Shadowgraph
TrendingBuzzTop stargazers
Chrome extension
Sign in

fabriceyhc/bert-base-uncased-ag_news

Model

should probably proofread and complete it, then remove this comment. -->

0

3 commits

1 linked in READMEs

updated Sep 21, 2021

See the code

README

bert-base-uncased-ag_news

This model is a fine-tuned version of bert-base-uncased on the ag_news dataset. It achieves the following results on the evaluation set:

  • Loss: 0.3284
  • Accuracy: 0.9375

Model description

More information needed

Intended uses & limitations

More information needed

Training and evaluation data

More information needed

Training procedure

Training hyperparameters

The following hyperparameters were used during training:

  • learning_rate: 5e-05
  • train_batch_size: 8
  • eval_batch_size: 16
  • seed: 42
  • optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
  • lr_scheduler_type: linear
  • lr_scheduler_warmup_steps: 7425
  • training_steps: 74250

Training results

Training LossEpochStepValidation LossAccuracy
0.57730.1320000.36270.8875
0.31010.2740000.29380.9208
0.30760.460000.31140.9092
0.31140.5480000.45450.9008
0.31540.67100000.38750.9083
0.30950.81120000.33900.9142
0.29480.94140000.33410.9133
0.25571.08160000.45730.9092
0.2581.21180000.33560.9217
0.24551.35200000.33480.9283
0.23611.48220000.32180.93
0.2541.62240000.38140.9033
0.25281.75260000.36280.9158
0.22821.89280000.33020.9308
0.2242.02300000.39670.9225
0.1742.15320000.36690.9333
0.18482.29340000.34350.9283
0.192.42360000.35520.93
0.18652.56380000.39960.9258
0.18772.69400000.37490.9258
0.19512.83420000.39630.9258
0.17022.96440000.36550.9317
0.14883.1460000.39420.9292
0.12313.23480000.39980.9267
0.13193.37500000.42920.9242
0.13343.5520000.49040.9192

Framework versions

  • Transformers 4.10.2
  • Pytorch 1.7.1
  • Datasets 1.6.1
  • Tokenizers 0.10.3
bert
endpoints_compatible
generated_from_trainer
model-index
pytorch
sibyl
text-classification
text-embeddings-inference
transformers

Similar repos

ungjus/Fake_News_BERT_Classifier

Model

should probably proofread and complete it, then remove this comment. -->

0

Pra-tham/results

Model

should probably proofread and complete it, then remove this comment. -->

0

lvwerra/distilbert-imdb

Model

should probably proofread and complete it, then remove this comment. -->

25

joelito/legal-german-roberta-base

Model

should probably proofread and complete it, then remove this comment. -->

1

joelito/legal-polish-roberta-base

Model

should probably proofread and complete it, then remove this comment. -->

2

joelito/legal-hungarian-roberta-base

Model

should probably proofread and complete it, then remove this comment. -->

0

joelito/legal-bulgarian-roberta-base

Model

should probably proofread and complete it, then remove this comment. -->

1

joelito/legal-slovenian-roberta-base

Model

should probably proofread and complete it, then remove this comment. -->

0

Linked in READMEs

kenhktsui/anyclassifier

One Line To Build Zero-Data Classifiers in Minutes

64

README badge

fabriceyhc/bert-base-uncased-ag_news

Model

should probably proofread and complete it, then remove this comment. -->

0

3 commits

1 linked in READMEs

updated Sep 21, 2021

See the code

README

bert-base-uncased-ag_news

This model is a fine-tuned version of bert-base-uncased on the ag_news dataset. It achieves the following results on the evaluation set:

  • Loss: 0.3284
  • Accuracy: 0.9375

Model description

More information needed

Intended uses & limitations

More information needed

Training and evaluation data

More information needed

Training procedure

Training hyperparameters

The following hyperparameters were used during training:

  • learning_rate: 5e-05
  • train_batch_size: 8
  • eval_batch_size: 16
  • seed: 42
  • optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
  • lr_scheduler_type: linear
  • lr_scheduler_warmup_steps: 7425
  • training_steps: 74250

Training results

Training LossEpochStepValidation LossAccuracy
0.57730.1320000.36270.8875
0.31010.2740000.29380.9208
0.30760.460000.31140.9092
0.31140.5480000.45450.9008
0.31540.67100000.38750.9083
0.30950.81120000.33900.9142
0.29480.94140000.33410.9133
0.25571.08160000.45730.9092
0.2581.21180000.33560.9217
0.24551.35200000.33480.9283
0.23611.48220000.32180.93
0.2541.62240000.38140.9033
0.25281.75260000.36280.9158
0.22821.89280000.33020.9308
0.2242.02300000.39670.9225
0.1742.15320000.36690.9333
0.18482.29340000.34350.9283
0.192.42360000.35520.93
0.18652.56380000.39960.9258
0.18772.69400000.37490.9258
0.19512.83420000.39630.9258
0.17022.96440000.36550.9317
0.14883.1460000.39420.9292
0.12313.23480000.39980.9267
0.13193.37500000.42920.9242
0.13343.5520000.49040.9192

Framework versions

  • Transformers 4.10.2
  • Pytorch 1.7.1
  • Datasets 1.6.1
  • Tokenizers 0.10.3
bert
endpoints_compatible
generated_from_trainer
model-index
pytorch
sibyl
text-classification
text-embeddings-inference
transformers

Similar repos

ungjus/Fake_News_BERT_Classifier

Model

should probably proofread and complete it, then remove this comment. -->

0

Pra-tham/results

Model

should probably proofread and complete it, then remove this comment. -->

0

lvwerra/distilbert-imdb

Model

should probably proofread and complete it, then remove this comment. -->

25

joelito/legal-german-roberta-base

Model

should probably proofread and complete it, then remove this comment. -->

1

joelito/legal-polish-roberta-base

Model

should probably proofread and complete it, then remove this comment. -->

2

joelito/legal-hungarian-roberta-base

Model

should probably proofread and complete it, then remove this comment. -->

0

joelito/legal-bulgarian-roberta-base

Model

should probably proofread and complete it, then remove this comment. -->

1

joelito/legal-slovenian-roberta-base

Model

should probably proofread and complete it, then remove this comment. -->

0

Linked in READMEs

kenhktsui/anyclassifier

One Line To Build Zero-Data Classifiers in Minutes

64

README badge