metadata
license: apache-2.0
base_model: mistralai/Mistral-7B-v0.1
tags:
- generated_from_trainer
model-index:
- name: zephyr-7b-dpo-lora
results: []
zephyr-7b-dpo-lora
This model is a fine-tuned version of mistralai/Mistral-7B-v0.1 on the None dataset. It achieves the following results on the evaluation set:
- Loss: 0.5264
- Rewards/chosen: -0.1560
- Rewards/rejected: -0.9014
- Rewards/accuracies: 0.7345
- Rewards/margins: 0.7454
- Logps/rejected: -232.6823
- Logps/chosen: -266.1183
- Logits/rejected: -1.9666
- Logits/chosen: -2.0208
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-07
- train_batch_size: 2
- eval_batch_size: 4
- seed: 42
- gradient_accumulation_steps: 32
- total_train_batch_size: 64
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
- lr_scheduler_type: linear
- lr_scheduler_warmup_ratio: 0.1
- num_epochs: 3
Training results
| Training Loss | Epoch | Step | Validation Loss | Rewards/chosen | Rewards/rejected | Rewards/accuracies | Rewards/margins | Logps/rejected | Logps/chosen | Logits/rejected | Logits/chosen |
|---|---|---|---|---|---|---|---|---|---|---|---|
| 0.5648 | 1.0 | 968 | 0.5548 | -0.1107 | -0.6467 | 0.7115 | 0.5360 | -230.1347 | -265.6653 | -2.0150 | -2.0685 |
| 0.5563 | 2.0 | 1936 | 0.5313 | -0.1530 | -0.8621 | 0.7315 | 0.7091 | -232.2889 | -266.0887 | -1.9774 | -2.0312 |
| 0.5172 | 3.0 | 2904 | 0.5264 | -0.1560 | -0.9014 | 0.7345 | 0.7454 | -232.6823 | -266.1183 | -1.9666 | -2.0208 |
Framework versions
- Transformers 4.35.0
- Pytorch 2.1.1+cu121
- Datasets 2.14.6
- Tokenizers 0.14.1