Hugging Face's logo

gokuls
/
HBERTv1_emb_compress_48_L12_H256_A4

Fill-Mask
Transformers
PyTorch
hybridbert
Generated from Trainer
Eval Results
Inference Endpoints
Model card Files Community
HBERTv1_emb_compress_48_L12_H256_A4 / README.md
gokuls's picture
gokuls
End of training
fbdccca
preview code
|
raw
history blame contribute delete
4.43 kB
metadata
tags:
  - generated_from_trainer
datasets:
  - gokuls/wiki_book_corpus_complete_processed_bert_dataset
metrics:
  - accuracy
model-index:
  - name: HBERTv1_emb_compress_48_L12_H256_A4
    results:
      - task:
          name: Masked Language Modeling
          type: fill-mask
        dataset:
          name: gokuls/wiki_book_corpus_complete_processed_bert_dataset
          type: gokuls/wiki_book_corpus_complete_processed_bert_dataset
        metrics:
          - name: Accuracy
            type: accuracy
            value: 0.15102291312237043

HBERTv1_emb_compress_48_L12_H256_A4

This model is a fine-tuned version of on the gokuls/wiki_book_corpus_complete_processed_bert_dataset dataset. It achieves the following results on the evaluation set:

  • Loss: 6.0478
  • Accuracy: 0.1510

Model description

More information needed

Intended uses & limitations

More information needed

Training and evaluation data

More information needed

Training procedure

Training hyperparameters

The following hyperparameters were used during training:

  • learning_rate: 1e-05
  • train_batch_size: 64
  • eval_batch_size: 64
  • seed: 10
  • distributed_type: multi-GPU
  • optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
  • lr_scheduler_type: linear
  • lr_scheduler_warmup_steps: 10000
  • num_epochs: 5

Training results

Training Loss Epoch Step Validation Loss Accuracy
7.1159 0.11 10000 7.0948 0.0805
6.698 0.22 20000 6.6913 0.1060
6.5481 0.33 30000 6.5473 0.1167
6.4589 0.44 40000 6.4576 0.1252
6.3925 0.55 50000 6.3858 0.1306
6.3433 0.66 60000 6.3356 0.1353
6.2983 0.76 70000 6.2965 0.1376
6.268 0.87 80000 6.2643 0.1397
6.2359 0.98 90000 6.2381 0.1411
6.2186 1.09 100000 6.2160 0.1429
6.1915 1.2 110000 6.1972 0.1439
6.1811 1.31 120000 6.1834 0.1440
6.1696 1.42 130000 6.1692 0.1455
6.1621 1.53 140000 6.1557 0.1454
6.1417 1.64 150000 6.1466 0.1468
6.1391 1.75 160000 6.1364 0.1466
6.1338 1.86 170000 6.1281 0.1476
6.1285 1.97 180000 6.1200 0.1477
6.1147 2.08 190000 6.1135 0.1483
6.1139 2.18 200000 6.1083 0.1486
6.1004 2.29 210000 6.1004 0.1487
6.0997 2.4 220000 6.0964 0.1489
6.092 2.51 230000 6.0922 0.1490
6.089 2.62 240000 6.0862 0.1490
6.0841 2.73 250000 6.0829 0.1498
6.0847 2.84 260000 6.0799 0.1496
6.0834 2.95 270000 6.0760 0.1501
6.0752 3.06 280000 6.0715 0.1502
6.0693 3.17 290000 6.0697 0.1502
6.0677 3.28 300000 6.0679 0.1502
6.0646 3.39 310000 6.0646 0.1503
6.0625 3.5 320000 6.0623 0.1503
6.0536 3.6 330000 6.0593 0.1507
6.0574 3.71 340000 6.0577 0.1507
6.0496 3.82 350000 6.0560 0.1508
6.0525 3.93 360000 6.0543 0.1507
6.0498 4.04 370000 6.0508 0.1509
6.0557 4.15 380000 6.0509 0.1508
6.0445 4.26 390000 6.0483 0.1509
6.0466 4.37 400000 6.0470 0.1510
6.0507 4.48 410000 6.0471 0.1510
6.0459 4.59 420000 6.0468 0.1510

Framework versions

  • Transformers 4.33.2
  • Pytorch 1.14.0a0+410ce96
  • Datasets 2.14.5
  • Tokenizers 0.13.3