Angelectronic's picture
End of training
78ea665 verified
metadata
license: apache-2.0
library_name: peft
tags:
  - trl
  - sft
  - unsloth
  - generated_from_trainer
base_model: unsloth/llama-3-8b-Instruct-bnb-4bit
model-index:
  - name: llama3-chat_10000_500
    results: []

llama3-chat_10000_500

This model is a fine-tuned version of unsloth/llama-3-8b-Instruct-bnb-4bit on the None dataset. It achieves the following results on the evaluation set:

  • Loss: 1.8491

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: 0.0002
  • train_batch_size: 8
  • eval_batch_size: 4
  • seed: 3407
  • gradient_accumulation_steps: 4
  • total_train_batch_size: 32
  • optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
  • lr_scheduler_type: cosine
  • lr_scheduler_warmup_steps: 5
  • num_epochs: 5
  • mixed_precision_training: Native AMP

Training results

Training Loss Epoch Step Validation Loss
1.765 0.33 104 1.4841
1.5146 0.67 208 1.4604
1.4912 1.0 312 1.4545
1.3584 1.33 416 1.4698
1.358 1.66 520 1.4671
1.3483 2.0 624 1.4637
1.1105 2.33 728 1.5471
1.101 2.66 832 1.5512
1.1007 3.0 936 1.5522
0.8526 3.33 1040 1.7081
0.8445 3.66 1144 1.7156
0.8463 3.99 1248 1.7115
0.6865 4.33 1352 1.8423
0.6811 4.66 1456 1.8458
0.6859 4.99 1560 1.8491

Framework versions

  • PEFT 0.10.0
  • Transformers 4.39.3
  • Pytorch 2.2.2+cu121
  • Datasets 2.16.0
  • Tokenizers 0.15.2