": 69}
model = ClassificationModel(
"roberta", "5roop/roberta-base-frenk-hate", use_cuda=True,
args=model_args
)
predictions, logit_output = model.predict(["Build the wall",
"Build the wall of trust"]
)
predictions
### Output:
### array([1, 0])
```
## Citation
If you use the model, please cite the following paper on which the original model is based:
```
@article{DBLP:journals/corr/abs-1907-11692,
author = {Yinhan Liu and
Myle Ott and
Naman Goyal and
Jingfei Du and
Mandar Joshi and
Danqi Chen and
Omer Levy and
Mike Lewis and
Luke Zettlemoyer and
Veselin Stoyanov},
title = {RoBERTa: {A} Robustly Optimized {BERT} Pretraining Approach},
journal = {CoRR},
volume = {abs/1907.11692},
year = {2019},
url = {http://arxiv.org/abs/1907.11692},
archivePrefix = {arXiv},
eprint = {1907.11692},
timestamp = {Thu, 01 Aug 2019 08:59:33 +0200},
biburl = {https://dblp.org/rec/journals/corr/abs-1907-11692.bib},
bibsource = {dblp computer science bibliography, https://dblp.org}
}
```
and the dataset used for fine-tuning:
```
@misc{ljubešić2019frenk,
title={The FRENK Datasets of Socially Unacceptable Discourse in Slovene and English},
author={Nikola Ljubešić and Darja Fišer and Tomaž Erjavec},
year={2019},
eprint={1906.02045},
archivePrefix={arXiv},
primaryClass={cs.CL},
url={https://arxiv.org/abs/1906.02045}
}
```
|
{"language": "en", "license": "cc-by-sa-4.0", "tags": ["text-classification", "hate-speech"], "widget": [{"text": "Gay is okay."}]}
|
classla/roberta-base-frenk-hate
| null |
[
"transformers",
"pytorch",
"safetensors",
"roberta",
"text-classification",
"hate-speech",
"en",
"arxiv:1907.11692",
"arxiv:1906.02045",
"license:cc-by-sa-4.0",
"autotrain_compatible",
"endpoints_compatible",
"has_space",
"region:us"
] | null |
2022-03-02T23:29:05+00:00
|
text-classification
|
transformers
|
Text classification model based on `EMBEDDIA/sloberta` and fine-tuned on the [FRENK dataset](https://www.clarin.si/repository/xmlui/handle/11356/1433) comprising of LGBT and migrant hatespeech. Only the slovenian subset of the data was used for fine-tuning and the dataset has been relabeled for binary classification (offensive or acceptable).
## Fine-tuning hyperparameters
Fine-tuning was performed with `simpletransformers`. Beforehand a brief hyperparameter optimisation was performed and the presumed optimal hyperparameters are:
```python
model_args = {
"num_train_epochs": 14,
"learning_rate": 1e-5,
"train_batch_size": 21,
}
```
## Performance
The same pipeline was run with two other transformer models and `fasttext` for comparison. Accuracy and macro F1 score were recorded for each of the 6 fine-tuning sessions and post festum analyzed.
| model | average accuracy | average macro F1|
|---|---|---|
|sloberta-frenk-hate|0.7785|0.7764|
|EMBEDDIA/crosloengual-bert |0.7616|0.7585|
|xlm-roberta-base |0.686|0.6827|
|fasttext|0.709 |0.701 |
From recorded accuracies and macro F1 scores p-values were also calculated:
Comparison with `crosloengual-bert`:
| test | accuracy p-value | macro F1 p-value|
| --- | --- | --- |
|Wilcoxon|0.00781|0.00781|
|Mann Whithney U test|0.00163|0.00108|
|Student t-test |0.000101|3.95e-05|
Comparison with `xlm-roberta-base`:
| test | accuracy p-value | macro F1 p-value|
| --- | --- | --- |
|Wilcoxon|0.00781|0.00781|
|Mann Whithney U test|0.00108|0.00108|
|Student t-test |9.46e-11|6.94e-11|
## Use examples
```python
from simpletransformers.classification import ClassificationModel
model_args = {
"num_train_epochs": 6,
"learning_rate": 3e-6,
"train_batch_size": 69}
model = ClassificationModel(
"camembert", "5roop/sloberta-frenk-hate", use_cuda=True,
args=model_args
)
predictions, logit_output = model.predict(["Silva, ti si grda in neprijazna", "Naša hiša ima dimnik"])
predictions
### Output:
### array([1, 0])
```
## Citation
If you use the model, please cite the following paper on which the original model is based:
```
@article{DBLP:journals/corr/abs-1907-11692,
author = {Yinhan Liu and
Myle Ott and
Naman Goyal and
Jingfei Du and
Mandar Joshi and
Danqi Chen and
Omer Levy and
Mike Lewis and
Luke Zettlemoyer and
Veselin Stoyanov},
title = {RoBERTa: {A} Robustly Optimized {BERT} Pretraining Approach},
journal = {CoRR},
volume = {abs/1907.11692},
year = {2019},
url = {http://arxiv.org/abs/1907.11692},
archivePrefix = {arXiv},
eprint = {1907.11692},
timestamp = {Thu, 01 Aug 2019 08:59:33 +0200},
biburl = {https://dblp.org/rec/journals/corr/abs-1907-11692.bib},
bibsource = {dblp computer science bibliography, https://dblp.org}
}
```
and the dataset used for fine-tuning:
```
@misc{ljubešić2019frenk,
title={The FRENK Datasets of Socially Unacceptable Discourse in Slovene and English},
author={Nikola Ljubešić and Darja Fišer and Tomaž Erjavec},
year={2019},
eprint={1906.02045},
archivePrefix={arXiv},
primaryClass={cs.CL},
url={https://arxiv.org/abs/1906.02045}
}
```
|
{"language": "sl", "license": "cc-by-sa-4.0", "tags": ["text-classification", "hate-speech"], "widget": [{"text": "Silva, ti si grda in neprijazna"}]}
|
classla/sloberta-frenk-hate
| null |
[
"transformers",
"pytorch",
"safetensors",
"camembert",
"text-classification",
"hate-speech",
"sl",
"arxiv:1907.11692",
"arxiv:1906.02045",
"license:cc-by-sa-4.0",
"autotrain_compatible",
"endpoints_compatible",
"region:us"
] | null |
2022-03-02T23:29:05+00:00
|
null |
transformers
|
{}
|
classla/swissbert-geo
| null |
[
"transformers",
"pytorch",
"bert",
"endpoints_compatible",
"region:us"
] | null |
2022-03-02T23:29:05+00:00
|
|
automatic-speech-recognition
|
transformers
|
# wav2vec2-xls-r-parlaspeech-hr
This model for Croatian ASR is based on the [facebook/wav2vec2-xls-r-300m model](https://huggingface.co/facebook/wav2vec2-xls-r-300m) and was fine-tuned with 300 hours of recordings and transcripts from the ASR Croatian parliament dataset [ParlaSpeech-HR v1.0](http://hdl.handle.net/11356/1494).
If you use this model, please cite the following paper:
Nikola Ljubešić, Danijel Koržinek, Peter Rupnik, Ivo-Pavao Jazbec. ParlaSpeech-HR -- a freely available ASR dataset for Croatian bootstrapped from the ParlaMint corpus. http://www.lrec-conf.org/proceedings/lrec2022/workshops/ParlaCLARINIII/pdf/2022.parlaclariniii-1.16.pdf
## Metrics
Evaluation is performed on the dev and test portions of the [ParlaSpeech-HR v1.0](http://hdl.handle.net/11356/1494) dataset.
|split|CER|WER|
|---|---|---|
|dev|0.0335|0.1046|
|test|0.0234|0.0761|
There are multiple models available, and in terms of CER and WER, the best-performing model is [wav2vec2-large-slavic-parlaspeech-hr-lm](https://huggingface.co/classla/wav2vec2-large-slavic-parlaspeech-hr-lm).
## Usage in `transformers`
```python
from transformers import Wav2Vec2Processor, Wav2Vec2ForCTC
import soundfile as sf
import torch
import os
device = torch.device("cuda:0" if torch.cuda.is_available() else "cpu")
# load model and tokenizer
processor = Wav2Vec2Processor.from_pretrained(
"classla/wav2vec2-xls-r-parlaspeech-hr")
model = Wav2Vec2ForCTC.from_pretrained("classla/wav2vec2-xls-r-parlaspeech-hr")
# download the example wav files:
os.system("wget https://huggingface.co/classla/wav2vec2-xls-r-parlaspeech-hr/raw/main/00020570a.flac.wav")
# read the wav file
speech, sample_rate = sf.read("00020570a.flac.wav")
input_values = processor(speech, sampling_rate=sample_rate, return_tensors="pt").input_values.to(device)
# remove the raw wav file
os.system("rm 00020570a.flac.wav")
# retrieve logits
logits = model.to(device)(input_values).logits
# take argmax and decode
predicted_ids = torch.argmax(logits, dim=-1)
transcription = processor.decode(predicted_ids[0]).lower()
# transcription: 'veliki broj poslovnih subjekata posluje sa minusom velik dio'
```
## Training hyperparameters
In fine-tuning, the following arguments were used:
| arg | value |
|-------------------------------|-------|
| `per_device_train_batch_size` | 16 |
| `gradient_accumulation_steps` | 4 |
| `num_train_epochs` | 8 |
| `learning_rate` | 3e-4 |
| `warmup_steps` | 500 |
|
{"language": "hr", "tags": ["audio", "automatic-speech-recognition", "parlaspeech"], "datasets": ["parlaspeech-hr"], "widget": [{"example_title": "example 1", "src": "https://huggingface.co/classla/wav2vec2-xls-r-parlaspeech-hr/raw/main/1800.m4a"}, {"example_title": "example 2", "src": "https://huggingface.co/classla/wav2vec2-xls-r-parlaspeech-hr/raw/main/00020578b.flac.wav"}]}
|
classla/wav2vec2-xls-r-parlaspeech-hr
| null |
[
"transformers",
"pytorch",
"safetensors",
"wav2vec2",
"automatic-speech-recognition",
"audio",
"parlaspeech",
"hr",
"dataset:parlaspeech-hr",
"endpoints_compatible",
"region:us"
] | null |
2022-03-02T23:29:05+00:00
|
text2text-generation
|
transformers
|
{}
|
claudelkros/T5_french_wiki_summarizer
| null |
[
"transformers",
"pytorch",
"jax",
"t5",
"text2text-generation",
"autotrain_compatible",
"endpoints_compatible",
"text-generation-inference",
"region:us"
] | null |
2022-03-02T23:29:05+00:00
|
|
null |
transformers
|
{}
|
claudelkros/bert-base-french
| null |
[
"transformers",
"pytorch",
"endpoints_compatible",
"region:us"
] | null |
2022-03-02T23:29:05+00:00
|
|
text-generation
|
transformers
|
# hiccupBot medium GPT
|
{"tags": ["conversational"]}
|
clayfox/DialoGPT-medium-Hiccup
| null |
[
"transformers",
"pytorch",
"gpt2",
"text-generation",
"conversational",
"autotrain_compatible",
"endpoints_compatible",
"text-generation-inference",
"region:us"
] | null |
2022-03-02T23:29:05+00:00
|
text-generation
|
transformers
|
# HiccupBot DialoGPT Model
|
{"tags": ["conversational"]}
|
clayfox/DialoGPT-small-Hiccup
| null |
[
"transformers",
"pytorch",
"gpt2",
"text-generation",
"conversational",
"autotrain_compatible",
"endpoints_compatible",
"text-generation-inference",
"region:us"
] | null |
2022-03-02T23:29:05+00:00
|
token-classification
|
transformers
|
{}
|
clee7/layoutlm-finetune-sroie
| null |
[
"transformers",
"pytorch",
"bert",
"token-classification",
"autotrain_compatible",
"endpoints_compatible",
"region:us"
] | null |
2022-03-02T23:29:05+00:00
|
|
text-classification
|
transformers
|
# Model Trained Using AutoNLP
- Problem type: Binary Classification
- Model ID: 2101779
## Validation Metrics
- Loss: 0.282466858625412
- Accuracy: 1.0
- Precision: 1.0
- Recall: 1.0
- AUC: 1.0
- F1: 1.0
## Usage
You can use cURL to access this model:
```
$ curl -X POST -H "Authorization: Bearer YOUR_API_KEY" -H "Content-Type: application/json" -d '{"inputs": "I love AutoNLP"}' https://api-inference.huggingface.co/models/clem/autonlp-test3-2101779
```
Or Python API:
```
from transformers import AutoModelForSequenceClassification, AutoTokenizer
model = AutoModelForSequenceClassification.from_pretrained("clem/autonlp-test3-2101779", use_auth_token=True)
tokenizer = AutoTokenizer.from_pretrained("clem/autonlp-test3-2101779", use_auth_token=True)
inputs = tokenizer("I love AutoNLP", return_tensors="pt")
outputs = model(**inputs)
```
|
{"language": "en", "tags": "autonlp", "datasets": ["clem/autonlp-data-test3"], "widget": [{"text": "I love AutoNLP \ud83e\udd17"}]}
|
clem/autonlp-test3-2101779
| null |
[
"transformers",
"pytorch",
"bert",
"text-classification",
"autonlp",
"en",
"dataset:clem/autonlp-data-test3",
"autotrain_compatible",
"endpoints_compatible",
"region:us"
] | null |
2022-03-02T23:29:05+00:00
|
text-classification
|
transformers
|
# Model Trained Using AutoNLP
- Problem type: Binary Classification
- Model ID: 2101782
## Validation Metrics
- Loss: 0.015991805121302605
- Accuracy: 1.0
- Precision: 1.0
- Recall: 1.0
- AUC: 1.0
- F1: 1.0
## Usage
You can use cURL to access this model:
```
$ curl -X POST -H "Authorization: Bearer YOUR_API_KEY" -H "Content-Type: application/json" -d '{"inputs": "I love AutoNLP"}' https://api-inference.huggingface.co/models/clem/autonlp-test3-2101782
```
Or Python API:
```
from transformers import AutoModelForSequenceClassification, AutoTokenizer
model = AutoModelForSequenceClassification.from_pretrained("clem/autonlp-test3-2101782", use_auth_token=True)
tokenizer = AutoTokenizer.from_pretrained("clem/autonlp-test3-2101782", use_auth_token=True)
inputs = tokenizer("I love AutoNLP", return_tensors="pt")
outputs = model(**inputs)
```
|
{"language": "en", "tags": "autonlp", "datasets": ["clem/autonlp-data-test3"], "widget": [{"text": "I love AutoNLP \ud83e\udd17"}]}
|
clem/autonlp-test3-2101782
| null |
[
"transformers",
"pytorch",
"bert",
"text-classification",
"autonlp",
"en",
"dataset:clem/autonlp-data-test3",
"autotrain_compatible",
"endpoints_compatible",
"region:us"
] | null |
2022-03-02T23:29:05+00:00
|
text-classification
|
transformers
|
# Model Trained Using AutoNLP
- Problem type: Binary Classification Urgent/Not Urgent
## Validation Metrics
- Loss: 0.08956164121627808
- Accuracy: 1.0
- Precision: 1.0
- Recall: 1.0
- AUC: 1.0
- F1: 1.0
## Usage
You can use cURL to access this model:
```
$ curl -X POST -H "Authorization: Bearer YOUR_API_KEY" -H "Content-Type: application/json" -d '{"inputs": "I love AutoNLP"}' https://api-inference.huggingface.co/models/clem/autonlp-test3-2101787
```
Or Python API:
```
from transformers import AutoModelForSequenceClassification, AutoTokenizer
model = AutoModelForSequenceClassification.from_pretrained("clem/autonlp-test3-2101787", use_auth_token=True)
tokenizer = AutoTokenizer.from_pretrained("clem/autonlp-test3-2101787", use_auth_token=True)
inputs = tokenizer("I love AutoNLP", return_tensors="pt")
outputs = model(**inputs)
```
|
{"language": "en", "tags": "autonlp", "datasets": ["clem/autonlp-data-test3"], "widget": [{"text": "this can wait"}]}
|
clem/autonlp-test3-2101787
| null |
[
"transformers",
"pytorch",
"distilbert",
"text-classification",
"autonlp",
"en",
"dataset:clem/autonlp-data-test3",
"autotrain_compatible",
"endpoints_compatible",
"region:us"
] | null |
2022-03-02T23:29:05+00:00
|
null | null |
{}
|
clem/bert_portuguese
| null |
[
"region:us"
] | null |
2022-03-02T23:29:05+00:00
|
|
null | null |
{}
|
cleoisstupid/cleobot
| null |
[
"region:us"
] | null |
2022-03-02T23:29:05+00:00
|
|
text-classification
|
transformers
|
# Model Card for distilroberta-base-climate-commitment
## Model Description
This is the fine-tuned ClimateBERT language model with a classification head for classifying climate-related paragraphs into paragraphs being about climate commitments and actions and paragraphs not being about climate commitments and actions.
Using the [climatebert/distilroberta-base-climate-f](https://huggingface.co/climatebert/distilroberta-base-climate-f) language model as starting point, the distilroberta-base-climate-commitment model is fine-tuned on our [climatebert/climate_commitments_actions](https://huggingface.co/climatebert/climate_commitments_actions) dataset.
*Note: This model is trained on paragraphs. It may not perform well on sentences.*
## Citation Information
```bibtex
@techreport{bingler2023cheaptalk,
title={How Cheap Talk in Climate Disclosures Relates to Climate Initiatives, Corporate Emissions, and Reputation Risk},
author={Bingler, Julia and Kraus, Mathias and Leippold, Markus and Webersinke, Nicolas},
type={Working paper},
institution={Available at SSRN 3998435},
year={2023}
}
```
## How to Get Started With the Model
You can use the model with a pipeline for text classification:
```python
from transformers import AutoModelForSequenceClassification, AutoTokenizer, pipeline
from transformers.pipelines.pt_utils import KeyDataset
import datasets
from tqdm.auto import tqdm
dataset_name = "climatebert/climate_commitments_actions"
model_name = "climatebert/distilroberta-base-climate-commitment"
# If you want to use your own data, simply load them as 🤗 Datasets dataset, see https://huggingface.co/docs/datasets/loading
dataset = datasets.load_dataset(dataset_name, split="test")
model = AutoModelForSequenceClassification.from_pretrained(model_name)
tokenizer = AutoTokenizer.from_pretrained(model_name, max_len=512)
pipe = pipeline("text-classification", model=model, tokenizer=tokenizer, device=0)
# See https://huggingface.co/docs/transformers/main_classes/pipelines#transformers.pipeline
for out in tqdm(pipe(KeyDataset(dataset, "text"), padding=True, truncation=True)):
print(out)
```
|
{"language": ["en"], "license": "apache-2.0", "datasets": ["climatebert/climate_commitments_actions"], "metrics": ["accuracy"]}
|
climatebert/distilroberta-base-climate-commitment
| null |
[
"transformers",
"pytorch",
"safetensors",
"roberta",
"text-classification",
"en",
"dataset:climatebert/climate_commitments_actions",
"license:apache-2.0",
"autotrain_compatible",
"endpoints_compatible",
"region:us"
] | null |
2022-03-02T23:29:05+00:00
|
fill-mask
|
transformers
|
# Model Card for distilroberta-base-climate-d-s
## Model Description
This is the ClimateBERT language model based on the DIV-SELECT and SIM-SELECT sample selection strategy.
*Note: We generally recommend choosing the [distilroberta-base-climate-f](https://huggingface.co/climatebert/distilroberta-base-climate-f) language model over this language model (unless you have good reasons not to).*
Using the [DistilRoBERTa](https://huggingface.co/distilroberta-base) model as starting point, the ClimateBERT Language Model is additionally pre-trained on a text corpus comprising climate-related research paper abstracts, corporate and general news and reports from companies. The underlying methodology can be found in our [language model research paper](https://arxiv.org/abs/2110.12010).
## Climate performance model card
| distilroberta-base-climate-d-s | |
|--------------------------------------------------------------------------|----------------|
| 1. Is the resulting model publicly available? | Yes |
| 2. How much time does the training of the final model take? | 48 hours |
| 3. How much time did all experiments take (incl. hyperparameter search)? | 350 hours |
| 4. What was the power of GPU and CPU? | 0.7 kW |
| 5. At which geo location were the computations performed? | Germany |
| 6. What was the energy mix at the geo location? | 470 gCO2eq/kWh |
| 7. How much CO2eq was emitted to train the final model? | 15.79 kg |
| 8. How much CO2eq was emitted for all experiments? | 115.15 kg |
| 9. What is the average CO2eq emission for the inference of one sample? | 0.62 mg |
| 10. Which positive environmental impact can be expected from this work? | This work can be categorized as a building block tools following Jin et al (2021). It supports the training of NLP models in the field of climate change and, thereby, have a positive environmental impact in the future. |
| 11. Comments | Block pruning could decrease CO2eq emissions |
## Citation Information
```bibtex
@inproceedings{wkbl2022climatebert,
title={{ClimateBERT: A Pretrained Language Model for Climate-Related Text}},
author={Webersinke, Nicolas and Kraus, Mathias and Bingler, Julia and Leippold, Markus},
booktitle={Proceedings of AAAI 2022 Fall Symposium: The Role of AI in Responding to Climate Challenges},
year={2022},
doi={https://doi.org/10.48550/arXiv.2212.13631},
}
```
|
{"language": "en", "license": "apache-2.0", "tags": ["climate"]}
|
climatebert/distilroberta-base-climate-d-s
| null |
[
"transformers",
"pytorch",
"safetensors",
"roberta",
"fill-mask",
"climate",
"en",
"arxiv:2110.12010",
"license:apache-2.0",
"autotrain_compatible",
"endpoints_compatible",
"region:us"
] | null |
2022-03-02T23:29:05+00:00
|
fill-mask
|
transformers
|
# Model Card for distilroberta-base-climate-d
## Model Description
This is the ClimateBERT language model based on the DIV-SELECT sample selection strategy.
*Note: We generally recommend choosing the [distilroberta-base-climate-f](https://huggingface.co/climatebert/distilroberta-base-climate-f) language model over this language model (unless you have good reasons not to).*
Using the [DistilRoBERTa](https://huggingface.co/distilroberta-base) model as starting point, the ClimateBERT Language Model is additionally pre-trained on a text corpus comprising climate-related research paper abstracts, corporate and general news and reports from companies. The underlying methodology can be found in our [language model research paper](https://arxiv.org/abs/2110.12010).
## Climate performance model card
| distilroberta-base-climate-d | |
|--------------------------------------------------------------------------|----------------|
| 1. Is the resulting model publicly available? | Yes |
| 2. How much time does the training of the final model take? | 48 hours |
| 3. How much time did all experiments take (incl. hyperparameter search)? | 350 hours |
| 4. What was the power of GPU and CPU? | 0.7 kW |
| 5. At which geo location were the computations performed? | Germany |
| 6. What was the energy mix at the geo location? | 470 gCO2eq/kWh |
| 7. How much CO2eq was emitted to train the final model? | 15.79 kg |
| 8. How much CO2eq was emitted for all experiments? | 115.15 kg |
| 9. What is the average CO2eq emission for the inference of one sample? | 0.62 mg |
| 10. Which positive environmental impact can be expected from this work? | This work can be categorized as a building block tools following Jin et al (2021). It supports the training of NLP models in the field of climate change and, thereby, have a positive environmental impact in the future. |
| 11. Comments | Block pruning could decrease CO2eq emissions |
## Citation Information
```bibtex
@inproceedings{wkbl2022climatebert,
title={{ClimateBERT: A Pretrained Language Model for Climate-Related Text}},
author={Webersinke, Nicolas and Kraus, Mathias and Bingler, Julia and Leippold, Markus},
booktitle={Proceedings of AAAI 2022 Fall Symposium: The Role of AI in Responding to Climate Challenges},
year={2022},
doi={https://doi.org/10.48550/arXiv.2212.13631},
}
```
|
{"language": "en", "license": "apache-2.0"}
|
climatebert/distilroberta-base-climate-d
| null |
[
"transformers",
"pytorch",
"roberta",
"fill-mask",
"en",
"arxiv:2110.12010",
"license:apache-2.0",
"autotrain_compatible",
"endpoints_compatible",
"region:us"
] | null |
2022-03-02T23:29:05+00:00
|
text-classification
|
transformers
|
# Model Card for distilroberta-base-climate-detector
## Model Description
This is the fine-tuned ClimateBERT language model with a classification head for detecting climate-related paragraphs.
Using the [climatebert/distilroberta-base-climate-f](https://huggingface.co/climatebert/distilroberta-base-climate-f) language model as starting point, the distilroberta-base-climate-detector model is fine-tuned on our [climatebert/climate_detection](https://huggingface.co/climatebert/climate_detection) dataset.
*Note: This model is trained on paragraphs. It may not perform well on sentences.*
## Citation Information
```bibtex
@techreport{bingler2023cheaptalk,
title={How Cheap Talk in Climate Disclosures Relates to Climate Initiatives, Corporate Emissions, and Reputation Risk},
author={Bingler, Julia and Kraus, Mathias and Leippold, Markus and Webersinke, Nicolas},
type={Working paper},
institution={Available at SSRN 3998435},
year={2023}
}
```
## How to Get Started With the Model
You can use the model with a pipeline for text classification:
```python
from transformers import AutoModelForSequenceClassification, AutoTokenizer, pipeline
from transformers.pipelines.pt_utils import KeyDataset
import datasets
from tqdm.auto import tqdm
dataset_name = "climatebert/climate_detection"
model_name = "climatebert/distilroberta-base-climate-detector"
# If you want to use your own data, simply load them as 🤗 Datasets dataset, see https://huggingface.co/docs/datasets/loading
dataset = datasets.load_dataset(dataset_name, split="test")
model = AutoModelForSequenceClassification.from_pretrained(model_name)
tokenizer = AutoTokenizer.from_pretrained(model_name, max_len=512)
pipe = pipeline("text-classification", model=model, tokenizer=tokenizer, device=0)
# See https://huggingface.co/docs/transformers/main_classes/pipelines#transformers.pipeline
for out in tqdm(pipe(KeyDataset(dataset, "text"), padding=True, truncation=True)):
print(out)
```
|
{"language": ["en"], "license": "apache-2.0", "datasets": ["climatebert/climate_detection"], "metrics": ["accuracy"]}
|
climatebert/distilroberta-base-climate-detector
| null |
[
"transformers",
"pytorch",
"safetensors",
"roberta",
"text-classification",
"en",
"dataset:climatebert/climate_detection",
"license:apache-2.0",
"autotrain_compatible",
"endpoints_compatible",
"has_space",
"region:us"
] | null |
2022-03-02T23:29:05+00:00
|
fill-mask
|
transformers
|
# Model Card for distilroberta-base-climate-f
## Model Description
This is the ClimateBERT language model based on the FULL-SELECT sample selection strategy.
*Note: We generally recommend choosing this language model over those based on the other sample selection strategies (unless you have good reasons not to). This is also the only language model we will update from time to time.*
Using the [DistilRoBERTa](https://huggingface.co/distilroberta-base) model as starting point, the ClimateBERT Language Model is additionally pre-trained on a text corpus comprising climate-related research paper abstracts, corporate and general news and reports from companies. The underlying methodology can be found in our [language model research paper](https://arxiv.org/abs/2110.12010).
*Update September 2, 2022: Now additionally pre-trained on an even larger text corpus, comprising >2M paragraphs. If you are looking for the language model before the update (i.e. for reproducibility), just use an older commit like [6be4fbd](https://huggingface.co/climatebert/distilroberta-base-climate-f/tree/6be4fbd3fedfd78ccb3c730c1f166947fbc940ba).*
## Climate performance model card
| distilroberta-base-climate-f | |
|--------------------------------------------------------------------------|----------------|
| 1. Is the resulting model publicly available? | Yes |
| 2. How much time does the training of the final model take? | 48 hours |
| 3. How much time did all experiments take (incl. hyperparameter search)? | 350 hours |
| 4. What was the power of GPU and CPU? | 0.7 kW |
| 5. At which geo location were the computations performed? | Germany |
| 6. What was the energy mix at the geo location? | 470 gCO2eq/kWh |
| 7. How much CO2eq was emitted to train the final model? | 15.79 kg |
| 8. How much CO2eq was emitted for all experiments? | 115.15 kg |
| 9. What is the average CO2eq emission for the inference of one sample? | 0.62 mg |
| 10. Which positive environmental impact can be expected from this work? | This work can be categorized as a building block tools following Jin et al (2021). It supports the training of NLP models in the field of climate change and, thereby, have a positive environmental impact in the future. |
| 11. Comments | Block pruning could decrease CO2eq emissions |
## Citation Information
```bibtex
@inproceedings{wkbl2022climatebert,
title={{ClimateBERT: A Pretrained Language Model for Climate-Related Text}},
author={Webersinke, Nicolas and Kraus, Mathias and Bingler, Julia and Leippold, Markus},
booktitle={Proceedings of AAAI 2022 Fall Symposium: The Role of AI in Responding to Climate Challenges},
year={2022},
doi={https://doi.org/10.48550/arXiv.2212.13631},
}
```
|
{"language": "en", "license": "apache-2.0", "tags": ["climate"]}
|
climatebert/distilroberta-base-climate-f
| null |
[
"transformers",
"pytorch",
"safetensors",
"roberta",
"fill-mask",
"climate",
"en",
"arxiv:2110.12010",
"license:apache-2.0",
"autotrain_compatible",
"endpoints_compatible",
"has_space",
"region:us"
] | null |
2022-03-02T23:29:05+00:00
|
fill-mask
|
transformers
|
# Model Card for distilroberta-base-climate-s
## Model Description
This is the ClimateBERT language model based on the SIM-SELECT sample selection strategy.
*Note: We generally recommend choosing the [distilroberta-base-climate-f](https://huggingface.co/climatebert/distilroberta-base-climate-f) language model over this language model (unless you have good reasons not to).*
Using the [DistilRoBERTa](https://huggingface.co/distilroberta-base) model as starting point, the ClimateBERT Language Model is additionally pre-trained on a text corpus comprising climate-related research paper abstracts, corporate and general news and reports from companies. The underlying methodology can be found in our [language model research paper](https://arxiv.org/abs/2110.12010).
## Climate performance model card
| distilroberta-base-climate-s | |
|--------------------------------------------------------------------------|----------------|
| 1. Is the resulting model publicly available? | Yes |
| 2. How much time does the training of the final model take? | 48 hours |
| 3. How much time did all experiments take (incl. hyperparameter search)? | 350 hours |
| 4. What was the power of GPU and CPU? | 0.7 kW |
| 5. At which geo location were the computations performed? | Germany |
| 6. What was the energy mix at the geo location? | 470 gCO2eq/kWh |
| 7. How much CO2eq was emitted to train the final model? | 15.79 kg |
| 8. How much CO2eq was emitted for all experiments? | 115.15 kg |
| 9. What is the average CO2eq emission for the inference of one sample? | 0.62 mg |
| 10. Which positive environmental impact can be expected from this work? | This work can be categorized as a building block tools following Jin et al (2021). It supports the training of NLP models in the field of climate change and, thereby, have a positive environmental impact in the future. |
| 11. Comments | Block pruning could decrease CO2eq emissions |
## Citation Information
```bibtex
@inproceedings{wkbl2022climatebert,
title={{ClimateBERT: A Pretrained Language Model for Climate-Related Text}},
author={Webersinke, Nicolas and Kraus, Mathias and Bingler, Julia and Leippold, Markus},
booktitle={Proceedings of AAAI 2022 Fall Symposium: The Role of AI in Responding to Climate Challenges},
year={2022},
doi={https://doi.org/10.48550/arXiv.2212.13631},
}
```
|
{"language": "en", "license": "apache-2.0"}
|
climatebert/distilroberta-base-climate-s
| null |
[
"transformers",
"pytorch",
"safetensors",
"roberta",
"fill-mask",
"en",
"arxiv:2110.12010",
"license:apache-2.0",
"autotrain_compatible",
"endpoints_compatible",
"region:us"
] | null |
2022-03-02T23:29:05+00:00
|
text-classification
|
transformers
|
# Model Card for distilroberta-base-climate-sentiment
## Model Description
This is the fine-tuned ClimateBERT language model with a classification head for classifying climate-related paragraphs into the climate-related sentiment classes opportunity, neutral, or risk.
Using the [climatebert/distilroberta-base-climate-f](https://huggingface.co/climatebert/distilroberta-base-climate-f) language model as starting point, the distilroberta-base-climate-sentiment model is fine-tuned on our [climatebert/climate_sentiment](https://huggingface.co/climatebert/climate_sentiment) dataset.
*Note: This model is trained on paragraphs. It may not perform well on sentences.*
## Citation Information
```bibtex
@techreport{bingler2023cheaptalk,
title={How Cheap Talk in Climate Disclosures Relates to Climate Initiatives, Corporate Emissions, and Reputation Risk},
author={Bingler, Julia and Kraus, Mathias and Leippold, Markus and Webersinke, Nicolas},
type={Working paper},
institution={Available at SSRN 3998435},
year={2023}
}
```
## How to Get Started With the Model
You can use the model with a pipeline for text classification:
```python
from transformers import AutoModelForSequenceClassification, AutoTokenizer, pipeline
from transformers.pipelines.pt_utils import KeyDataset
import datasets
from tqdm.auto import tqdm
dataset_name = "climatebert/climate_sentiment"
model_name = "climatebert/distilroberta-base-climate-sentiment"
# If you want to use your own data, simply load them as 🤗 Datasets dataset, see https://huggingface.co/docs/datasets/loading
dataset = datasets.load_dataset(dataset_name, split="test")
model = AutoModelForSequenceClassification.from_pretrained(model_name)
tokenizer = AutoTokenizer.from_pretrained(model_name, max_len=512)
pipe = pipeline("text-classification", model=model, tokenizer=tokenizer, device=0)
# See https://huggingface.co/docs/transformers/main_classes/pipelines#transformers.pipeline
for out in tqdm(pipe(KeyDataset(dataset, "text"), padding=True, truncation=True)):
print(out)
```
|
{"language": ["en"], "license": "apache-2.0", "datasets": ["climatebert/climate_sentiment"], "metrics": ["accuracy"]}
|
climatebert/distilroberta-base-climate-sentiment
| null |
[
"transformers",
"pytorch",
"safetensors",
"roberta",
"text-classification",
"en",
"dataset:climatebert/climate_sentiment",
"license:apache-2.0",
"autotrain_compatible",
"endpoints_compatible",
"region:us"
] | null |
2022-03-02T23:29:05+00:00
|
text-classification
|
transformers
|
# Model Card for distilroberta-base-climate-specificity
## Model Description
This is the fine-tuned ClimateBERT language model with a classification head for classifying climate-related paragraphs into specific and non-specific paragraphs.
Using the [climatebert/distilroberta-base-climate-f](https://huggingface.co/climatebert/distilroberta-base-climate-f) language model as starting point, the distilroberta-base-climate-specificity model is fine-tuned on our [climatebert/climate_specificity](https://huggingface.co/climatebert/climate_specificity) dataset.
*Note: This model is trained on paragraphs. It may not perform well on sentences.*
## Citation Information
```bibtex
@techreport{bingler2023cheaptalk,
title={How Cheap Talk in Climate Disclosures Relates to Climate Initiatives, Corporate Emissions, and Reputation Risk},
author={Bingler, Julia and Kraus, Mathias and Leippold, Markus and Webersinke, Nicolas},
type={Working paper},
institution={Available at SSRN 3998435},
year={2023}
}
```
## How to Get Started With the Model
You can use the model with a pipeline for text classification:
```python
from transformers import AutoModelForSequenceClassification, AutoTokenizer, pipeline
from transformers.pipelines.pt_utils import KeyDataset
import datasets
from tqdm.auto import tqdm
dataset_name = "climatebert/climate_specificity"
model_name = "climatebert/distilroberta-base-climate-specificity"
# If you want to use your own data, simply load them as 🤗 Datasets dataset, see https://huggingface.co/docs/datasets/loading
dataset = datasets.load_dataset(dataset_name, split="test")
model = AutoModelForSequenceClassification.from_pretrained(model_name)
tokenizer = AutoTokenizer.from_pretrained(model_name, max_len=512)
pipe = pipeline("text-classification", model=model, tokenizer=tokenizer, device=0)
# See https://huggingface.co/docs/transformers/main_classes/pipelines#transformers.pipeline
for out in tqdm(pipe(KeyDataset(dataset, "text"), padding=True, truncation=True)):
print(out)
```
|
{"language": ["en"], "license": "apache-2.0", "tags": ["climate"], "datasets": ["climatebert/climate_specificity"], "metrics": ["accuracy"]}
|
climatebert/distilroberta-base-climate-specificity
| null |
[
"transformers",
"pytorch",
"safetensors",
"roberta",
"text-classification",
"climate",
"en",
"dataset:climatebert/climate_specificity",
"license:apache-2.0",
"autotrain_compatible",
"endpoints_compatible",
"has_space",
"region:us"
] | null |
2022-03-02T23:29:05+00:00
|
text-classification
|
transformers
|
# Model Card for distilroberta-base-climate-tcfd
## Model Description
This is the fine-tuned ClimateBERT language model with a classification head for classifying climate-related paragraphs into the four TCFD recommendation categories ([fsb-tcfd.org](https://www.fsb-tcfd.org)).
Using the [climatebert/distilroberta-base-climate-f](https://huggingface.co/climatebert/distilroberta-base-climate-f) language model as starting point, the distilroberta-base-climate-tcfd model is fine-tuned on our [climatebert/tcfd_recommendations](https://huggingface.co/climatebert/tcfd_recommendations) dataset using only the four recommendation categories (i.e., we remove the non-climate-related class from the dataset).
*Note: This model is trained on paragraphs. It may not perform well on sentences.*
## Citation Information
```bibtex
@techreport{bingler2023cheaptalk,
title={How Cheap Talk in Climate Disclosures Relates to Climate Initiatives, Corporate Emissions, and Reputation Risk},
author={Bingler, Julia and Kraus, Mathias and Leippold, Markus and Webersinke, Nicolas},
type={Working paper},
institution={Available at SSRN 3998435},
year={2023}
}
```
## How to Get Started With the Model
You can use the model with a pipeline for text classification:
```python
from transformers import AutoModelForSequenceClassification, AutoTokenizer, pipeline
from transformers.pipelines.pt_utils import KeyDataset
import datasets
from tqdm.auto import tqdm
dataset_name = "climatebert/tcfd_recommendations"
model_name = "climatebert/distilroberta-base-climate-tcfd"
# If you want to use your own data, simply load them as 🤗 Datasets dataset, see https://huggingface.co/docs/datasets/loading
dataset = datasets.load_dataset(dataset_name, split="test")
model = AutoModelForSequenceClassification.from_pretrained(model_name)
tokenizer = AutoTokenizer.from_pretrained(model_name, max_len=512)
pipe = pipeline("text-classification", model=model, tokenizer=tokenizer, device=0)
# See https://huggingface.co/docs/transformers/main_classes/pipelines#transformers.pipeline
for out in tqdm(pipe(KeyDataset(dataset, "text"), padding=True, truncation=True)):
print(out)
```
|
{"language": ["en"], "license": "apache-2.0", "tags": ["climate"], "datasets": ["climatebert/tcfd_recommendations"], "metrics": ["accuracy"]}
|
climatebert/distilroberta-base-climate-tcfd
| null |
[
"transformers",
"pytorch",
"safetensors",
"roberta",
"text-classification",
"climate",
"en",
"dataset:climatebert/tcfd_recommendations",
"license:apache-2.0",
"autotrain_compatible",
"endpoints_compatible",
"region:us"
] | null |
2022-03-02T23:29:05+00:00
|
fill-mask
|
transformers
|
{}
|
cling371/modeling_test
| null |
[
"transformers",
"pytorch",
"distilbert",
"fill-mask",
"autotrain_compatible",
"endpoints_compatible",
"region:us"
] | null |
2022-03-02T23:29:05+00:00
|
|
null |
transformers
|
{}
|
clip-italian/clip-it-margin
| null |
[
"transformers",
"jax",
"hybrid-clip",
"endpoints_compatible",
"region:us"
] | null |
2022-03-02T23:29:05+00:00
|
|
null |
transformers
|
# CLIP-Italian
CLIP Italian is a CLIP-like Model for Italian. The CLIP model (Contrastive Language–Image Pre-training) was developed by researchers at OpenAI and is able to efficiently learn visual concepts from natural language supervision.
We fine-tuned a competitive Italian CLIP model with only ~1.4 million Italian image-text pairs. This model is part of the [Flax/Jax Community Week](https://discuss.huggingface.co/t/open-to-the-community-community-week-using-jax-flax-for-nlp-cv/7104), organized by [HuggingFace](https://huggingface.co/) and TPU usage sponsored by Google.
## Training Data
We considered three main sources of data:
- [WIT](https://github.com/google-research-datasets/wit)
- [MSCOCO-IT](https://github.com/crux82/mscoco-it)
- [Conceptual Captions](https://ai.google.com/research/ConceptualCaptions/)
## Training Procedure
Preprocessing, hardware used, hyperparameters...
## Evaluation Performance
## Limitations
## Usage
## Team members
- Federico Bianchi ([vinid](https://huggingface.co/vinid))
- Raphael Pisoni ([4rtemi5](https://huggingface.co/4rtemi5))
- Giuseppe Attanasio ([g8a9](https://huggingface.co/g8a9))
- Silvia Terragni ([silviatti](https://huggingface.co/silviatti))
- Dario Balestri ([D3Reo](https://huggingface.co/D3Reo))
- Gabriele Sarti ([gsarti](https://huggingface.co/gsarti))
- Sri Lakshmi ([srisweet](https://huggingface.co/srisweet))
## Useful links
- [CLIP Blog post](https://openai.com/blog/clip/)
- [CLIP paper](https://arxiv.org/abs/2103.00020)
- [Community Week README](https://github.com/huggingface/transformers/blob/master/examples/research_projects/jax-projects/README.md)
- [Community Week channel](https://discord.com/channels/858019234139602994/859711887520038933)
- [Hybrid CLIP example scripts](https://github.com/huggingface/transformers/tree/master/examples/research_projects/jax-projects/hybrid_clip)
- [Model Repository](https://huggingface.co/clip-italian/clip-italian-final/)
|
{"language": "it", "tags": ["italian", "bert", "vit", "vision"], "datasets": ["wit", "ctl/conceptualCaptions", "mscoco-it"]}
|
clip-italian/clip-italian-final
| null |
[
"transformers",
"jax",
"hybrid-clip",
"italian",
"bert",
"vit",
"vision",
"it",
"dataset:wit",
"dataset:ctl/conceptualCaptions",
"dataset:mscoco-it",
"arxiv:2103.00020",
"endpoints_compatible",
"region:us"
] | null |
2022-03-02T23:29:05+00:00
|
feature-extraction
|
transformers
|
# Italian CLIP
Paper: [Contrastive Language-Image Pre-training for the Italian Language](https://arxiv.org/abs/2108.08688)
With a few tricks, we have been able to fine-tune a competitive Italian CLIP model with **only 1.4 million** training samples. Our Italian CLIP model is built upon the [Italian BERT](https://huggingface.co/dbmdz/bert-base-italian-xxl-cased) model provided by [dbmdz](https://huggingface.co/dbmdz) and the OpenAI [vision transformer](https://huggingface.co/openai/clip-vit-base-patch32).
Do you want to test our model right away? We got you covered! You just need to head to our [demo application](https://huggingface.co/spaces/clip-italian/clip-italian-demo).
The demo also contains all the details of the project, from training tricks to our most impressive results, and much more!
# Training data
We considered four main sources of data:
+ [WIT](https://github.com/google-research-datasets/wit) is an image-caption dataset collected from Wikipedia (see,
[Srinivasan et al., 2021](https://arxiv.org/pdf/2103.01913.pdf)).
+ [MSCOCO-IT](https://github.com/crux82/mscoco-it). This image-caption dataset comes from the work by [Scaiella et al., 2019](http://www.ai-lc.it/IJCoL/v5n2/IJCOL_5_2_3___scaiella_et_al.pdf).
+ [Conceptual Captions](https://ai.google.com/research/ConceptualCaptions/). This image-caption dataset comes from
the work by [Sharma et al., 2018](https://aclanthology.org/P18-1238.pdf).
+ [La Foto del Giorno](https://www.ilpost.it/foto-del-giorno/). This image-caption dataset is collected from [Il Post](https://www.ilpost.it/), a prominent Italian online newspaper.
We used better data augmentation, strategic training choices (we have way less data than the original CLIP paper), and backbone-freezing pre-training. For all the details on that, please refer to our [demo](https://huggingface.co/spaces/clip-italian/clip-italian-demo).
# Experiments
## Quantitative Evaluation
To better understand how well our clip-italian model works we run an experimental evaluation. Since this is the first clip-based model in Italian, we used the multilingual CLIP model as a comparison baseline.
### mCLIP
The multilingual CLIP (henceforth, mCLIP), is a model introduced by [Nils Reimers](https://www.sbert.net/docs/pretrained_models.html) in his
[sentence-transformer](https://www.sbert.net/index.html) library. mCLIP is based on a multilingual encoder
that was created through multilingual knowledge distillation (see [Reimers et al., 2020](https://aclanthology.org/2020.emnlp-main.365/)).
### Tasks
We selected two different tasks:
+ image-retrieval
+ zero-shot classification
### Reproducibiliy
Both experiments should be very easy to replicate, we share the two colab notebook we used to compute the two results
+ [Image Retrieval](https://colab.research.google.com/drive/1bLVwVKpAndpEDHqjzxVPr_9nGrSbuOQd?usp=sharing)
+ [ImageNet Zero Shot Evaluation](https://colab.research.google.com/drive/1zfWeVWY79XXH63Ci-pk8xxx3Vu_RRgW-?usp=sharing)
### Image Retrieval
This experiment is run against the MSCOCO-IT validation set (that we haven't used in training). Given in input
a caption, we search for the most similar image in the MSCOCO-IT validation set. As evaluation metrics
we use the MRR@K.
| MRR | CLIP-Italian | mCLIP |
| --------------- | ------------ |-------|
| MRR@1 | **0.3797** | 0.2874|
| MRR@5 | **0.5039** | 0.3957|
| MRR@10 | **0.5204** | 0.4129|
It is true that we used MSCOCO-IT in training, and this might give us an advantage. However the original CLIP model was trained
on 400million images (and some of them probably were from MSCOCO).
### Zero-shot image classification
This experiment replicates the original one run by OpenAI on zero-shot image classification on ImageNet.
To do this, we used DeepL to translate the image labels in ImageNet. We evaluate the models computing the accuracy at different levels.
| Accuracy | CLIP-Italian | mCLIP |
| --------------- | ------------ |-------|
| Accuracy@1 | **22.11** | 20.15 |
| Accuracy@5 | **43.69** | 36.57 |
| Accuracy@10 | **52.55** | 42.91 |
| Accuracy@100 | **81.08** | 67.11 |
Our results confirm that CLIP-Italian is very competitive and beats mCLIP on the two different task
we have been testing. Note, however, that our results are lower than those shown in the original OpenAI
paper (see, [Radford et al., 2021](https://arxiv.org/abs/2103.00020)). However, considering that our results are in line with those obtained by mCLIP we think that
the translated image labels might have had an impact on the final scores.
# Team members
- Federico Bianchi ([vinid](https://huggingface.co/vinid))
- Raphael Pisoni ([4rtemi5](https://huggingface.co/4rtemi5))
- Giuseppe Attanasio ([g8a9](https://huggingface.co/g8a9))
- Silvia Terragni ([silviatti](https://huggingface.co/silviatti))
- Dario Balestri ([D3Reo](https://huggingface.co/D3Reo))
- Gabriele Sarti ([gsarti](https://huggingface.co/gsarti))
- Sri Lakshmi ([srisweet](https://huggingface.co/srisweet))
|
{"language": "it", "license": "gpl-3.0", "tags": ["italian", "bert", "vit", "vision"], "datasets": ["wit", "ctl/conceptualCaptions", "mscoco-it"]}
|
clip-italian/clip-italian
| null |
[
"transformers",
"pytorch",
"jax",
"vision-text-dual-encoder",
"feature-extraction",
"italian",
"bert",
"vit",
"vision",
"it",
"dataset:wit",
"dataset:ctl/conceptualCaptions",
"dataset:mscoco-it",
"arxiv:2108.08688",
"arxiv:2103.01913",
"arxiv:2103.00020",
"license:gpl-3.0",
"endpoints_compatible",
"has_space",
"region:us"
] | null |
2022-03-02T23:29:05+00:00
|
feature-extraction
|
transformers
|
# CoNTACT
### Model description
<u>Co</u>ntextual <u>N</u>eural <u>T</u>ransformer <u>A</u>dapted to <u>C</u>OVID-19 <u>T</u>weets or **CoNTACT** is a Dutch RobBERT model (```pdelobelle/robbert-v2-dutch-base```) adapted to the domain of COVID-19 tweets. The model was developed at [CLiPS](https://www.uantwerpen.be/en/research-groups/clips/) by Jens Lemmens, Jens Van Nooten, Tim Kreutz and Walter Daelemans. A full description of the model, the data that was used and the experiments that were conducted can be found in this ArXiv preprint: https://arxiv.org/abs/2203.07362
### Intended use
The model was developed with the intention of achieving high results on NLP tasks involving Dutch social media messages related to COVID-19.
### How to use
CoNTACT should be fine-tuned on a downstream task. This can be achieved by referring to ```clips/contact``` in the ```--model_name_or_path``` argument in Huggingface/Transformers' example scripts, or by loading CoNTACT (as shown below) and fine-tuning it using your own code:
```
from transformers import AutoModel, AutoTokenizer
model = AutoModel.from_pretrained('clips/contact')
tokenizer = AutoTokenizer.from_pretrained('clips/contact')
...
```
### Training data
CoNTACT was trained on 2.8M Dutch tweets related to COVID-19 that were posted in 2021.
### Training Procedure
The model's pre-training phase was extended by performing Masked Language Modeling (MLM) on the training data described above. This was done for 4 epochs, using the largest possible batch size that fit working memory (32).
### Evaluation
The model was evaluated on two tasks using data from two social media platforms: Twitter and Facebook. Task 1 involved the binary classification of COVID-19 vaccine stance (hesitant vs. not hesitant), whereas task 2 consisted of the mulilabel, multiclass classification of arguments for vaccine hesitancy. CoNTACT outperformed out-of-the-box RobBERT in virtually all our experiments, and with statistical significance in most cases.
### How to cite
```
@misc{lemmens2022contact,
title={CoNTACT: A Dutch COVID-19 Adapted BERT for Vaccine Hesitancy and Argumentation Detection},
author={Jens Lemmens and Jens Van Nooten and Tim Kreutz and Walter Daelemans},
year={2022},
eprint={2203.07362},
archivePrefix={arXiv},
primaryClass={cs.CL}
}
```
|
{}
|
clips/contact
| null |
[
"transformers",
"pytorch",
"roberta",
"feature-extraction",
"arxiv:2203.07362",
"endpoints_compatible",
"region:us"
] | null |
2022-03-02T23:29:05+00:00
|
sentence-similarity
|
sentence-transformers
|
# MFAQ
We present a multilingual FAQ retrieval model trained on the [MFAQ dataset](https://huggingface.co/datasets/clips/mfaq), it ranks candidate answers according to a given question.
## Installation
```
pip install sentence-transformers transformers
```
## Usage
You can use MFAQ with sentence-transformers or directly with a HuggingFace model.
In both cases, questions need to be prepended with `<Q>`, and answers with `<A>`.
#### Sentence Transformers
```python
from sentence_transformers import SentenceTransformer
question = "<Q>How many models can I host on HuggingFace?"
answer_1 = "<A>All plans come with unlimited private models and datasets."
answer_2 = "<A>AutoNLP is an automatic way to train and deploy state-of-the-art NLP models, seamlessly integrated with the Hugging Face ecosystem."
answer_3 = "<A>Based on how much training data and model variants are created, we send you a compute cost and payment link - as low as $10 per job."
model = SentenceTransformer('clips/mfaq')
embeddings = model.encode([question, answer_1, answer_3, answer_3])
print(embeddings)
```
#### HuggingFace Transformers
```python
from transformers import AutoTokenizer, AutoModel
import torch
def mean_pooling(model_output, attention_mask):
token_embeddings = model_output[0] #First element of model_output contains all token embeddings
input_mask_expanded = attention_mask.unsqueeze(-1).expand(token_embeddings.size()).float()
return torch.sum(token_embeddings * input_mask_expanded, 1) / torch.clamp(input_mask_expanded.sum(1), min=1e-9)
question = "<Q>How many models can I host on HuggingFace?"
answer_1 = "<A>All plans come with unlimited private models and datasets."
answer_2 = "<A>AutoNLP is an automatic way to train and deploy state-of-the-art NLP models, seamlessly integrated with the Hugging Face ecosystem."
answer_3 = "<A>Based on how much training data and model variants are created, we send you a compute cost and payment link - as low as $10 per job."
tokenizer = AutoTokenizer.from_pretrained('clips/mfaq')
model = AutoModel.from_pretrained('clips/mfaq')
# Tokenize sentences
encoded_input = tokenizer([question, answer_1, answer_3, answer_3], padding=True, truncation=True, return_tensors='pt')
# Compute token embeddings
with torch.no_grad():
model_output = model(**encoded_input)
# Perform pooling. In this case, max pooling.
sentence_embeddings = mean_pooling(model_output, encoded_input['attention_mask'])
```
## Training
You can find the training script for the model [here](https://github.com/clips/mfaq).
## People
This model was developed by [Maxime De Bruyn](https://www.linkedin.com/in/maximedebruyn/), Ehsan Lotfi, Jeska Buhmann and Walter Daelemans.
## Citation information
```
@misc{debruyn2021mfaq,
title={MFAQ: a Multilingual FAQ Dataset},
author={Maxime De Bruyn and Ehsan Lotfi and Jeska Buhmann and Walter Daelemans},
year={2021},
eprint={2109.12870},
archivePrefix={arXiv},
primaryClass={cs.CL}
}
```
|
{"language": ["cs", "da", "de", "en", "es", "fi", "fr", "he", "hr", "hu", "id", "it", "nl", "no", "pl", "pt", "ro", "ru", "sv", "tr", "vi"], "license": "apache-2.0", "tags": ["sentence-transformers", "feature-extraction", "sentence-similarity", "transformers"], "datasets": ["clips/mfaq"], "pipeline_tag": "sentence-similarity", "widget": {"source_sentence": "<Q>How many models can I host on HuggingFace?", "sentences": ["<A>All plans come with unlimited private models and datasets.", "<A>AutoNLP is an automatic way to train and deploy state-of-the-art NLP models, seamlessly integrated with the Hugging Face ecosystem.", "<A>Based on how much training data and model variants are created, we send you a compute cost and payment link - as low as $10 per job."]}}
|
clips/mfaq
| null |
[
"sentence-transformers",
"pytorch",
"tf",
"xlm-roberta",
"feature-extraction",
"sentence-similarity",
"transformers",
"cs",
"da",
"de",
"en",
"es",
"fi",
"fr",
"he",
"hr",
"hu",
"id",
"it",
"nl",
"no",
"pl",
"pt",
"ro",
"ru",
"sv",
"tr",
"vi",
"dataset:clips/mfaq",
"arxiv:2109.12870",
"license:apache-2.0",
"endpoints_compatible",
"has_space",
"region:us"
] | null |
2022-03-02T23:29:05+00:00
|
null |
transformers
|
## albert_chinese_small
### Overview
**Language model:** albert-small
**Model size:** 18.5M
**Language:** Chinese
**Training data:** [CLUECorpusSmall](https://github.com/CLUEbenchmark/CLUECorpus2020)
**Eval data:** [CLUE dataset](https://github.com/CLUEbenchmark/CLUE)
### Results
For results on downstream tasks like text classification, please refer to [this repository](https://github.com/CLUEbenchmark/CLUE).
### Usage
**NOTE:**Since sentencepiece is not used in `albert_chinese_small` model, you have to call **BertTokenizer** instead of AlbertTokenizer !!!
```
import torch
from transformers import BertTokenizer, AlbertModel
tokenizer = BertTokenizer.from_pretrained("clue/albert_chinese_small")
albert = AlbertModel.from_pretrained("clue/albert_chinese_small")
```
### About CLUE benchmark
Organization of Language Understanding Evaluation benchmark for Chinese: tasks & datasets, baselines, pre-trained Chinese models, corpus and leaderboard.
Github: https://github.com/CLUEbenchmark
Website: https://www.cluebenchmarks.com/
|
{"language": "zh"}
|
clue/albert_chinese_small
| null |
[
"transformers",
"pytorch",
"albert",
"zh",
"endpoints_compatible",
"region:us"
] | null |
2022-03-02T23:29:05+00:00
|
null |
transformers
|
## albert_chinese_tiny
### Overview
**Language model:** albert-tiny
**Model size:** 16M
**Language:** Chinese
**Training data:** [CLUECorpusSmall](https://github.com/CLUEbenchmark/CLUECorpus2020)
**Eval data:** [CLUE dataset](https://github.com/CLUEbenchmark/CLUE)
### Results
For results on downstream tasks like text classification, please refer to [this repository](https://github.com/CLUEbenchmark/CLUE).
### Usage
**NOTE:**Since sentencepiece is not used in `albert_chinese_tiny` model, you have to call **BertTokenizer** instead of AlbertTokenizer !!!
```
import torch
from transformers import BertTokenizer, AlbertModel
tokenizer = BertTokenizer.from_pretrained("clue/albert_chinese_tiny")
albert = AlbertModel.from_pretrained("clue/albert_chinese_tiny")
```
### About CLUE benchmark
Organization of Language Understanding Evaluation benchmark for Chinese: tasks & datasets, baselines, pre-trained Chinese models, corpus and leaderboard.
Github: https://github.com/CLUEbenchmark
Website: https://www.cluebenchmarks.com/
|
{"language": "zh"}
|
clue/albert_chinese_tiny
| null |
[
"transformers",
"pytorch",
"albert",
"zh",
"endpoints_compatible",
"region:us"
] | null |
2022-03-02T23:29:05+00:00
|
null |
transformers
|
# Introduction
This model was trained on TPU and the details are as follows:
## Model
##
| Model_name | params | size | Training_corpus | Vocab |
| :------------------------------------------ | :----- | :------- | :----------------- | :-----------: |
| **`RoBERTa-tiny-clue`** <br/>Super_small_model | 7.5M | 28.3M | **CLUECorpus2020** | **CLUEVocab** |
| **`RoBERTa-tiny-pair`** <br/>Super_small_sentence_pair_model | 7.5M | 28.3M | **CLUECorpus2020** | **CLUEVocab** |
| **`RoBERTa-tiny3L768-clue`** <br/>small_model | 38M | 110M | **CLUECorpus2020** | **CLUEVocab** |
| **`RoBERTa-tiny3L312-clue`** <br/>small_model | <7.5M | 24M | **CLUECorpus2020** | **CLUEVocab** |
| **`RoBERTa-large-clue`** <br/> Large_model | 290M | 1.20G | **CLUECorpus2020** | **CLUEVocab** |
| **`RoBERTa-large-pair`** <br/>Large_sentence_pair_model | 290M | 1.20G | **CLUECorpus2020** | **CLUEVocab** |
### Usage
With the help of[Huggingface-Transformers 2.5.1](https://github.com/huggingface/transformers), you could use these model as follows
```
tokenizer = BertTokenizer.from_pretrained("MODEL_NAME")
model = BertModel.from_pretrained("MODEL_NAME")
```
`MODEL_NAME`：
| Model_NAME | MODEL_LINK |
| -------------------------- | ------------------------------------------------------------ |
| **RoBERTa-tiny-clue** | [`clue/roberta_chinese_clue_tiny`](https://huggingface.co/clue/roberta_chinese_clue_tiny) |
| **RoBERTa-tiny-pair** | [`clue/roberta_chinese_pair_tiny`](https://huggingface.co/clue/roberta_chinese_pair_tiny) |
| **RoBERTa-tiny3L768-clue** | [`clue/roberta_chinese_3L768_clue_tiny`](https://huggingface.co/clue/roberta_chinese_3L768_clue_tiny) |
| **RoBERTa-tiny3L312-clue** | [`clue/roberta_chinese_3L312_clue_tiny`](https://huggingface.co/clue/roberta_chinese_3L312_clue_tiny) |
| **RoBERTa-large-clue** | [`clue/roberta_chinese_clue_large`](https://huggingface.co/clue/roberta_chinese_clue_large) |
| **RoBERTa-large-pair** | [`clue/roberta_chinese_pair_large`](https://huggingface.co/clue/roberta_chinese_pair_large) |
## Details
Please read <a href='https://arxiv.org/pdf/2003.01355'>https://arxiv.org/pdf/2003.01355.
Please visit our repository: https://github.com/CLUEbenchmark/CLUEPretrainedModels.git
|
{"language": "zh"}
|
clue/roberta_chinese_3L312_clue_tiny
| null |
[
"transformers",
"pytorch",
"jax",
"roberta",
"zh",
"arxiv:2003.01355",
"endpoints_compatible",
"region:us"
] | null |
2022-03-02T23:29:05+00:00
|
null |
transformers
|
{}
|
clue/roberta_chinese_3L768_clue_tiny
| null |
[
"transformers",
"pytorch",
"jax",
"roberta",
"endpoints_compatible",
"region:us"
] | null |
2022-03-02T23:29:05+00:00
|
|
null |
transformers
|
## roberta_chinese_base
### Overview
**Language model:** roberta-base
**Model size:** 392M
**Language:** Chinese
**Training data:** [CLUECorpusSmall](https://github.com/CLUEbenchmark/CLUECorpus2020)
**Eval data:** [CLUE dataset](https://github.com/CLUEbenchmark/CLUE)
### Results
For results on downstream tasks like text classification, please refer to [this repository](https://github.com/CLUEbenchmark/CLUE).
### Usage
**NOTE:** You have to call **BertTokenizer** instead of RobertaTokenizer !!!
```
import torch
from transformers import BertTokenizer, BertModel
tokenizer = BertTokenizer.from_pretrained("clue/roberta_chinese_base")
roberta = BertModel.from_pretrained("clue/roberta_chinese_base")
```
### About CLUE benchmark
Organization of Language Understanding Evaluation benchmark for Chinese: tasks & datasets, baselines, pre-trained Chinese models, corpus and leaderboard.
Github: https://github.com/CLUEbenchmark
Website: https://www.cluebenchmarks.com/
|
{"language": "zh"}
|
clue/roberta_chinese_base
| null |
[
"transformers",
"pytorch",
"jax",
"roberta",
"zh",
"endpoints_compatible",
"region:us"
] | null |
2022-03-02T23:29:05+00:00
|
null |
transformers
|
{}
|
clue/roberta_chinese_clue_large
| null |
[
"transformers",
"pytorch",
"jax",
"roberta",
"endpoints_compatible",
"region:us"
] | null |
2022-03-02T23:29:05+00:00
|
|
null |
transformers
|
{}
|
clue/roberta_chinese_clue_tiny
| null |
[
"transformers",
"pytorch",
"jax",
"roberta",
"endpoints_compatible",
"has_space",
"region:us"
] | null |
2022-03-02T23:29:05+00:00
|
|
null |
transformers
|
## roberta_chinese_large
### Overview
**Language model:** roberta-large
**Model size:** 1.2G
**Language:** Chinese
**Training data:** [CLUECorpusSmall](https://github.com/CLUEbenchmark/CLUECorpus2020)
**Eval data:** [CLUE dataset](https://github.com/CLUEbenchmark/CLUE)
### Results
For results on downstream tasks like text classification, please refer to [this repository](https://github.com/CLUEbenchmark/CLUE).
### Usage
**NOTE:** You have to call **BertTokenizer** instead of RobertaTokenizer !!!
```
import torch
from transformers import BertTokenizer, BertModel
tokenizer = BertTokenizer.from_pretrained("clue/roberta_chinese_large")
roberta = BertModel.from_pretrained("clue/roberta_chinese_large")
```
### About CLUE benchmark
Organization of Language Understanding Evaluation benchmark for Chinese: tasks & datasets, baselines, pre-trained Chinese models, corpus and leaderboard.
Github: https://github.com/CLUEbenchmark
Website: https://www.cluebenchmarks.com/
|
{"language": "zh"}
|
clue/roberta_chinese_large
| null |
[
"transformers",
"pytorch",
"jax",
"roberta",
"zh",
"endpoints_compatible",
"region:us"
] | null |
2022-03-02T23:29:05+00:00
|
null |
transformers
|
{}
|
clue/roberta_chinese_pair_large
| null |
[
"transformers",
"pytorch",
"jax",
"roberta",
"endpoints_compatible",
"region:us"
] | null |
2022-03-02T23:29:05+00:00
|
|
null |
transformers
|
{}
|
clue/roberta_chinese_pair_tiny
| null |
[
"transformers",
"pytorch",
"jax",
"roberta",
"endpoints_compatible",
"region:us"
] | null |
2022-03-02T23:29:05+00:00
|
|
null |
transformers
|
## xlnet_chinese_large
### Overview
**Language model:** xlnet-large
**Model size:** 1.3G
**Language:** Chinese
**Training data:** [CLUECorpusSmall](https://github.com/CLUEbenchmark/CLUECorpus2020)
**Eval data:** [CLUE dataset](https://github.com/CLUEbenchmark/CLUE)
### Results
For results on downstream tasks like text classification, please refer to [this repository](https://github.com/CLUEbenchmark/CLUE).
### Usage
```
import torch
from transformers import XLNetTokenizer,XLNetModel
tokenizer = XLNetTokenizer.from_pretrained("clue/xlnet_chinese_large")
xlnet = XLNetModel.from_pretrained("clue/xlnet_chinese_large")
```
### About CLUE benchmark
Organization of Language Understanding Evaluation benchmark for Chinese: tasks & datasets, baselines, pre-trained Chinese models, corpus and leaderboard.
Github: https://github.com/CLUEbenchmark
Website: https://www.cluebenchmarks.com/
|
{"language": "zh"}
|
clue/xlnet_chinese_large
| null |
[
"transformers",
"pytorch",
"xlnet",
"zh",
"endpoints_compatible",
"region:us"
] | null |
2022-03-02T23:29:05+00:00
|
token-classification
|
transformers
|
{}
|
clulab/roberta-timex-semeval
| null |
[
"transformers",
"pytorch",
"jax",
"roberta",
"token-classification",
"autotrain_compatible",
"endpoints_compatible",
"region:us"
] | null |
2022-03-02T23:29:05+00:00
|
|
token-classification
|
transformers
|
DistilCamemBERT-NER
===================
We present DistilCamemBERT-NER, which is [DistilCamemBERT](https://huggingface.co/cmarkea/distilcamembert-base) fine-tuned for the NER (Named Entity Recognition) task for the French language. The work is inspired by [Jean-Baptiste/camembert-ner](https://huggingface.co/Jean-Baptiste/camembert-ner) based on the [CamemBERT](https://huggingface.co/camembert-base) model. The problem of the modelizations based on CamemBERT is at the scaling moment, for the production phase, for example. Indeed, inference cost can be a technological issue. To counteract this effect, we propose this modelization which **divides the inference time by two** with the same consumption power thanks to [DistilCamemBERT](https://huggingface.co/cmarkea/distilcamembert-base).
Dataset
-------
The dataset used is [wikiner_fr](https://huggingface.co/datasets/Jean-Baptiste/wikiner_fr), which represents ~170k sentences labeled in 5 categories :
* PER: personality ;
* LOC: location ;
* ORG: organization ;
* MISC: miscellaneous entities (movies title, books, etc.) ;
* O: background (Outside entity).
Evaluation results
------------------
| **class** | **precision (%)** | **recall (%)** | **f1 (%)** | **support (#sub-word)** |
| :------------: | :---------------: | :------------: | :--------: | :---------------------: |
| **global** | 98.17 | 98.19 | 98.18 | 378,776 |
| **PER** | 96.78 | 96.87 | 96.82 | 23,754 |
| **LOC** | 94.05 | 93.59 | 93.82 | 27,196 |
| **ORG** | 86.05 | 85.92 | 85.98 | 6,526 |
| **MISC** | 88.78 | 84.69 | 86.69 | 11,891 |
| **O** | 99.26 | 99.47 | 99.37 | 309,409 |
Benchmark
---------
This model performance is compared to 2 reference models (see below) with the metric f1 score. For the mean inference time measure, an AMD Ryzen 5 4500U @ 2.3GHz with 6 cores was used:
| **model** | **time (ms)** | **PER (%)** | **LOC (%)** | **ORG (%)** | **MISC (%)** | **O (%)** |
| :---------------------------------------------------------------------------------------------------------------: | :-----------: | :---------: | :---------: | :---------: | :-----------: | :-------: |
| [cmarkea/distilcamembert-base-ner](https://huggingface.co/cmarkea/distilcamembert-base-ner) | **43.44** | **96.82** | **93.82** | **85.98** | **86.69** | **99.37** |
| [Davlan/bert-base-multilingual-cased-ner-hrl](https://huggingface.co/Davlan/bert-base-multilingual-cased-ner-hrl) | 87.56 | 79.93 | 72.89 | 61.34 | n/a | 96.04 |
| [flair/ner-french](https://huggingface.co/flair/ner-french) | 314.96 | 82.91 | 76.17 | 70.96 | 76.29 | 97.65 |
How to use DistilCamemBERT-NER
------------------------------
```python
from transformers import pipeline
ner = pipeline(
task='ner',
model="cmarkea/distilcamembert-base-ner",
tokenizer="cmarkea/distilcamembert-base-ner",
aggregation_strategy="simple"
)
result = ner(
"Le Crédit Mutuel Arkéa est une banque Française, elle comprend le CMB "
"qui est une banque située en Bretagne et le CMSO qui est une banque "
"qui se situe principalement en Aquitaine. C'est sous la présidence de "
"Louis Lichou, dans les années 1980 que différentes filiales sont créées "
"au sein du CMB et forment les principales filiales du groupe qui "
"existent encore aujourd'hui (Federal Finance, Suravenir, Financo, etc.)."
)
result
[{'entity_group': 'ORG',
'score': 0.9974479,
'word': 'Crédit Mutuel Arkéa',
'start': 3,
'end': 22},
{'entity_group': 'LOC',
'score': 0.9000358,
'word': 'Française',
'start': 38,
'end': 47},
{'entity_group': 'ORG',
'score': 0.9788757,
'word': 'CMB',
'start': 66,
'end': 69},
{'entity_group': 'LOC',
'score': 0.99919766,
'word': 'Bretagne',
'start': 99,
'end': 107},
{'entity_group': 'ORG',
'score': 0.9594884,
'word': 'CMSO',
'start': 114,
'end': 118},
{'entity_group': 'LOC',
'score': 0.99935514,
'word': 'Aquitaine',
'start': 169,
'end': 178},
{'entity_group': 'PER',
'score': 0.99911094,
'word': 'Louis Lichou',
'start': 208,
'end': 220},
{'entity_group': 'ORG',
'score': 0.96226394,
'word': 'CMB',
'start': 291,
'end': 294},
{'entity_group': 'ORG',
'score': 0.9983959,
'word': 'Federal Finance',
'start': 374,
'end': 389},
{'entity_group': 'ORG',
'score': 0.9984454,
'word': 'Suravenir',
'start': 391,
'end': 400},
{'entity_group': 'ORG',
'score': 0.9985084,
'word': 'Financo',
'start': 402,
'end': 409}]
```
### Optimum + ONNX
```python
from optimum.onnxruntime import ORTModelForTokenClassification
from transformers import AutoTokenizer, pipeline
HUB_MODEL = "cmarkea/distilcamembert-base-nli"
tokenizer = AutoTokenizer.from_pretrained(HUB_MODEL)
model = ORTModelForTokenClassification.from_pretrained(HUB_MODEL)
onnx_qa = pipeline("token-classification", model=model, tokenizer=tokenizer)
# Quantized onnx model
quantized_model = ORTModelForTokenClassification.from_pretrained(
HUB_MODEL, file_name="model_quantized.onnx"
)
```
Citation
--------
```bibtex
@inproceedings{delestre:hal-03674695,
TITLE = {{DistilCamemBERT : une distillation du mod{\`e}le fran{\c c}ais CamemBERT}},
AUTHOR = {Delestre, Cyrile and Amar, Abibatou},
URL = {https://hal.archives-ouvertes.fr/hal-03674695},
BOOKTITLE = {{CAp (Conf{\'e}rence sur l'Apprentissage automatique)}},
ADDRESS = {Vannes, France},
YEAR = {2022},
MONTH = Jul,
KEYWORDS = {NLP ; Transformers ; CamemBERT ; Distillation},
PDF = {https://hal.archives-ouvertes.fr/hal-03674695/file/cap2022.pdf},
HAL_ID = {hal-03674695},
HAL_VERSION = {v1},
}
```
|
{"language": "fr", "license": "mit", "datasets": ["Jean-Baptiste/wikiner_fr"], "widget": [{"text": "Boulanger, habitant \u00e0 Boulanger et travaillant dans le magasin Boulanger situ\u00e9 dans la ville de Boulanger. Boulanger a \u00e9crit le livre \u00e9ponyme Boulanger \u00e9dit\u00e9 par la maison d'\u00e9dition Boulanger."}, {"text": "Quentin Jerome Tarantino na\u00eet le 27 mars 1963 \u00e0 Knoxville, dans le Tennessee. Il est le fils de Connie McHugh, une infirmi\u00e8re, n\u00e9e le 3 septembre 1946, et de Tony Tarantino, acteur et musicien amateur n\u00e9 \u00e0 New York. Ce dernier est d'origine italienne par son p\u00e8re ; sa m\u00e8re a des ascendances irlandaises et cherokees. Il est pr\u00e9nomm\u00e9 d'apr\u00e8s Quint Asper, le personnage jou\u00e9 par Burt Reynolds dans la s\u00e9rie Gunsmoke et Quentin Compson, personnage du roman Le Bruit et la Fureur. Son p\u00e8re quitte le domicile familial avant m\u00eame sa naissance. En 1965, sa m\u00e8re d\u00e9m\u00e9nage \u00e0 Torrance, dans la banlieue sud de Los Angeles, et se remarie avec Curtis Zastoupil, un pianiste de bar, qui lui fait d\u00e9couvrir le cin\u00e9ma. Le couple divorce alors que le jeune Quentin a une dizaine d'ann\u00e9es."}]}
|
cmarkea/distilcamembert-base-ner
| null |
[
"transformers",
"pytorch",
"tf",
"onnx",
"safetensors",
"camembert",
"token-classification",
"fr",
"dataset:Jean-Baptiste/wikiner_fr",
"license:mit",
"autotrain_compatible",
"endpoints_compatible",
"has_space",
"region:us"
] | null |
2022-03-02T23:29:05+00:00
|
zero-shot-classification
|
transformers
|
DistilCamemBERT-NLI
===================
We present DistilCamemBERT-NLI, which is [DistilCamemBERT](https://huggingface.co/cmarkea/distilcamembert-base) fine-tuned for the Natural Language Inference (NLI) task for the french language, also known as recognizing textual entailment (RTE). This model is constructed on the XNLI dataset, which determines whether a premise entails, contradicts or neither entails or contradicts a hypothesis.
This modelization is close to [BaptisteDoyen/camembert-base-xnli](https://huggingface.co/BaptisteDoyen/camembert-base-xnli) based on [CamemBERT](https://huggingface.co/camembert-base) model. The problem of the modelizations based on CamemBERT is at the scaling moment, for the production phase, for example. Indeed, inference cost can be a technological issue especially in the context of cross-encoding like this task. To counteract this effect, we propose this modelization which divides the inference time by 2 with the same consumption power, thanks to DistilCamemBERT.
Dataset
-------
The dataset XNLI from [FLUE](https://huggingface.co/datasets/flue) comprises 392,702 premises with their hypothesis for the train and 5,010 couples for the test. The goal is to predict textual entailment (does sentence A imply/contradict/neither sentence B?) and is a classification task (given two sentences, predict one of three labels). Sentence A is called *premise*, and sentence B is called *hypothesis*, then the goal of modelization is determined as follows:
$$P(premise=c\in\{contradiction, entailment, neutral\}\vert hypothesis)$$
Evaluation results
------------------
| **class** | **precision (%)** | **f1-score (%)** | **support** |
| :----------------: | :---------------: | :--------------: | :---------: |
| **global** | 77.70 | 77.45 | 5,010 |
| **contradiction** | 78.00 | 79.54 | 1,670 |
| **entailment** | 82.90 | 78.87 | 1,670 |
| **neutral** | 72.18 | 74.04 | 1,670 |
Benchmark
---------
We compare the [DistilCamemBERT](https://huggingface.co/cmarkea/distilcamembert-base) model to 2 other modelizations working on the french language. The first one [BaptisteDoyen/camembert-base-xnli](https://huggingface.co/BaptisteDoyen/camembert-base-xnli) is based on well named [CamemBERT](https://huggingface.co/camembert-base), the french RoBERTa model and the second one [MoritzLaurer/mDeBERTa-v3-base-mnli-xnli](https://huggingface.co/MoritzLaurer/mDeBERTa-v3-base-mnli-xnli) based on [mDeBERTav3](https://huggingface.co/microsoft/mdeberta-v3-base) a multilingual model. To compare the performances, the metrics of accuracy and [MCC (Matthews Correlation Coefficient)](https://en.wikipedia.org/wiki/Phi_coefficient) were used. We used an **AMD Ryzen 5 4500U @ 2.3GHz with 6 cores** for mean inference time measure.
| **model** | **time (ms)** | **accuracy (%)** | **MCC (x100)** |
| :--------------: | :-----------: | :--------------: | :------------: |
| [cmarkea/distilcamembert-base-nli](https://huggingface.co/cmarkea/distilcamembert-base-nli) | **51.35** | 77.45 | 66.24 |
| [BaptisteDoyen/camembert-base-xnli](https://huggingface.co/BaptisteDoyen/camembert-base-xnli) | 105.0 | 81.72 | 72.67 |
| [MoritzLaurer/mDeBERTa-v3-base-mnli-xnli](https://huggingface.co/MoritzLaurer/mDeBERTa-v3-base-mnli-xnli) | 299.18 | **83.43** | **75.15** |
Zero-shot classification
------------------------
The main advantage of such modelization is to create a zero-shot classifier allowing text classification without training. This task can be summarized by:
$$P(hypothesis=i\in\mathcal{C}|premise)=\frac{e^{P(premise=entailment\vert hypothesis=i)}}{\sum_{j\in\mathcal{C}}e^{P(premise=entailment\vert hypothesis=j)}}$$
For this part, we use two datasets, the first one: [allocine](https://huggingface.co/datasets/allocine) used to train the sentiment analysis models. The dataset comprises two classes: "positif" and "négatif" appreciation of movie reviews. Here we use "Ce commentaire est {}." as the hypothesis template and "positif" and "négatif" as candidate labels.
| **model** | **time (ms)** | **accuracy (%)** | **MCC (x100)** |
| :--------------: | :-----------: | :--------------: | :------------: |
| [cmarkea/distilcamembert-base-nli](https://huggingface.co/cmarkea/distilcamembert-base-nli) | **195.54** | 80.59 | 63.71 |
| [BaptisteDoyen/camembert-base-xnli](https://huggingface.co/BaptisteDoyen/camembert-base-xnli) | 378.39 | **86.37** | **73.74** |
| [MoritzLaurer/mDeBERTa-v3-base-mnli-xnli](https://huggingface.co/MoritzLaurer/mDeBERTa-v3-base-mnli-xnli) | 520.58 | 84.97 | 70.05 |
The second one: [mlsum](https://huggingface.co/datasets/mlsum) used to train the summarization models. In this aim, we aggregate sub-topics and select a few of them. We use the articles summary part to predict their topics. In this case, the hypothesis template used is "C'est un article traitant de {}." and the candidate labels are: "économie", "politique", "sport" and "science".
| **model** | **time (ms)** | **accuracy (%)** | **MCC (x100)** |
| :--------------: | :-----------: | :--------------: | :------------: |
| [cmarkea/distilcamembert-base-nli](https://huggingface.co/cmarkea/distilcamembert-base-nli) | **217.77** | **79.30** | **70.55** |
| [BaptisteDoyen/camembert-base-xnli](https://huggingface.co/BaptisteDoyen/camembert-base-xnli) | 448.27 | 70.7 | 64.10 |
| [MoritzLaurer/mDeBERTa-v3-base-mnli-xnli](https://huggingface.co/MoritzLaurer/mDeBERTa-v3-base-mnli-xnli) | 591.34 | 64.45 | 58.67 |
How to use DistilCamemBERT-NLI
------------------------------
```python
from transformers import pipeline
classifier = pipeline(
task='zero-shot-classification',
model="cmarkea/distilcamembert-base-nli",
tokenizer="cmarkea/distilcamembert-base-nli"
)
result = classifier (
sequences="Le style très cinéphile de Quentin Tarantino "
"se reconnaît entre autres par sa narration postmoderne "
"et non linéaire, ses dialogues travaillés souvent "
"émaillés de références à la culture populaire, et ses "
"scènes hautement esthétiques mais d'une violence "
"extrême, inspirées de films d'exploitation, d'arts "
"martiaux ou de western spaghetti.",
candidate_labels="cinéma, technologie, littérature, politique",
hypothesis_template="Ce texte parle de {}."
)
result
{"labels": ["cinéma",
"littérature",
"technologie",
"politique"],
"scores": [0.7164115309715271,
0.12878799438476562,
0.1092301607131958,
0.0455702543258667]}
```
### Optimum + ONNX
```python
from optimum.onnxruntime import ORTModelForSequenceClassification
from transformers import AutoTokenizer, pipeline
HUB_MODEL = "cmarkea/distilcamembert-base-nli"
tokenizer = AutoTokenizer.from_pretrained(HUB_MODEL)
model = ORTModelForSequenceClassification.from_pretrained(HUB_MODEL)
onnx_qa = pipeline("zero-shot-classification", model=model, tokenizer=tokenizer)
# Quantized onnx model
quantized_model = ORTModelForSequenceClassification.from_pretrained(
HUB_MODEL, file_name="model_quantized.onnx"
)
```
Citation
--------
```bibtex
@inproceedings{delestre:hal-03674695,
TITLE = {{DistilCamemBERT : une distillation du mod{\`e}le fran{\c c}ais CamemBERT}},
AUTHOR = {Delestre, Cyrile and Amar, Abibatou},
URL = {https://hal.archives-ouvertes.fr/hal-03674695},
BOOKTITLE = {{CAp (Conf{\'e}rence sur l'Apprentissage automatique)}},
ADDRESS = {Vannes, France},
YEAR = {2022},
MONTH = Jul,
KEYWORDS = {NLP ; Transformers ; CamemBERT ; Distillation},
PDF = {https://hal.archives-ouvertes.fr/hal-03674695/file/cap2022.pdf},
HAL_ID = {hal-03674695},
HAL_VERSION = {v1},
}
```
|
{"language": "fr", "license": "mit", "tags": ["zero-shot-classification", "sentence-similarity", "nli"], "datasets": ["flue"], "pipeline_tag": "zero-shot-classification", "widget": [{"text": "Selon certains physiciens, un univers parall\u00e8le, miroir du n\u00f4tre ou relevant de ce que l'on appelle la th\u00e9orie des branes, autoriserait des neutrons \u00e0 sortir de notre Univers pour y entrer \u00e0 nouveau. L'id\u00e9e a \u00e9t\u00e9 test\u00e9e une nouvelle fois avec le r\u00e9acteur nucl\u00e9aire de l'Institut Laue-Langevin \u00e0 Grenoble, plus pr\u00e9cis\u00e9ment en utilisant le d\u00e9tecteur de l'exp\u00e9rience Stereo initialement con\u00e7u pour chasser des particules de mati\u00e8re noire potentielles, les neutrinos st\u00e9riles.", "candidate_labels": "politique, science, sport, sant\u00e9", "hypothesis_template": "Ce texte parle de {}."}]}
|
cmarkea/distilcamembert-base-nli
| null |
[
"transformers",
"pytorch",
"tf",
"onnx",
"safetensors",
"camembert",
"text-classification",
"zero-shot-classification",
"sentence-similarity",
"nli",
"fr",
"dataset:flue",
"license:mit",
"autotrain_compatible",
"endpoints_compatible",
"region:us"
] | null |
2022-03-02T23:29:05+00:00
|
question-answering
|
transformers
|
DistilCamemBERT-QA
==================
We present DistilCamemBERT-QA, which is [DistilCamemBERT](https://huggingface.co/cmarkea/distilcamembert-base) fine-tuned for the Question-Answering task for the french language. This model is built using two datasets, FQuAD v1.0 and Piaf, composed of contexts and questions with their answers inside the context.
This modelization is close to [etalab-ia/camembert-base-squadFR-fquad-piaf](https://huggingface.co/etalab-ia/camembert-base-squadFR-fquad-piaf) based on [CamemBERT](https://huggingface.co/camembert-base) model. The problem of the modelizations based on CamemBERT is at the scaling moment, for the production phase, for example. Indeed, inference cost can be a technological issue, especially in cross-encoding like this task. To counteract this effect, we propose this modelization which divides the inference time by 2 with the same consumption power, thanks to DistilCamemBERT.
Dataset
-------
The dataset comprises FQuAD v1.0 and Piaf with 24'566 questions and answers for the training set and 3'188 for the evaluation set.
Evaluation results and benchmark
--------------------------------
We compare [DistilCamemBERT-QA](https://huggingface.co/cmarkea/distilcamembert-base-qa) to two other modelizations working on the french language. The first one [etalab-ia/camembert-base-squadFR-fquad-piaf](https://huggingface.co/etalab-ia/camembert-base-squadFR-fquad-piaf) is based on well named [CamemBERT](https://huggingface.co/camembert-base), the french RoBERTa model and the second one [fmikaelian/flaubert-base-uncased-squad](https://huggingface.co/fmikaelian/flaubert-base-uncased-squad) is based on [FlauBERT](https://huggingface.co/flaubert/flaubert_base_uncased) another french model based on BERT architecture this time.
For our benchmarks, we do a word-to-word comparison between words that are matching between the predicted answer and the ground truth. We also use f1-score, which measures the intersection quality between predicted responses and ground truth. Finally, we use inclusion score, which measures if the ground truth answer is included in the predicted answer. An **AMD Ryzen 5 4500U @ 2.3GHz with 6 cores** was used for the mean inference time measure.
| **model** | **time (ms)** | **exact match (%)** | **f1-score (%)** | **inclusion-score (%)** |
| :--------------: | :-----------: | :--------------: | :------------: | :------------: |
| [cmarkea/distilcamembert-base-qa](https://huggingface.co/cmarkea/distilcamembert-base-qa) | **216.96** | 25.66 | 62.65 | 59.82 |
| [etalab-ia/camembert-base-squadFR-fquad-piaf](https://huggingface.co/etalab-ia/camembert-base-squadFR-fquad-piaf) | 432.17 | **59.76** | **79.57** | **69.23** |
| [fmikaelian/flaubert-base-uncased-squad](https://huggingface.co/fmikaelian/flaubert-base-uncased-squad) | 875.84 | 0.22 | 5.21 | 3.68 |
Do not take into account the results of the FlauBERT model. The modeling seems to be a problem, as the results seem very low.
How to use DistilCamemBERT-QA
------------------------------
```python
from transformers import pipeline
qa_engine = pipeline(
"question-answering",
model="cmarkea/distilcamembert-base-qa",
tokenizer="cmarkea/distilcamembert-base-qa"
)
result = qa_engine(
context="David Fincher, né le 28 août 1962 à Denver (Colorado), "
"est un réalisateur et producteur américain. Il est principalement "
"connu pour avoir réalisé les films Seven, Fight Club, L'Étrange "
"Histoire de Benjamin Button, The Social Network et Gone Girl qui "
"lui ont valu diverses récompenses et nominations aux Oscars du "
"cinéma ou aux Golden Globes. Réputé pour son perfectionnisme, il "
"peut tourner un très grand nombre de prises de ses plans et "
"séquences afin d'obtenir le rendu visuel qu'il désire. Il a "
"également développé et produit les séries télévisées House of "
"Cards (pour laquelle il remporte l'Emmy Award de la meilleure "
"réalisation pour une série dramatique en 2013) et Mindhunter, "
"diffusées sur Netflix.",
question="Quel est le métier de David Fincher ?"
)
result
{'score': 0.7981914281845093,
'start': 61,
'end': 98,
'answer': ' réalisateur et producteur américain.'}
```
### Optimum + ONNX
```python
from optimum.onnxruntime import ORTModelForQuestionAnswering
from transformers import AutoTokenizer, pipeline
HUB_MODEL = "cmarkea/distilcamembert-base-qa"
tokenizer = AutoTokenizer.from_pretrained(HUB_MODEL)
model = ORTModelForQuestionAnswering.from_pretrained(HUB_MODEL)
onnx_qa = pipeline("question-answering", model=model, tokenizer=tokenizer)
# Quantized onnx model
quantized_model = ORTModelForQuestionAnswering.from_pretrained(
HUB_MODEL, file_name="model_quantized.onnx"
)
```
Citation
--------
```bibtex
@inproceedings{delestre:hal-03674695,
TITLE = {{DistilCamemBERT : une distillation du mod{\`e}le fran{\c c}ais CamemBERT}},
AUTHOR = {Delestre, Cyrile and Amar, Abibatou},
URL = {https://hal.archives-ouvertes.fr/hal-03674695},
BOOKTITLE = {{CAp (Conf{\'e}rence sur l'Apprentissage automatique)}},
ADDRESS = {Vannes, France},
YEAR = {2022},
MONTH = Jul,
KEYWORDS = {NLP ; Transformers ; CamemBERT ; Distillation},
PDF = {https://hal.archives-ouvertes.fr/hal-03674695/file/cap2022.pdf},
HAL_ID = {hal-03674695},
HAL_VERSION = {v1},
}
```
|
{"language": "fr", "license": "cc-by-nc-sa-3.0", "datasets": ["fquad", "piaf"], "widget": [{"text": "Quand et o\u00f9 est sorti Toy Story ?", "context": "Pixar Animation Studios, ou simplement Pixar dans le langage courant, est une soci\u00e9t\u00e9 am\u00e9ricaine de production de films en images tridimensionnelles de synth\u00e8se. Elle a acquis sa notori\u00e9t\u00e9 gr\u00e2ce \u00e0 Toy Story, premier long m\u00e9trage de ce type, sorti aux \u00c9tats-Unis en 1995. \u00c0 ce jour, le studio d'animation a remport\u00e9 dix-neuf Oscars, quatre Golden Globes et trois Grammy Awards ainsi que de nombreuses autres r\u00e9compenses. Le studio travaille avec PhotoRealistic RenderMan, sa propre version de l'interface de programmation de rendu RenderMan utilis\u00e9e pour cr\u00e9er des images de haute qualit\u00e9. Ses studios de production et son si\u00e8ge social se trouvent au Pixar Campus situ\u00e9 \u00e0 Emeryville pr\u00e8s de San Francisco en Californie."}, {"text": "Quel est le premier long m\u00e9trage du studio ?", "context": "Pixar Animation Studios, ou simplement Pixar dans le langage courant, est une soci\u00e9t\u00e9 am\u00e9ricaine de production de films en images tridimensionnelles de synth\u00e8se. Elle a acquis sa notori\u00e9t\u00e9 gr\u00e2ce \u00e0 Toy Story, premier long m\u00e9trage de ce type, sorti aux \u00c9tats-Unis en 1995. \u00c0 ce jour, le studio d'animation a remport\u00e9 dix-neuf Oscars, quatre Golden Globes et trois Grammy Awards ainsi que de nombreuses autres r\u00e9compenses. Le studio travaille avec PhotoRealistic RenderMan, sa propre version de l'interface de programmation de rendu RenderMan utilis\u00e9e pour cr\u00e9er des images de haute qualit\u00e9. Ses studios de production et son si\u00e8ge social se trouvent au Pixar Campus situ\u00e9 \u00e0 Emeryville pr\u00e8s de San Francisco en Californie."}]}
|
cmarkea/distilcamembert-base-qa
| null |
[
"transformers",
"pytorch",
"tf",
"onnx",
"safetensors",
"camembert",
"question-answering",
"fr",
"dataset:fquad",
"dataset:piaf",
"license:cc-by-nc-sa-3.0",
"endpoints_compatible",
"region:us"
] | null |
2022-03-02T23:29:05+00:00
|
text-classification
|
transformers
|
DistilCamemBERT-Sentiment
=========================
We present DistilCamemBERT-Sentiment, which is [DistilCamemBERT](https://huggingface.co/cmarkea/distilcamembert-base) fine-tuned for the sentiment analysis task for the French language. This model is built using two datasets: [Amazon Reviews](https://huggingface.co/datasets/amazon_reviews_multi) and [Allociné.fr](https://huggingface.co/datasets/allocine) to minimize the bias. Indeed, Amazon reviews are similar in messages and relatively shorts, contrary to Allociné critics, who are long and rich texts.
This modelization is close to [tblard/tf-allocine](https://huggingface.co/tblard/tf-allocine) based on [CamemBERT](https://huggingface.co/camembert-base) model. The problem of the modelizations based on CamemBERT is at the scaling moment, for the production phase, for example. Indeed, inference cost can be a technological issue. To counteract this effect, we propose this modelization which **divides the inference time by two** with the same consumption power thanks to [DistilCamemBERT](https://huggingface.co/cmarkea/distilcamembert-base).
Dataset
-------
The dataset comprises 204,993 reviews for training and 4,999 reviews for the test from Amazon, and 235,516 and 4,729 critics from [Allocine website](https://www.allocine.fr/). The dataset is labeled into five categories:
* 1 star: represents a terrible appreciation,
* 2 stars: bad appreciation,
* 3 stars: neutral appreciation,
* 4 stars: good appreciation,
* 5 stars: excellent appreciation.
Evaluation results
------------------
In addition of accuracy (called here *exact accuracy*) in order to be robust to +/-1 star estimation errors, we will take the following definition as a performance measure:
$$\mathrm{top\!-\!2\; acc}=\frac{1}{|\mathcal{O}|}\sum_{i\in\mathcal{O}}\sum_{0\leq l < 2}\mathbb{1}(\hat{f}_{i,l}=y_i)$$
where \\(\hat{f}_l\\) is the l-th largest predicted label, \\(y\\) the true label, \\(\mathcal{O}\\) is the test set of the observations and \\(\mathbb{1}\\) is the indicator function.
| **class** | **exact accuracy (%)** | **top-2 acc (%)** | **support** |
| :---------: | :--------------------: | :---------------: | :---------: |
| **global** | 61.01 | 88.80 | 9,698 |
| **1 star** | 87.21 | 77.17 | 1,905 |
| **2 stars** | 79.19 | 84.75 | 1,935 |
| **3 stars** | 77.85 | 78.98 | 1,974 |
| **4 stars** | 78.61 | 90.22 | 1,952 |
| **5 stars** | 85.96 | 82.92 | 1,932 |
Benchmark
---------
This model is compared to 3 reference models (see below). As each model doesn't have the exact definition of targets, we detail the performance measure used for each. An **AMD Ryzen 5 4500U @ 2.3GHz with 6 cores** was used for the mean inference time measure.
#### bert-base-multilingual-uncased-sentiment
[nlptown/bert-base-multilingual-uncased-sentiment](https://huggingface.co/nlptown/bert-base-multilingual-uncased-sentiment) is based on BERT model in the multilingual and uncased version. This sentiment analyzer is trained on Amazon reviews, similar to our model. Hence the targets and their definitions are the same.
| **model** | **time (ms)** | **exact accuracy (%)** | **top-2 acc (%)** |
| :-------: | :-----------: | :--------------------: | :---------------: |
| [cmarkea/distilcamembert-base-sentiment](https://huggingface.co/cmarkea/distilcamembert-base-sentiment) | **95.56** | **61.01** | **88.80** |
| [nlptown/bert-base-multilingual-uncased-sentiment](https://huggingface.co/nlptown/bert-base-multilingual-uncased-sentiment) | 187.70 | 54.41 | 82.82 |
#### tf-allociné and barthez-sentiment-classification
[tblard/tf-allocine](https://huggingface.co/tblard/tf-allocine) based on [CamemBERT](https://huggingface.co/camembert-base) model and [moussaKam/barthez-sentiment-classification](https://huggingface.co/moussaKam/barthez-sentiment-classification) based on [BARThez](https://huggingface.co/moussaKam/barthez) use the same bi-class definition between them. To bring this back to a two-class problem, we will only consider the *"1 star"* and *"2 stars"* labels for the *negative* sentiments and *"4 stars"* and *"5 stars"* for *positive* sentiments. We exclude the *"3 stars"* which can be interpreted as a *neutral* class. In this context, the problem of +/-1 star estimation errors disappears. Then we use only the classical accuracy definition.
| **model** | **time (ms)** | **exact accuracy (%)** |
| :-------: | :-----------: | :--------------------: |
| [cmarkea/distilcamembert-base-sentiment](https://huggingface.co/cmarkea/distilcamembert-base-sentiment) | **95.56** | **97.52** |
| [tblard/tf-allocine](https://huggingface.co/tblard/tf-allocine) | 329.74 | 95.69 |
| [moussaKam/barthez-sentiment-classification](https://huggingface.co/moussaKam/barthez-sentiment-classification) | 197.95 | 94.29 |
How to use DistilCamemBERT-Sentiment
------------------------------------
```python
from transformers import pipeline
analyzer = pipeline(
task='text-classification',
model="cmarkea/distilcamembert-base-sentiment",
tokenizer="cmarkea/distilcamembert-base-sentiment"
)
result = analyzer(
"J'aime me promener en forêt même si ça me donne mal aux pieds.",
return_all_scores=True
)
result
[{'label': '1 star',
'score': 0.047529436647892},
{'label': '2 stars',
'score': 0.14150355756282806},
{'label': '3 stars',
'score': 0.3586442470550537},
{'label': '4 stars',
'score': 0.3181498646736145},
{'label': '5 stars',
'score': 0.13417290151119232}]
```
### Optimum + ONNX
```python
from optimum.onnxruntime import ORTModelForSequenceClassification
from transformers import AutoTokenizer, pipeline
HUB_MODEL = "cmarkea/distilcamembert-base-sentiment"
tokenizer = AutoTokenizer.from_pretrained(HUB_MODEL)
model = ORTModelForSequenceClassification.from_pretrained(HUB_MODEL)
onnx_qa = pipeline("text-classification", model=model, tokenizer=tokenizer)
# Quantized onnx model
quantized_model = ORTModelForSequenceClassification.from_pretrained(
HUB_MODEL, file_name="model_quantized.onnx"
)
```
Citation
--------
```bibtex
@inproceedings{delestre:hal-03674695,
TITLE = {{DistilCamemBERT : une distillation du mod{\`e}le fran{\c c}ais CamemBERT}},
AUTHOR = {Delestre, Cyrile and Amar, Abibatou},
URL = {https://hal.archives-ouvertes.fr/hal-03674695},
BOOKTITLE = {{CAp (Conf{\'e}rence sur l'Apprentissage automatique)}},
ADDRESS = {Vannes, France},
YEAR = {2022},
MONTH = Jul,
KEYWORDS = {NLP ; Transformers ; CamemBERT ; Distillation},
PDF = {https://hal.archives-ouvertes.fr/hal-03674695/file/cap2022.pdf},
HAL_ID = {hal-03674695},
HAL_VERSION = {v1},
}
```
|
{"language": "fr", "license": "mit", "datasets": ["amazon_reviews_multi", "allocine"], "widget": [{"text": "Je pensais lire un livre nul, mais finalement je l'ai trouv\u00e9 super !"}, {"text": "Cette banque est tr\u00e8s bien, mais elle n'offre pas les services de paiements sans contact."}, {"text": "Cette banque est tr\u00e8s bien et elle offre en plus les services de paiements sans contact."}]}
|
cmarkea/distilcamembert-base-sentiment
| null |
[
"transformers",
"pytorch",
"tf",
"onnx",
"safetensors",
"camembert",
"text-classification",
"fr",
"dataset:amazon_reviews_multi",
"dataset:allocine",
"license:mit",
"autotrain_compatible",
"endpoints_compatible",
"has_space",
"region:us"
] | null |
2022-03-02T23:29:05+00:00
|
fill-mask
|
transformers
|
DistilCamemBERT
===============
We present a distillation version of the well named [CamemBERT](https://huggingface.co/camembert-base), a RoBERTa French model version, alias DistilCamemBERT. The aim of distillation is to drastically reduce the complexity of the model while preserving the performances. The proof of concept is shown in the [DistilBERT paper](https://arxiv.org/abs/1910.01108) and the code used for the training is inspired by the code of [DistilBERT](https://github.com/huggingface/transformers/tree/master/examples/research_projects/distillation).
Loss function
-------------
The training for the distilled model (student model) is designed to be the closest as possible to the original model (teacher model). To perform this the loss function is composed of 3 parts:
* DistilLoss: a distillation loss which measures the silimarity between the probabilities at the outputs of the student and teacher models with a cross-entropy loss on the MLM task ;
* CosineLoss: a cosine embedding loss. This loss function is applied on the last hidden layers of student and teacher models to guarantee a collinearity between them ;
* MLMLoss: and finaly a Masked Language Modeling (MLM) task loss to perform the student model with the original task of the teacher model.
The final loss function is a combination of these three losses functions. We use the following ponderation:
$$Loss = 0.5 \times DistilLoss + 0.3 \times CosineLoss + 0.2 \times MLMLoss$$
Dataset
-------
To limit the bias between the student and teacher models, the dataset used for the DstilCamemBERT training is the same as the camembert-base training one: OSCAR. The French part of this dataset approximately represents 140 GB on a hard drive disk.
Training
--------
We pre-trained the model on a nVidia Titan RTX during 18 days.
Evaluation results
------------------
| Dataset name | f1-score |
| :----------: | :------: |
| [FLUE](https://huggingface.co/datasets/flue) CLS | 83% |
| [FLUE](https://huggingface.co/datasets/flue) PAWS-X | 77% |
| [FLUE](https://huggingface.co/datasets/flue) XNLI | 77% |
| [wikiner_fr](https://huggingface.co/datasets/Jean-Baptiste/wikiner_fr) NER | 98% |
How to use DistilCamemBERT
--------------------------
Load DistilCamemBERT and its sub-word tokenizer :
```python
from transformers import AutoTokenizer, AutoModel
tokenizer = AutoTokenizer.from_pretrained("cmarkea/distilcamembert-base")
model = AutoModel.from_pretrained("cmarkea/distilcamembert-base")
model.eval()
...
```
Filling masks using pipeline :
```python
from transformers import pipeline
model_fill_mask = pipeline("fill-mask", model="cmarkea/distilcamembert-base", tokenizer="cmarkea/distilcamembert-base")
results = model_fill_mask("Le camembert est <mask> :)")
results
[{'sequence': '<s> Le camembert est délicieux :)</s>', 'score': 0.3878222405910492, 'token': 7200},
{'sequence': '<s> Le camembert est excellent :)</s>', 'score': 0.06469205021858215, 'token': 2183},
{'sequence': '<s> Le camembert est parfait :)</s>', 'score': 0.04534877464175224, 'token': 1654},
{'sequence': '<s> Le camembert est succulent :)</s>', 'score': 0.04128391295671463, 'token': 26202},
{'sequence': '<s> Le camembert est magnifique :)</s>', 'score': 0.02425697259604931, 'token': 1509}]
```
Citation
--------
```bibtex
@inproceedings{delestre:hal-03674695,
TITLE = {{DistilCamemBERT : une distillation du mod{\`e}le fran{\c c}ais CamemBERT}},
AUTHOR = {Delestre, Cyrile and Amar, Abibatou},
URL = {https://hal.archives-ouvertes.fr/hal-03674695},
BOOKTITLE = {{CAp (Conf{\'e}rence sur l'Apprentissage automatique)}},
ADDRESS = {Vannes, France},
YEAR = {2022},
MONTH = Jul,
KEYWORDS = {NLP ; Transformers ; CamemBERT ; Distillation},
PDF = {https://hal.archives-ouvertes.fr/hal-03674695/file/cap2022.pdf},
HAL_ID = {hal-03674695},
HAL_VERSION = {v1},
}
```
|
{"language": "fr", "license": "mit", "datasets": ["oscar"], "widget": [{"text": "J'aime lire les <mask> de SF."}]}
|
cmarkea/distilcamembert-base
| null |
[
"transformers",
"pytorch",
"tf",
"safetensors",
"camembert",
"fill-mask",
"fr",
"dataset:oscar",
"arxiv:1910.01108",
"license:mit",
"autotrain_compatible",
"endpoints_compatible",
"region:us"
] | null |
2022-03-02T23:29:05+00:00
|
Subsets and Splits
No community queries yet
The top public SQL queries from the community will appear here once available.