---
license: apache-2.0
library_name: peft
tags:
- generated_from_trainer
base_model: mnoukhov/pythia410m-sft-tldr
model-index:
- name: pythia410m-dpo2-tldr
  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. -->

# pythia410m-dpo2-tldr

This model is a fine-tuned version of [mnoukhov/pythia410m-sft-tldr](https://huggingface.co/mnoukhov/pythia410m-sft-tldr) on an unknown dataset.
It achieves the following results on the evaluation set:
- Loss: 0.7753
- Rewards/chosen: -6.3555
- Rewards/rejected: -6.7803
- Rewards/accuracies: 0.5989
- Rewards/margins: 0.4248
- Logps/rejected: -192.7698
- Logps/chosen: -192.7698
- Logps/ref Rejected: -59.5615
- Logps/ref Chosen: -65.6594

## 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: 1e-05
- train_batch_size: 4
- eval_batch_size: 8
- seed: 42
- distributed_type: multi-GPU
- num_devices: 4
- gradient_accumulation_steps: 4
- total_train_batch_size: 64
- total_eval_batch_size: 32
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
- lr_scheduler_type: cosine
- num_epochs: 1.0
- 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 | Logps/ref Rejected | Logps/ref Chosen |
|:-------------:|:------:|:----:|:---------------:|:--------------:|:----------------:|:------------------:|:---------------:|:--------------:|:------------:|:------------------:|:----------------:|
| 0.412         | 0.1999 | 335  | 0.6786          | -4.5428        | -4.9111          | 0.6222             | 0.3683          | -156.5151      | -156.5151    | -59.5615           | -65.6594         |
| 0.3588        | 0.3999 | 670  | 0.7264          | -5.6339        | -6.0502          | 0.6107             | 0.4163          | -178.3372      | -178.3372    | -59.5615           | -65.6594         |
| 0.345         | 0.5998 | 1005 | 0.7470          | -6.0062        | -6.4506          | 0.6086             | 0.4444          | -185.7831      | -185.7831    | -59.5615           | -65.6594         |
| 0.3419        | 0.7998 | 1340 | 0.7738          | -6.3469        | -6.7796          | 0.6012             | 0.4327          | -192.5978      | -192.5978    | -59.5615           | -65.6594         |
| 0.3384        | 0.9997 | 1675 | 0.7753          | -6.3555        | -6.7803          | 0.5989             | 0.4248          | -192.7698      | -192.7698    | -59.5615           | -65.6594         |


### Framework versions

- PEFT 0.10.0
- Transformers 4.40.2
- Pytorch 2.2.1+cu121
- Datasets 2.19.1
- Tokenizers 0.19.1