mixvideo-v2/cargos/tvai-v2/视觉-语言模型配置/ViT-M-16-alt.json

{
    "embed_dim": 384,
    "vision_cfg": {
        "image_size": 224,
        "layers": 12,
        "width": 512,
        "patch_size": 16,
        "ls_init_value": 1e-4
    },
    "text_cfg": {
        "context_length": 77,
        "vocab_size": 49408,
        "width": 384,
        "heads": 6,
        "layers": 12
    }
}