metadata
base_model: meta-llama/Meta-Llama-3-8B-Instruct
library_name: peft
license: llama3
tags:
- trl
- kto
- generated_from_trainer
model-index:
- name: 'null'
results: []
null
This model is a fine-tuned version of meta-llama/Meta-Llama-3-8B-Instruct on the None dataset. It achieves the following results on the evaluation set:
- Loss: 0.5456
- Eval/rewards/chosen: 3.9143
- Eval/logps/chosen: -136.1776
- Eval/rewards/rejected: 3.5677
- Eval/logps/rejected: -182.5892
- Eval/rewards/margins: 0.3466
- Eval/kl: 35.7288
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: 0.0001
- train_batch_size: 1
- eval_batch_size: 2
- 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_ratio: 0.1
- num_epochs: 5.0
Training results
| Training Loss | Epoch | Step | Validation Loss | |
|---|---|---|---|---|
| 0.6027 | 0.9677 | 15 | 0.5987 | 1.5787 |
| 0.2879 | 2.0 | 31 | 0.6263 | 29.2556 |
| 0.2962 | 2.9677 | 46 | 0.5909 | 33.4994 |
| 0.132 | 4.0 | 62 | 0.5446 | 35.5494 |
| 0.2602 | 4.8387 | 75 | 0.5456 | 35.7288 |
Framework versions
- PEFT 0.11.1
- Transformers 4.42.2
- Pytorch 2.2.0
- Datasets 2.20.0
- Tokenizers 0.19.1