---
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_extractchosenFalse
  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_extractchosenFalse

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: 0.0236
- Accuracy: 0.9907

## 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 |
|:-------------:|:-----:|:----:|:---------------:|:--------:|
| 0.5999        | 0.03  | 250  | 0.1410          | 0.9717   |
| 0.4197        | 0.06  | 500  | 0.0394          | 0.9882   |
| 0.4246        | 0.08  | 750  | 0.0347          | 0.9895   |
| 0.4616        | 0.11  | 1000 | 0.0345          | 0.9885   |
| 0.4141        | 0.14  | 1250 | 0.0308          | 0.9900   |
| 0.3989        | 0.17  | 1500 | 0.0311          | 0.9887   |
| 0.4122        | 0.19  | 1750 | 0.0299          | 0.9895   |
| 0.4106        | 0.22  | 2000 | 0.0298          | 0.9892   |
| 0.4657        | 0.25  | 2250 | 0.0270          | 0.9905   |
| 0.4311        | 0.28  | 2500 | 0.0304          | 0.9890   |
| 0.4474        | 0.31  | 2750 | 0.0277          | 0.9905   |
| 0.4202        | 0.33  | 3000 | 0.0293          | 0.9892   |
| 0.4487        | 0.36  | 3250 | 0.0287          | 0.9902   |
| 0.4219        | 0.39  | 3500 | 0.0257          | 0.9910   |
| 0.4525        | 0.42  | 3750 | 0.0264          | 0.9910   |
| 0.3805        | 0.44  | 4000 | 0.0277          | 0.9897   |
| 0.3824        | 0.47  | 4250 | 0.0241          | 0.9910   |
| 0.4217        | 0.5   | 4500 | 0.0235          | 0.9912   |
| 0.4275        | 0.53  | 4750 | 0.0259          | 0.9905   |
| 0.4395        | 0.56  | 5000 | 0.0247          | 0.9910   |
| 0.3848        | 0.58  | 5250 | 0.0250          | 0.9910   |
| 0.4297        | 0.61  | 5500 | 0.0249          | 0.9900   |
| 0.4167        | 0.64  | 5750 | 0.0258          | 0.9892   |
| 0.4205        | 0.67  | 6000 | 0.0244          | 0.9902   |
| 0.4072        | 0.69  | 6250 | 0.0264          | 0.9890   |
| 0.4033        | 0.72  | 6500 | 0.0253          | 0.9892   |
| 0.3699        | 0.75  | 6750 | 0.0244          | 0.9905   |
| 0.4101        | 0.78  | 7000 | 0.0259          | 0.9887   |
| 0.3969        | 0.81  | 7250 | 0.0249          | 0.9892   |
| 0.3845        | 0.83  | 7500 | 0.0236          | 0.9907   |
| 0.4208        | 0.86  | 7750 | 0.0232          | 0.9907   |
| 0.3925        | 0.89  | 8000 | 0.0232          | 0.9907   |
| 0.3769        | 0.92  | 8250 | 0.0231          | 0.9912   |
| 0.4323        | 0.94  | 8500 | 0.0232          | 0.9912   |
| 0.3999        | 0.97  | 8750 | 0.0236          | 0.9907   |


### Framework versions

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