metadata
library_name: transformers
base_model: halcyon-llm/SmolLM2-360M-japanese_patch-11000
tags:
- generated_from_trainer
datasets:
- kajuma/training_01-09_token
model-index:
- name: results
results: []
results
This model is a fine-tuned version of halcyon-llm/SmolLM2-360M-japanese_patch-11000 on the kajuma/training_01-09_token dataset. It achieves the following results on the evaluation set:
- Loss: 1.1685
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: 0.003
- train_batch_size: 4
- eval_batch_size: 4
- seed: 42
- gradient_accumulation_steps: 64
- total_train_batch_size: 256
- optimizer: Use OptimizerNames.ADAMW_BNB with betas=(0.9,0.95) and epsilon=1e-08 and optimizer_args=No additional optimizer arguments
- lr_scheduler_type: cosine_with_min_lr
- lr_scheduler_warmup_steps: 1000
- num_epochs: 1.0
Training results
| Training Loss | Epoch | Step | Validation Loss |
|---|---|---|---|
| 1.3365 | 0.0438 | 500 | 1.3520 |
| 1.3108 | 0.0877 | 1000 | 1.3424 |
| 1.2777 | 0.1315 | 1500 | 1.3183 |
| 1.3173 | 0.1754 | 2000 | 1.2995 |
| 1.2953 | 0.2192 | 2500 | 1.2839 |
| 1.2651 | 0.2631 | 3000 | 1.2729 |
| 1.2416 | 0.3069 | 3500 | 1.2610 |
| 1.2501 | 0.3508 | 4000 | 1.2496 |
| 1.2258 | 0.3946 | 4500 | 1.2393 |
| 1.1961 | 0.4385 | 5000 | 1.2292 |
| 1.2401 | 0.4823 | 5500 | 1.2193 |
| 1.2089 | 0.5262 | 6000 | 1.2098 |
| 1.1854 | 0.5700 | 6500 | 1.2019 |
| 1.1716 | 0.6138 | 7000 | 1.1943 |
| 1.2056 | 0.6577 | 7500 | 1.1877 |
| 1.1998 | 0.7015 | 8000 | 1.1821 |
| 1.1582 | 0.7454 | 8500 | 1.1777 |
| 1.1667 | 0.7892 | 9000 | 1.1744 |
| 1.1042 | 0.8331 | 9500 | 1.1722 |
| 1.1436 | 0.8769 | 10000 | 1.1705 |
| 1.1224 | 0.9208 | 10500 | 1.1695 |
| 1.1215 | 0.9646 | 11000 | 1.1688 |
Framework versions
- Transformers 4.48.0.dev0
- Pytorch 2.5.1+cu124
- Datasets 3.2.0
- Tokenizers 0.21.0