V0409MP2 / README.md
Litzy619's picture
End of training
94e73f4 verified
metadata
license: mit
base_model: microsoft/phi-2
tags:
  - generated_from_trainer
model-index:
  - name: V0409MP2
    results: []

V0409MP2

This model is a fine-tuned version of microsoft/phi-2 on an unknown dataset. It achieves the following results on the evaluation set:

  • Loss: 0.3412

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: 0.0003
  • train_batch_size: 8
  • eval_batch_size: 8
  • seed: 42
  • gradient_accumulation_steps: 16
  • total_train_batch_size: 128
  • optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
  • lr_scheduler_type: cosine_with_restarts
  • lr_scheduler_warmup_steps: 20
  • num_epochs: 2
  • mixed_precision_training: Native AMP

Training results

Training Loss Epoch Step Validation Loss
6.1578 0.09 10 5.3816
5.3721 0.18 20 4.1461
3.968 0.27 30 2.7849
2.7382 0.36 40 1.8394
1.863 0.45 50 1.2646
1.3779 0.54 60 0.9405
1.0695 0.63 70 0.7297
0.8284 0.73 80 0.5808
0.6698 0.82 90 0.4740
0.5725 0.91 100 0.3968
0.4905 1.0 110 0.3449
0.4426 1.09 120 0.3412
0.4444 1.18 130 0.3414
0.4746 1.27 140 0.3414
0.4367 1.36 150 0.3413
0.4518 1.45 160 0.3414
0.4524 1.54 170 0.3413
0.4507 1.63 180 0.3414
0.4424 1.72 190 0.3413
0.4429 1.81 200 0.3414
0.4571 1.9 210 0.3414
0.4601 1.99 220 0.3412

Framework versions

  • Transformers 4.36.0.dev0
  • Pytorch 2.1.2+cu121
  • Datasets 2.14.6
  • Tokenizers 0.14.1