Llama-360M-RUN1 / README.md
ninagroot's picture
ninagroot/Llama-360Mtest
96f66c0 verified
|
raw
history blame
2.03 kB
metadata
tags:
  - generated_from_trainer
model-index:
  - name: Llama-360M
    results: []

Llama-360M

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

  • Loss: 4.3606

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: 1
  • eval_batch_size: 8
  • seed: 42
  • gradient_accumulation_steps: 8
  • total_train_batch_size: 8
  • optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
  • lr_scheduler_type: cosine
  • lr_scheduler_warmup_steps: 300
  • num_epochs: 15
  • mixed_precision_training: Native AMP

Training results

Training Loss Epoch Step Validation Loss
8.3772 0.99 44 8.2600
7.4813 1.99 88 7.1917
6.7108 2.98 132 6.4281
6.0293 4.0 177 5.8969
5.1002 4.99 221 5.3150
4.5767 5.99 265 4.8598
4.3085 6.98 309 4.6251
4.1028 8.0 354 4.4794
3.765 8.99 398 4.4287
3.6087 9.99 442 4.3823
3.4937 10.98 486 4.3547
3.2339 12.0 531 4.3534
3.0636 12.99 575 4.3613
2.9827 13.99 619 4.3627
3.0038 14.92 660 4.3606

Framework versions

  • Transformers 4.37.2
  • Pytorch 2.1.2+cu121
  • Datasets 2.16.1
  • Tokenizers 0.15.0