mfai.pytorch.models.llms¶ class mfai.pytorch.models.llms.FreezeMLMMixin[source]¶ Bases: object A Mixin for (un)freezing llm and vision stages of a multimodal model. backend: GPT2 | Llama2 | CrossAttentionGPT2 | Llama3¶ freeze_llm()[source]¶ Freeze the LLM layers (not the vision layers). Return type: None freeze_vision()[source]¶ Freeze the vision encoder layers. Return type: None unfreeze_llm()[source]¶ Unfreeze the LLM layers. Return type: None unfreeze_vision()[source]¶ Unfreeze the vision encoder layers. Return type: None vision_encoder: Module¶