---
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-magpi-reward-scale-1
  results: []
---

<!-- This model card has been generated automatically according to the information the Trainer had access to. You
should probably proofread and complete it, then remove this comment. -->

# zephyr-7b-dpo-full-magpi-reward-scale-1

This model is a fine-tuned version of [alignment-handbook/zephyr-7b-sft-full](https://huggingface.co/alignment-handbook/zephyr-7b-sft-full) on an unknown dataset.
It achieves the following results on the evaluation set:
- Loss: 0.0001
- Rewards/chosen: -2.4218
- Rewards/rejected: -70.3140
- Rewards/accuracies: 1.0
- Rewards/margins: 67.8922
- Logps/rejected: -7672.1890
- Logps/chosen: -609.1660
- Logits/rejected: 2.5753
- Logits/chosen: -0.2188

## 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: 55
- distributed_type: multi-GPU
- num_devices: 8
- gradient_accumulation_steps: 2
- total_train_batch_size: 128
- total_eval_batch_size: 64
- 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

### Training results

| Training Loss | Epoch  | Step | Validation Loss | Rewards/chosen | Rewards/rejected | Rewards/accuracies | Rewards/margins | Logps/rejected | Logps/chosen | Logits/rejected | Logits/chosen |
|:-------------:|:------:|:----:|:---------------:|:--------------:|:----------------:|:------------------:|:---------------:|:--------------:|:------------:|:---------------:|:-------------:|
| 0.0055        | 0.1420 | 50   | 0.0031          | -0.8850        | -42.4340         | 0.9960             | 41.5490         | -4884.1895     | -455.4866    | -2.5280         | -2.9176       |
| 0.0014        | 0.2841 | 100  | 0.0003          | -1.7467        | -57.2050         | 1.0                | 55.4583         | -6361.2881     | -541.6496    | -0.1436         | -2.5818       |
| 0.001         | 0.4261 | 150  | 0.0002          | -2.4059        | -66.6788         | 1.0                | 64.2728         | -7308.6660     | -607.5745    | 1.9422          | -1.9097       |
| 0.0018        | 0.5682 | 200  | 0.0002          | -2.1797        | -67.6111         | 1.0                | 65.4314         | -7401.8965     | -584.9550    | 2.3189          | -1.0395       |
| 0.0009        | 0.7102 | 250  | 0.0001          | -2.4169        | -67.5787         | 1.0                | 65.1618         | -7398.6553     | -608.6732    | 2.5585          | -0.3354       |
| 0.0009        | 0.8523 | 300  | 0.0001          | -2.4125        | -70.2443         | 1.0                | 67.8319         | -7665.2217     | -608.2272    | 2.5751          | -0.2549       |
| 0.0024        | 0.9943 | 350  | 0.0001          | -2.4218        | -70.3140         | 1.0                | 67.8922         | -7672.1890     | -609.1660    | 2.5753          | -0.2188       |


### Framework versions

- Transformers 4.44.0.dev0
- Pytorch 2.1.2
- Datasets 2.20.0
- Tokenizers 0.19.1