Finetune with `LLaMA-Factory`

`dataset_info.json` add

```json
{
    "llm-complex-reasoning-train-qwen2-72b-instruct-correct": {
        "file_name": "/data/songhaoyang/llm-complex-reasoning/data/llm-complex-reasoning-train-qwen2-72b-instruct-correct/train.jsonl",
        "formatting": "sharegpt",
        "columns": {
            "messages": "messages"
        },
        "tags": {
            "role_tag": "role",
            "content_tag": "content",
            "user_tag": "user",
            "assistant_tag": "assistant",
            "system_tag": "system"
        }
    },
    "Infinity-Instruct-0625": {
        "hf_hub_url": "BAAI/Infinity-Instruct",
        "subset": "0625",
        "formatting": "sharegpt"
    }
}
```

Usage: `llamafactory-cli train qwen2_7b_instruct_lora_sft.yaml`