egogpt/model/multimodal_encoder/builder.py

import os

from .clip_encoder import CLIPVisionTower, CLIPVisionTowerS2
from .siglip_encoder import SigLipVisionTower

# 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 is_absolute_path_exists or vision_tower.startswith("openai") or vision_tower.startswith("laion") or "ShareGPT4V" in vision_tower:
    if (
        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 (
        "siglip" in vision_tower.lower()
        or "open_clip_pytorch_model.bin" in vision_tower
    ):
        return SigLipVisionTower(
            vision_tower, vision_tower_cfg=vision_tower_cfg, **kwargs
        )

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