Brunobkr/llama.cpp_AlgMor24_github

1from __future__ import annotations2 3from typing import Callable, TYPE_CHECKING4 5if TYPE_CHECKING:6    from torch import Tensor7 8from .base import ModelBase, gguf9 10from .llama import LlamaModel11from .qwenvl import Qwen2VLVisionModel12 13 14@ModelBase.register("Sarashina2VisionForCausalLM")15class Sarashina2VLTextModel(LlamaModel):16    model_arch = gguf.MODEL_ARCH.LLAMA17 18    @classmethod19    def filter_tensors(cls, item: tuple[str, Callable[[], Tensor]]) -> tuple[str, Callable[[], Tensor]] | None:20        name, gen = item21        if name.startswith("llm."):22            name = name.replace("llm.", "", 1)23        elif name.startswith("norm."):24            return None25        return super().filter_tensors((name, gen))26 27 28@ModelBase.register("Sarashina2VisionForCausalLM")29class Sarashina2VLVisionModel(Qwen2VLVisionModel):30    def __init__(self, *args, **kwargs):31        super().__init__(*args, **kwargs)32        self.global_config['model_type'] = "qwen2_vl"33