metadata
base_model: meta-llama/Meta-Llama-3.1-8B
library_name: peft
license: llama3.1
tags:
- trl
- dpo
- generated_from_trainer
model-index:
- name: dpo-llama3-8b-sample-rules
results: []
dpo-llama3-8b-sample-rules
This model is a fine-tuned version of meta-llama/Meta-Llama-3.1-8B on the None dataset. It achieves the following results on the evaluation set:
- Loss: 0.1841
- Rewards/chosen: 0.0866
- Rewards/rejected: -1.6352
- Rewards/accuracies: 1.0
- Rewards/margins: 1.7218
- Logps/rejected: -545.1958
- Logps/chosen: -202.5652
- Logits/rejected: -1.3753
- Logits/chosen: -1.2073
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: 5e-06
- 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: 1
- mixed_precision_training: Native AMP
Training results
| Training Loss | Epoch | Step | Validation Loss | Rewards/chosen | Rewards/rejected | Rewards/accuracies | Rewards/margins | Logps/rejected | Logps/chosen | Logits/rejected | Logits/chosen |
|---|---|---|---|---|---|---|---|---|---|---|---|
| 0.4281 | 0.4444 | 50 | 0.3902 | 0.1843 | -0.5815 | 1.0 | 0.7658 | -439.8294 | -192.8043 | -1.3883 | -1.2422 |
| 0.193 | 0.8889 | 100 | 0.1841 | 0.0866 | -1.6352 | 1.0 | 1.7218 | -545.1958 | -202.5652 | -1.3753 | -1.2073 |
Framework versions
- PEFT 0.12.0
- Transformers 4.44.0
- Pytorch 2.3.1+cu121
- Datasets 2.21.0
- Tokenizers 0.19.1