qlora_llama3_1_405b¶
- torchtune.models.llama3_1.qlora_llama3_1_405b(lora_attn_modules: List[Literal['q_proj', 'k_proj', 'v_proj', 'output_proj']], apply_lora_to_mlp: bool = False, apply_lora_to_output: bool = False, lora_rank: int = 8, lora_alpha: float = 16, lora_dropout: float = 0.0, *, quantize_base: bool = True) TransformerDecoder¶
用于创建启用了QLoRA的Llama3.1 405B模型的构建器。根据QLoRA论文,LoRA应用于线性层的基础模型权重被量化:https://arxiv.org/abs/2305.14314。请参阅lora_llama3_1_405b以获取完整的API参数。