VLMEvalKit-sudoku/llava/model/multimodal_encoder/builder.py

import os
from .imagebind import ImageBindWrapper
from .open_clip_encoder import OpenCLIPVisionTower
from .hf_vision import HFVisionTower
from .siglip_encoder import SigLipVisionTower
from .modeling_siglip2 import SigLip2VisionTower
from .modeling_swin_siglip2 import NaFlexSigLip2SwinVisionTower
from .modeling_swin_siglip2_zyc import SigLip2SwinVisionTower
from .clip_encoder import CLIPVisionTower, CLIPVisionTowerS2
from .modeling_moonvit import MoonViTVisionTower
from .modeling_qwen2_5vl import Qwen2_5VLVisionTower

# from .eva_clip.eva_clip_encoder import EvaClipVisionTower
# from .dev_eva_clip.eva_vit import EvaViTWrapper


def build_vision_tower(vision_tower_cfg, **kwargs):
    vision_tower = getattr(vision_tower_cfg, "mm_vision_tower", getattr(vision_tower_cfg, "vision_tower", None))
    is_absolute_path_exists = os.path.exists(vision_tower)
    use_s2 = getattr(vision_tower_cfg, "s2", False)

    if "siglip2" in vision_tower and "swin" in vision_tower:
        return SigLip2SwinVisionTower(vision_tower, vision_tower_cfg=vision_tower_cfg, **kwargs)
        # return NaFlexSigLip2SwinVisionTower(vision_tower, vision_tower_cfg=vision_tower_cfg, **kwargs)
    elif "siglip2" in vision_tower:
        return SigLip2VisionTower(vision_tower, vision_tower_cfg=vision_tower_cfg, **kwargs)
    elif "moonvit" in vision_tower:
        return MoonViTVisionTower(vision_tower, vision_tower_cfg=vision_tower_cfg, **kwargs)
    elif "qwen2_5vl" in vision_tower:
        return Qwen2_5VLVisionTower(vision_tower, vision_tower_cfg=vision_tower_cfg, **kwargs)
    elif "siglip" in vision_tower:
        return SigLipVisionTower(vision_tower, vision_tower_cfg=vision_tower_cfg, **kwargs)
    elif is_absolute_path_exists or vision_tower.startswith("openai") or vision_tower.startswith("laion") or "ShareGPT4V" in vision_tower:
        if use_s2:
            return CLIPVisionTowerS2(vision_tower, args=vision_tower_cfg, **kwargs)
        else:
            return CLIPVisionTower(vision_tower, args=vision_tower_cfg, **kwargs)
    elif vision_tower.startswith("hf:"):
        return HFVisionTower(vision_tower, args=vision_tower_cfg, **kwargs)
    elif vision_tower in ["imagebind_huge"]:
        return ImageBindWrapper(vision_tower, args=vision_tower_cfg, **kwargs)
    elif vision_tower.startswith("open_clip_hub"):
        return OpenCLIPVisionTower(vision_tower, args=vision_tower_cfg, **kwargs)
    # elif "internal-eva" in vision_tower.lower() or "eva02" in vision_tower.lower():
    #     return EvaClipVisionTower(vision_tower, args=vision_tower_cfg, **kwargs)
    # elif vision_tower in ["EVA-CLIP-8B", "EVA-CLIP-8B-plus"]:
    #     return EvaViTWrapper(vision_tower, args=vision_tower_cfg, **kwargs)

    raise ValueError(f"Unknown vision tower: {vision_tower}")