|
--- |
|
license: other |
|
library_name: peft |
|
tags: |
|
- generated_from_trainer |
|
base_model: beomi/gemma-ko-2b |
|
model-index: |
|
- name: gemma2_on_korean_summary |
|
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_summary |
|
|
|
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: 0.9622 |
|
|
|
## 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: 800 |
|
- mixed_precision_training: Native AMP |
|
|
|
### Training results |
|
|
|
| Training Loss | Epoch | Step | Validation Loss | |
|
|:-------------:|:-----:|:----:|:---------------:| |
|
| 1.6606 | 0.26 | 20 | 1.5454 | |
|
| 1.4381 | 0.53 | 40 | 1.3247 | |
|
| 1.2548 | 0.79 | 60 | 1.1921 | |
|
| 1.1574 | 1.05 | 80 | 1.1227 | |
|
| 1.0968 | 1.32 | 100 | 1.0727 | |
|
| 1.0485 | 1.58 | 120 | 1.0316 | |
|
| 1.0258 | 1.84 | 140 | 1.0019 | |
|
| 0.9582 | 2.11 | 160 | 0.9785 | |
|
| 0.906 | 2.37 | 180 | 0.9575 | |
|
| 0.8837 | 2.63 | 200 | 0.9429 | |
|
| 0.8763 | 2.89 | 220 | 0.9247 | |
|
| 0.8295 | 3.16 | 240 | 0.9213 | |
|
| 0.7799 | 3.42 | 260 | 0.9122 | |
|
| 0.7742 | 3.68 | 280 | 0.8992 | |
|
| 0.7708 | 3.95 | 300 | 0.8918 | |
|
| 0.7196 | 4.21 | 320 | 0.8952 | |
|
| 0.6908 | 4.47 | 340 | 0.8917 | |
|
| 0.6977 | 4.74 | 360 | 0.8841 | |
|
| 0.6789 | 5.0 | 380 | 0.8764 | |
|
| 0.6198 | 5.26 | 400 | 0.9003 | |
|
| 0.6203 | 5.53 | 420 | 0.9030 | |
|
| 0.6169 | 5.79 | 440 | 0.8913 | |
|
| 0.6111 | 6.05 | 460 | 0.8935 | |
|
| 0.564 | 6.32 | 480 | 0.9096 | |
|
| 0.5819 | 6.58 | 500 | 0.9027 | |
|
| 0.5673 | 6.84 | 520 | 0.8997 | |
|
| 0.5382 | 7.11 | 540 | 0.9181 | |
|
| 0.5228 | 7.37 | 560 | 0.9197 | |
|
| 0.5197 | 7.63 | 580 | 0.9254 | |
|
| 0.5319 | 7.89 | 600 | 0.9158 | |
|
| 0.4954 | 8.16 | 620 | 0.9394 | |
|
| 0.4954 | 8.42 | 640 | 0.9345 | |
|
| 0.4719 | 8.68 | 660 | 0.9388 | |
|
| 0.4866 | 8.95 | 680 | 0.9390 | |
|
| 0.4739 | 9.21 | 700 | 0.9528 | |
|
| 0.4591 | 9.47 | 720 | 0.9534 | |
|
| 0.4478 | 9.74 | 740 | 0.9542 | |
|
| 0.4547 | 10.0 | 760 | 0.9539 | |
|
| 0.4379 | 10.26 | 780 | 0.9617 | |
|
| 0.4348 | 10.53 | 800 | 0.9622 | |
|
|
|
|
|
### Framework versions |
|
|
|
- PEFT 0.8.2 |
|
- Transformers 4.38.2 |
|
- Pytorch 2.1.2+cu121 |
|
- Datasets 2.17.0 |
|
- Tokenizers 0.15.0 |
