sft_2025-02-14-2
This model is a fine-tuned version of /app/data/models/Llama-3.2-Taiwan-1B on the tw-reasoning, the skythought, the openthought and the dolphin-r1 datasets.
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-06
- train_batch_size: 3
- eval_batch_size: 1
- seed: 42
- distributed_type: multi-GPU
- num_devices: 48
- gradient_accumulation_steps: 10
- total_train_batch_size: 1440
- total_eval_batch_size: 48
- optimizer: Use adamw_torch_fused with betas=(0.9,0.999) and epsilon=1e-08 and optimizer_args=min_lr_ratio=0.1
- lr_scheduler_type: cosine
- lr_scheduler_warmup_ratio: 0.01
- num_epochs: 10.0
Training results
Framework versions
- Transformers 4.48.3
- Pytorch 2.5.1+cu124
- Datasets 3.2.0
- Tokenizers 0.21.0
- Downloads last month
- 6
Inference Providers
NEW
This model is not currently available via any of the supported Inference Providers.