---
library_name: peft
tags:
- trl
- dpo
- DPO
- WeniGPT
- generated_from_trainer
base_model: Weni/WeniGPT-Agents-Mistral-1.0.11-SFT-merged
model-index:
- name: WeniGPT-Agents-Mistral-1.0.11-SFT-1.0.27-DPO
  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. -->

# WeniGPT-Agents-Mistral-1.0.11-SFT-1.0.27-DPO

This model is a fine-tuned version of [Weni/WeniGPT-Agents-Mistral-1.0.11-SFT-merged](https://huggingface.co/Weni/WeniGPT-Agents-Mistral-1.0.11-SFT-merged) on an unknown dataset.
It achieves the following results on the evaluation set:
- Loss: 0.0855
- Rewards/chosen: 1.8718
- Rewards/rejected: -1.2443
- Rewards/accuracies: 1.0
- Rewards/margins: 3.1161
- Logps/rejected: -191.7649
- Logps/chosen: -127.9738
- Logits/rejected: -1.8982
- Logits/chosen: -1.8696

## 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: 1
- seed: 42
- distributed_type: multi-GPU
- num_devices: 4
- gradient_accumulation_steps: 2
- total_train_batch_size: 8
- total_eval_batch_size: 4
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
- lr_scheduler_type: linear
- lr_scheduler_warmup_ratio: 0.03
- training_steps: 180
- 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 |
|:-------------:|:------:|:----:|:---------------:|:--------------:|:----------------:|:------------------:|:---------------:|:--------------:|:------------:|:---------------:|:-------------:|
| 0.5138        | 0.9677 | 30   | 0.4536          | 0.5540         | -0.0631          | 1.0                | 0.6171          | -179.9529      | -141.1520    | -1.8497         | -1.8200       |
| 0.2915        | 1.9355 | 60   | 0.2489          | 1.1740         | -0.2675          | 1.0                | 1.4414          | -181.9961      | -134.9520    | -1.8636         | -1.8333       |
| 0.1995        | 2.9032 | 90   | 0.1573          | 1.5574         | -0.6071          | 1.0                | 2.1646          | -185.3929      | -131.1177    | -1.8785         | -1.8481       |
| 0.1172        | 3.8710 | 120  | 0.1100          | 1.7744         | -0.9240          | 1.0                | 2.6984          | -188.5615      | -128.9482    | -1.8899         | -1.8606       |
| 0.1162        | 4.8387 | 150  | 0.0910          | 1.8558         | -1.1622          | 1.0                | 3.0180          | -190.9434      | -128.1337    | -1.8967         | -1.8678       |
| 0.1057        | 5.8065 | 180  | 0.0855          | 1.8718         | -1.2443          | 1.0                | 3.1161          | -191.7649      | -127.9738    | -1.8982         | -1.8696       |


### Framework versions

- PEFT 0.10.0
- Transformers 4.40.0
- Pytorch 2.1.0+cu118
- Datasets 2.18.0
- Tokenizers 0.19.1