---
license: gemma
library_name: peft
tags:
- trl
- sft
- alignment-handbook
- generated_from_trainer
base_model: google/gemma-2b
datasets:
- generator
model-index:
- name: gemma2b-summarize-gpt4o-64k
  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. -->

# gemma2b-summarize-gpt4o-64k

This model is a fine-tuned version of [google/gemma-2b](https://huggingface.co/google/gemma-2b) on the generator dataset.
It achieves the following results on the evaluation set:
- Loss: 2.6852

## 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: 0.0002
- train_batch_size: 16
- eval_batch_size: 16
- seed: 42
- distributed_type: multi-GPU
- num_devices: 3
- gradient_accumulation_steps: 2
- total_train_batch_size: 96
- total_eval_batch_size: 48
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
- lr_scheduler_type: cosine
- lr_scheduler_warmup_ratio: 0.1
- num_epochs: 10

### Training results

| Training Loss | Epoch | Step | Validation Loss |
|:-------------:|:-----:|:----:|:---------------:|
| 1.2474        | 1.0   | 146  | 2.5237          |
| 1.1269        | 2.0   | 292  | 2.4805          |
| 1.0909        | 3.0   | 438  | 2.4893          |
| 1.0354        | 4.0   | 584  | 2.5017          |
| 1.0016        | 5.0   | 730  | 2.5295          |
| 0.9823        | 6.0   | 876  | 2.5500          |
| 0.955         | 7.0   | 1022 | 2.5866          |
| 0.9214        | 8.0   | 1168 | 2.6224          |
| 0.913         | 9.0   | 1314 | 2.6512          |
| 0.889         | 10.0  | 1460 | 2.6852          |


### Framework versions

- PEFT 0.11.1
- Transformers 4.41.2
- Pytorch 2.3.0+cu121
- Datasets 2.19.2
- Tokenizers 0.19.1