Method load_compress_model

fastchat/model/model_adapter.py:114–121

(self, model_path, device, torch_dtype, revision="main")

        return model, tokenizer

    def load_compress_model(self, model_path, device, torch_dtype, revision="main"):
        return load_compress_model(
            model_path,
            device,
            torch_dtype,
            use_fast=self.use_fast_tokenizer,
            revision=revision,
        )

    def get_default_conv_template(self, model_path: str) -> Conversation:
        return get_conv_template("one_shot")