HachiML's picture
End of training
582b115 verified
metadata
tags:
  - generated_from_trainer
model-index:
  - name: myBit-Llama2-jp-127M-8
    results: []

myBit-Llama2-jp-127M-8

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

  • Loss: 2.8102

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: 0.0024
  • train_batch_size: 96
  • eval_batch_size: 96
  • seed: 42
  • optimizer: Adam with betas=(0.9,0.95) and epsilon=1e-08
  • lr_scheduler_type: linear
  • lr_scheduler_warmup_steps: 5000
  • num_epochs: 1

Training results

Training Loss Epoch Step Validation Loss
4.7094 0.05 2000 3.7099
3.5644 0.1 4000 3.4754
3.4187 0.15 6000 3.3482
3.3026 0.2 8000 3.2653
3.2405 0.25 10000 3.2143
3.1966 0.29 12000 3.1806
3.1666 0.34 14000 3.1533
3.1408 0.39 16000 3.1344
3.12 0.44 18000 3.1123
3.1005 0.49 20000 3.0934
3.0802 0.54 22000 3.0769
3.0629 0.59 24000 3.0545
3.0427 0.64 26000 3.0319
3.0206 0.69 28000 3.0111
3.0008 0.74 30000 2.9897
2.9735 0.79 32000 2.9632
2.9466 0.83 34000 2.9335
2.9165 0.88 36000 2.9039
2.8816 0.93 38000 2.8623
2.8345 0.98 40000 2.8102

Framework versions

  • Transformers 4.38.2
  • Pytorch 2.2.1+cu121
  • Datasets 2.18.0
  • Tokenizers 0.15.2