| 12345678910111213141516171819 |
- model:
- base_model: /path/to/Baichuan2-7B-Chat
- data:
- train_file: data/processed/opinion_train.jsonl
- lora:
- r: 8
- alpha: 32
- dropout: 0.05
- target_modules: [q_proj, k_proj, v_proj]
- train:
- output_dir: checkpoints/lora-opinion
- epochs: 3
- batch_size: 1
- gradient_accumulation_steps: 8
- learning_rate: 0.0002
- max_length: 4096
|