---
library_name: peft
tags:
- alignment-handbook
- trl
- dpo
- generated_from_trainer
base_model: NorLLM-AI/NorMistral-7B
datasets:
- hugodk-sch/aftonposten_title_prefs
model-index:
- name: norllm-ai-normistral-7b-align-scan
  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. -->

# norllm-ai-normistral-7b-align-scan

This model is a fine-tuned version of [data/norllm-ai-normistral-7b-sft-qlora](https://huggingface.co/data/norllm-ai-normistral-7b-sft-qlora) on the hugodk-sch/aftonposten_title_prefs dataset.
It achieves the following results on the evaluation set:
- Loss: 0.4787
- Rewards/chosen: -0.2528
- Rewards/rejected: -0.4471
- Rewards/accuracies: 0.5602
- Rewards/margins: 0.1943
- Logps/rejected: -35.4433
- Logps/chosen: -31.7037
- Logits/rejected: -2.8306
- Logits/chosen: -2.8338

## 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: 4
- eval_batch_size: 8
- seed: 42
- distributed_type: multi-GPU
- gradient_accumulation_steps: 2
- total_train_batch_size: 8
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
- lr_scheduler_type: cosine
- lr_scheduler_warmup_ratio: 0.1
- num_epochs: 4

### Training results

| Training Loss | Epoch | Step | Validation Loss | Rewards/chosen | Rewards/rejected | Rewards/accuracies | Rewards/margins | Logps/rejected | Logps/chosen | Logits/rejected | Logits/chosen |
|:-------------:|:-----:|:----:|:---------------:|:--------------:|:----------------:|:------------------:|:---------------:|:--------------:|:------------:|:---------------:|:-------------:|
| 0.4852        | 0.26  | 100  | 0.4894          | 0.0835         | 0.0380           | 0.5689             | 0.0456          | -34.6349       | -31.1432     | -2.8023         | -2.8050       |
| 0.4094        | 0.52  | 200  | 0.4824          | 0.1068         | 0.0177           | 0.5432             | 0.0891          | -34.6687       | -31.1045     | -2.8122         | -2.8150       |
| 0.3408        | 0.78  | 300  | 0.4827          | 0.0196         | -0.0908          | 0.5694             | 0.1105          | -34.8496       | -31.2497     | -2.8188         | -2.8212       |
| 0.1572        | 1.04  | 400  | 0.4798          | -0.0175        | -0.1496          | 0.5839             | 0.1321          | -34.9475       | -31.3117     | -2.8156         | -2.8172       |
| 0.1463        | 1.3   | 500  | 0.4816          | -0.0382        | -0.1668          | 0.5843             | 0.1285          | -34.9761       | -31.3462     | -2.8282         | -2.8294       |
| 0.1315        | 1.56  | 600  | 0.4786          | -0.0124        | -0.1617          | 0.5511             | 0.1493          | -34.9677       | -31.3031     | -2.8038         | -2.8065       |
| 0.1119        | 1.82  | 700  | 0.4778          | -0.0846        | -0.2536          | 0.5893             | 0.1691          | -35.1209       | -31.4234     | -2.8174         | -2.8202       |
| 0.0639        | 2.08  | 800  | 0.4746          | -0.1169        | -0.3129          | 0.5689             | 0.1961          | -35.2197       | -31.4772     | -2.8306         | -2.8335       |
| 0.0635        | 2.34  | 900  | 0.4768          | -0.1886        | -0.3820          | 0.5777             | 0.1934          | -35.3348       | -31.5967     | -2.8313         | -2.8346       |
| 0.0517        | 2.6   | 1000 | 0.4804          | -0.2063        | -0.3767          | 0.5544             | 0.1704          | -35.3260       | -31.6262     | -2.8305         | -2.8336       |
| 0.0446        | 2.86  | 1100 | 0.4794          | -0.2379        | -0.4230          | 0.5689             | 0.1851          | -35.4032       | -31.6790     | -2.8298         | -2.8332       |
| 0.0511        | 3.12  | 1200 | 0.4783          | -0.2347        | -0.4328          | 0.5806             | 0.1982          | -35.4196       | -31.6736     | -2.8297         | -2.8330       |
| 0.0414        | 3.38  | 1300 | 0.4778          | -0.2496        | -0.4539          | 0.5660             | 0.2043          | -35.4547       | -31.6984     | -2.8309         | -2.8340       |
| 0.0334        | 3.64  | 1400 | 0.4783          | -0.2495        | -0.4524          | 0.5777             | 0.2028          | -35.4521       | -31.6984     | -2.8308         | -2.8338       |
| 0.0416        | 3.9   | 1500 | 0.4785          | -0.2460        | -0.4418          | 0.5631             | 0.1958          | -35.4345       | -31.6925     | -2.8303         | -2.8337       |


### Framework versions

- PEFT 0.10.0
- Transformers 4.39.0.dev0
- Pytorch 2.1.2+cu121
- Datasets 2.14.6
- Tokenizers 0.15.1