roberta-tiny-8l-10M / README.md
g8a9's picture
update model card README.md
191c1bc
metadata
tags:
  - generated_from_trainer
metrics:
  - accuracy
model-index:
  - name: roberta-tiny-8l-10M
    results: []

roberta-tiny-8l-10M

This model was trained from scratch on an unknown dataset. It achieves the following results on the evaluation set:

  • Loss: 7.3389
  • Accuracy: 0.0516

Model description

More information needed

Intended uses & limitations

More information needed

Training and evaluation data

More information needed

Training procedure

Training hyperparameters

The following hyperparameters were used during training:

  • learning_rate: 0.0004
  • train_batch_size: 16
  • eval_batch_size: 32
  • seed: 42
  • gradient_accumulation_steps: 32
  • total_train_batch_size: 512
  • optimizer: Adam with betas=(0.9,0.98) and epsilon=1e-08
  • lr_scheduler_type: cosine
  • lr_scheduler_warmup_steps: 50
  • num_epochs: 100.0
  • mixed_precision_training: Native AMP

Training results

Training Loss Epoch Step Validation Loss Accuracy
7.8102 1.04 50 7.3747 0.0514
7.805 2.08 100 7.3699 0.0517
7.7907 3.12 150 7.3595 0.0517
7.7838 4.16 200 7.3617 0.0514
7.7706 5.21 250 7.3586 0.0514
7.2933 6.25 300 7.3566 0.0513
7.2932 7.29 350 7.3527 0.0516
7.2986 8.33 400 7.3561 0.0516
7.289 9.37 450 7.3495 0.0515
7.2879 10.41 500 7.3455 0.0514
7.276 11.45 550 7.3477 0.0513
7.3072 12.49 600 7.3446 0.0516
7.2978 13.53 650 7.3463 0.0514
7.2857 14.58 700 7.3426 0.0515
7.2868 15.62 750 7.3438 0.0515
7.2973 16.66 800 7.3442 0.0517
7.2988 17.7 850 7.3437 0.0512

Framework versions

  • Transformers 4.24.0
  • Pytorch 1.11.0+cu113
  • Datasets 2.6.1
  • Tokenizers 0.12.1