---
library_name: transformers
base_model: W-61/llama-3-8b-base-sft-ultrachat-8xh200
tags:
- alignment-handbook
- new-dpo
- generated_from_trainer
datasets:
- HuggingFaceH4/ultrafeedback_binarized
model-index:
- name: llama-3-8b-base-new-dpo-ultrafeedback-4xh200-batch-128-q_t-0.45-s_star-0.35-20260428-045924
  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. -->

# llama-3-8b-base-new-dpo-ultrafeedback-4xh200-batch-128-q_t-0.45-s_star-0.35-20260428-045924

This model is a fine-tuned version of [W-61/llama-3-8b-base-sft-ultrachat-8xh200](https://huggingface.co/W-61/llama-3-8b-base-sft-ultrachat-8xh200) on the HuggingFaceH4/ultrafeedback_binarized dataset.
It achieves the following results on the evaluation set:
- Loss: 0.5985
- Fcm Dpo/beta: 0.0028
- Margin Dpo/margin Mean: 99.3391
- Margin Dpo/margin Std: 172.5638
- Logps/chosen: -539.1735
- Logps/rejected: -617.6157
- Logps/ref Chosen: -287.8268
- Logps/ref Rejected: -266.9300
- Kl/chosen Kl Mean: -251.3467
- Kl/rejected Kl Mean: -350.6858
- Kl/mean: -301.0162
- Kl/std: 151.9403
- Logits/chosen: -0.8488
- Logits/rejected: -0.8314

## 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: 4
- eval_batch_size: 2
- seed: 42
- distributed_type: multi-GPU
- num_devices: 4
- gradient_accumulation_steps: 8
- total_train_batch_size: 128
- total_eval_batch_size: 8
- optimizer: Use OptimizerNames.ADAMW_TORCH with betas=(0.9,0.999) and epsilon=1e-08 and optimizer_args=No additional optimizer arguments
- lr_scheduler_type: cosine
- lr_scheduler_warmup_ratio: 0.1
- num_epochs: 1

### Training results

| Training Loss | Epoch  | Step | Validation Loss | Fcm Dpo/beta | Margin Dpo/margin Mean | Margin Dpo/margin Std | Logps/chosen | Logps/rejected | Logps/ref Chosen | Logps/ref Rejected | Kl/chosen Kl Mean | Kl/rejected Kl Mean | Kl/mean   | Kl/std   | Logits/chosen | Logits/rejected |
|:-------------:|:------:|:----:|:---------------:|:------------:|:----------------------:|:---------------------:|:------------:|:--------------:|:----------------:|:------------------:|:-----------------:|:-------------------:|:---------:|:--------:|:-------------:|:---------------:|
| 4.959         | 0.4188 | 200  | 0.5908          | 0.0053       | 56.5030                | 93.3615               | -400.3537    | -435.9599      | -287.8268        | -266.9300          | -112.5269         | -169.0299           | -140.7784 | 83.7769  | -0.8930       | -0.8754         |
| 4.8212        | 0.8377 | 400  | 0.5985          | 0.0028       | 99.3391                | 172.5638              | -539.1735    | -617.6157      | -287.8268        | -266.9300          | -251.3467         | -350.6858           | -301.0162 | 151.9403 | -0.8488       | -0.8314         |


### Framework versions

- Transformers 4.51.0
- Pytorch 2.3.1+cu121
- Datasets 2.21.0
- Tokenizers 0.21.4