metadata
license: apache-2.0
base_model: alignment-handbook/zephyr-7b-sft-full
tags:
- trl
- dpo
- generated_from_trainer
model-index:
- name: zephyr-7b-dpo-full
results: []
zephyr-7b-dpo-full
This model is a fine-tuned version of alignment-handbook/zephyr-7b-sft-full on the None dataset. It achieves the following results on the evaluation set:
- Loss: 0.5714
- Rewards/chosen: -0.4765
- Rewards/rejected: -1.7868
- Rewards/accuracies: 0.7619
- Rewards/margins: 1.3103
- Logps/rejected: -294.2741
- Logps/chosen: -305.4321
- Logits/rejected: -2.4275
- Logits/chosen: -2.4841
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: 8
- eval_batch_size: 8
- seed: 42
- distributed_type: multi-GPU
- num_devices: 4
- total_train_batch_size: 32
- total_eval_batch_size: 32
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
- lr_scheduler_type: linear
- lr_scheduler_warmup_ratio: 0.1
- num_epochs: 1
Training results
| Training Loss | Epoch | Step | Validation Loss | Rewards/chosen | Rewards/rejected | Rewards/accuracies | Rewards/margins | Logps/rejected | Logps/chosen | Logits/rejected | Logits/chosen |
|---|---|---|---|---|---|---|---|---|---|---|---|
| 0.5509 | 0.26 | 500 | 0.6311 | 0.8114 | -0.1181 | 0.7619 | 0.9294 | -277.5868 | -292.5533 | -2.4249 | -2.4869 |
| 0.6191 | 0.52 | 1000 | 0.5984 | 0.0381 | -1.0554 | 0.7599 | 1.0935 | -286.9601 | -300.2859 | -2.4280 | -2.4907 |
| 0.5342 | 0.78 | 1500 | 0.5819 | 0.0994 | -1.1560 | 0.7560 | 1.2554 | -287.9663 | -299.6732 | -2.4208 | -2.4793 |
Framework versions
- Transformers 4.36.1
- Pytorch 2.0.1+cu117
- Datasets 2.16.1
- Tokenizers 0.15.0