---
license: other
library_name: peft
tags:
- generated_from_trainer
base_model: beomi/gemma-ko-2b
model-index:
- name: gemma2_on_korean_events
  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. -->

# gemma2_on_korean_events

This model is a fine-tuned version of [beomi/gemma-ko-2b](https://huggingface.co/beomi/gemma-ko-2b) on the None dataset.
It achieves the following results on the evaluation set:
- Loss: 0.4946

## 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-05
- train_batch_size: 2
- eval_batch_size: 2
- seed: 42
- gradient_accumulation_steps: 5
- total_train_batch_size: 10
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
- lr_scheduler_type: linear
- lr_scheduler_warmup_steps: 50
- training_steps: 760
- mixed_precision_training: Native AMP

### Training results

| Training Loss | Epoch | Step | Validation Loss |
|:-------------:|:-----:|:----:|:---------------:|
| 1.4589        | 0.26  | 20   | 1.1191          |
| 0.8348        | 0.53  | 40   | 0.5145          |
| 0.5149        | 0.79  | 60   | 0.3695          |
| 0.3757        | 1.05  | 80   | 0.3677          |
| 0.3665        | 1.32  | 100  | 0.3245          |
| 0.3607        | 1.58  | 120  | 0.3221          |
| 0.3277        | 1.84  | 140  | 0.3016          |
| 0.2741        | 2.11  | 160  | 0.2971          |
| 0.2529        | 2.37  | 180  | 0.3052          |
| 0.255         | 2.63  | 200  | 0.3021          |
| 0.2727        | 2.89  | 220  | 0.2888          |
| 0.1846        | 3.16  | 240  | 0.3174          |
| 0.1775        | 3.42  | 260  | 0.3055          |
| 0.1586        | 3.68  | 280  | 0.3140          |
| 0.1444        | 3.95  | 300  | 0.3146          |
| 0.1134        | 4.21  | 320  | 0.3357          |
| 0.1031        | 4.47  | 340  | 0.3334          |
| 0.115         | 4.74  | 360  | 0.3208          |
| 0.1012        | 5.0   | 380  | 0.3350          |
| 0.0613        | 5.26  | 400  | 0.3807          |
| 0.0548        | 5.53  | 420  | 0.3641          |
| 0.0642        | 5.79  | 440  | 0.3625          |
| 0.0575        | 6.05  | 460  | 0.3730          |
| 0.033         | 6.32  | 480  | 0.4076          |
| 0.0369        | 6.58  | 500  | 0.3900          |
| 0.0357        | 6.84  | 520  | 0.3915          |
| 0.0265        | 7.11  | 540  | 0.4345          |
| 0.0241        | 7.37  | 560  | 0.4440          |
| 0.0184        | 7.63  | 580  | 0.4233          |
| 0.0167        | 7.89  | 600  | 0.4478          |
| 0.0168        | 8.16  | 620  | 0.4914          |
| 0.0097        | 8.42  | 640  | 0.4625          |
| 0.012         | 8.68  | 660  | 0.4916          |
| 0.0116        | 8.95  | 680  | 0.4730          |
| 0.0077        | 9.21  | 700  | 0.4892          |
| 0.0083        | 9.47  | 720  | 0.4939          |
| 0.0072        | 9.74  | 740  | 0.4902          |
| 0.0077        | 10.0  | 760  | 0.4946          |


### Framework versions

- PEFT 0.11.1
- Transformers 4.38.2
- Pytorch 2.1.2+cu121
- Datasets 2.17.0
- Tokenizers 0.15.0