metadata
license: other
library_name: peft
tags:
- generated_from_trainer
base_model: beomi/gemma-ko-2b
model-index:
- name: gemma2_on_korean_conv-base-stm
results: []
gemma2_on_korean_conv-base-stm
This model is a fine-tuned version of beomi/gemma-ko-2b on the None dataset. It achieves the following results on the evaluation set:
- Loss: 1.1954
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: 10
- total_train_batch_size: 20
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
- lr_scheduler_type: linear
- lr_scheduler_warmup_steps: 100
- training_steps: 2000
Training results
Training Loss | Epoch | Step | Validation Loss |
---|---|---|---|
1.4804 | 0.2563 | 100 | 1.4702 |
1.3147 | 0.5126 | 200 | 1.3067 |
1.2424 | 0.7688 | 300 | 1.2374 |
1.0918 | 1.0251 | 400 | 1.1792 |
0.9923 | 1.2814 | 500 | 1.1563 |
0.9586 | 1.5377 | 600 | 1.1370 |
0.9639 | 1.7940 | 700 | 1.1093 |
0.7698 | 2.0502 | 800 | 1.1266 |
0.7274 | 2.3065 | 900 | 1.1393 |
0.7416 | 2.5628 | 1000 | 1.1022 |
0.759 | 2.8191 | 1100 | 1.0903 |
0.5755 | 3.0753 | 1200 | 1.1302 |
0.5839 | 3.3316 | 1300 | 1.1443 |
0.6008 | 3.5879 | 1400 | 1.1300 |
0.5773 | 3.8442 | 1500 | 1.1326 |
0.4565 | 4.1005 | 1600 | 1.1757 |
0.4436 | 4.3567 | 1700 | 1.1832 |
0.4462 | 4.6130 | 1800 | 1.1816 |
0.4502 | 4.8693 | 1900 | 1.1878 |
0.3731 | 5.1256 | 2000 | 1.1954 |
Framework versions
- PEFT 0.11.1
- Transformers 4.41.1
- Pytorch 2.3.0+cu121
- Datasets 2.19.1
- Tokenizers 0.19.1