saiga2_7b.json

{
    "trainer": {
        "evaluation_strategy": "steps",
        "per_device_train_batch_size": 1,
        "per_device_eval_batch_size": 1,
        "gradient_accumulation_steps": 128,
        "eval_steps": 50,
        "save_steps": 50,
        "logging_steps": 2,
        "learning_rate": 0.0003,
        "num_train_epochs": 5,
        "lr_scheduler_type": "cosine",
        "warmup_steps": 50,
        "fp16": false,
        "bf16": true,
        "torch_compile": false,
        "optim": "adamw_torch"
    },
    "lora": {
        "r": 16,
        "lora_alpha": 16,
        "lora_dropout": 0.05,
        "bias": "none",
        "target_modules": ["q_proj", "v_proj", "k_proj", "o_proj"],
        "task_type": "CAUSAL_LM"
    },
    "load_in_8bit": true,
    "only_target_loss": true,
    "mode": "chat",
    "templates_path": "internal_prompts/saiga_v2.json",
    "model_name": "models/llama2-7b",
    "model_type": "causal",
    "max_tokens_count": 4096
}