Hello author. I noticed that the tokenizer_config.json file for your llama 3.2 base model provided on Hugging Face doesn't have a chat template. I'd like to understand how you fine-tune the llama 3.2 base model—is it based on llama-factory? If I want to add a corresponding chat template to the llama 3.2 base model in tokenizer_config.json, what should I add to ensure consistency with the template you're using during the fine-tuning process? @yifei-he
Hello author. I noticed that the tokenizer_config.json file for your llama 3.2 base model provided on Hugging Face doesn't have a chat template. I'd like to understand how you fine-tune the llama 3.2 base model—is it based on llama-factory? If I want to add a corresponding chat template to the llama 3.2 base model in tokenizer_config.json, what should I add to ensure consistency with the template you're using during the fine-tuning process? @yifei-he