---
license: gemma
library_name: peft
tags:
- trl
- reward-trainer
- generated_from_trainer
metrics:
- accuracy
base_model: google/gemma-2b
model-index:
- name: RM-HH-Gemma_harmless_gpt3_20000_gemma2b_shuffleFalse_extractchosenTrue
  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. -->

# RM-HH-Gemma_harmless_gpt3_20000_gemma2b_shuffleFalse_extractchosenTrue

This model is a fine-tuned version of [google/gemma-2b](https://huggingface.co/google/gemma-2b) on the None dataset.
It achieves the following results on the evaluation set:
- Loss: 0.0495
- Accuracy: 0.9820

## 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: 1.41e-05
- train_batch_size: 1
- eval_batch_size: 8
- seed: 42
- gradient_accumulation_steps: 4
- total_train_batch_size: 4
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
- lr_scheduler_type: linear
- num_epochs: 1.0

### Training results

| Training Loss | Epoch | Step | Validation Loss | Accuracy |
|:-------------:|:-----:|:----:|:---------------:|:--------:|
| 0.9068        | 0.03  | 250  | 0.5546          | 0.7177   |
| 0.5566        | 0.06  | 500  | 0.2048          | 0.9170   |
| 0.5143        | 0.08  | 750  | 0.1646          | 0.9370   |
| 0.4865        | 0.11  | 1000 | 0.1396          | 0.9457   |
| 0.4771        | 0.14  | 1250 | 0.1204          | 0.9510   |
| 0.4452        | 0.17  | 1500 | 0.1118          | 0.9565   |
| 0.436         | 0.19  | 1750 | 0.1063          | 0.9570   |
| 0.4433        | 0.22  | 2000 | 0.0942          | 0.9615   |
| 0.4541        | 0.25  | 2250 | 0.0878          | 0.9647   |
| 0.4361        | 0.28  | 2500 | 0.0822          | 0.9672   |
| 0.4626        | 0.31  | 2750 | 0.0766          | 0.9700   |
| 0.4595        | 0.33  | 3000 | 0.0714          | 0.9720   |
| 0.4375        | 0.36  | 3250 | 0.0720          | 0.9715   |
| 0.4338        | 0.39  | 3500 | 0.0693          | 0.9727   |
| 0.4082        | 0.42  | 3750 | 0.0675          | 0.9720   |
| 0.4306        | 0.44  | 4000 | 0.0635          | 0.9745   |
| 0.4296        | 0.47  | 4250 | 0.0629          | 0.9750   |
| 0.4318        | 0.5   | 4500 | 0.0590          | 0.9767   |
| 0.4226        | 0.53  | 4750 | 0.0575          | 0.9775   |
| 0.435         | 0.56  | 5000 | 0.0556          | 0.9785   |
| 0.4501        | 0.58  | 5250 | 0.0557          | 0.9790   |
| 0.3923        | 0.61  | 5500 | 0.0542          | 0.9785   |
| 0.4222        | 0.64  | 5750 | 0.0541          | 0.9790   |
| 0.3891        | 0.67  | 6000 | 0.0538          | 0.9787   |
| 0.4123        | 0.69  | 6250 | 0.0551          | 0.9790   |
| 0.3805        | 0.72  | 6500 | 0.0521          | 0.9805   |
| 0.4269        | 0.75  | 6750 | 0.0529          | 0.9800   |
| 0.382         | 0.78  | 7000 | 0.0530          | 0.9802   |
| 0.422         | 0.81  | 7250 | 0.0517          | 0.9812   |
| 0.4621        | 0.83  | 7500 | 0.0506          | 0.9812   |
| 0.3963        | 0.86  | 7750 | 0.0498          | 0.9820   |
| 0.4097        | 0.89  | 8000 | 0.0495          | 0.9820   |
| 0.4705        | 0.92  | 8250 | 0.0492          | 0.9822   |
| 0.4248        | 0.94  | 8500 | 0.0493          | 0.9820   |
| 0.3938        | 0.97  | 8750 | 0.0495          | 0.9820   |


### Framework versions

- PEFT 0.9.0
- Transformers 4.38.2
- Pytorch 2.1.2
- Datasets 2.18.0
- Tokenizers 0.15.2