tiny-mlm-imdb / README.md
muhtasham's picture
update model card README.md
0cc5986
metadata
license: apache-2.0
tags:
  - generated_from_trainer
model-index:
  - name: tiny-mlm-imdb
    results: []

tiny-mlm-imdb

This model is a fine-tuned version of google/bert_uncased_L-2_H-128_A-2 on the None dataset. It achieves the following results on the evaluation set:

  • Loss: 3.5540

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: 3e-05
  • train_batch_size: 32
  • eval_batch_size: 32
  • seed: 42
  • optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
  • lr_scheduler_type: constant
  • num_epochs: 200

Training results

Training Loss Epoch Step Validation Loss
4.2358 0.16 500 3.8225
4.1206 0.32 1000 3.7793
4.0857 0.48 1500 3.7520
4.0699 0.64 2000 3.7277
4.0378 0.8 2500 3.7125
4.0191 0.96 3000 3.7019
3.9747 1.12 3500 3.6871
3.9647 1.28 4000 3.6735
3.956 1.44 4500 3.6773
3.9574 1.6 5000 3.6580
3.9408 1.76 5500 3.6435
3.9421 1.92 6000 3.6419
3.9265 2.08 6500 3.6343
3.9198 2.24 7000 3.6306
3.9205 2.4 7500 3.6198
3.8985 2.56 8000 3.6158
3.9167 2.72 8500 3.6091
3.9111 2.88 9000 3.6073
3.8882 3.04 9500 3.5922
3.8761 3.2 10000 3.5908
3.8603 3.36 10500 3.5841
3.8621 3.52 11000 3.5835
3.8332 3.68 11500 3.5883
3.8523 3.84 12000 3.5798
3.8449 4.0 12500 3.5771
3.8284 4.16 13000 3.5653
3.8253 4.32 13500 3.5701
3.8021 4.48 14000 3.5681
3.8316 4.64 14500 3.5537
3.8318 4.8 15000 3.5609
3.82 4.96 15500 3.5579
3.8094 5.12 16000 3.5540

Framework versions

  • Transformers 4.25.1
  • Pytorch 1.12.1
  • Datasets 2.7.1
  • Tokenizers 0.13.2