VTP.Config(
        encoder=Encoder.Config(
            patch_size=16,
            channels_hidden=1_024,
            num_layers=24,
            heads=16,
            expansion=4.0,
            channels_out=64,
            checkpoint=HubFile.Config(
                repo_id='MiniMaxAI/VTP-Large-f16d64',
                filename='model.safetensors',
                revision='d1726ca9abd5ed0b3ad938c508c635d8b1a51a31',
                sha256='de5df2006083a9536c4d3ea36c6ae2181ec604e06550f1b2d7cece3b16aac32f'
            )
        ),
        decoder=Decoder.Config(
            channels_in=64,
            channels_hidden=1_024,
            num_layers=24,
            heads=16,
            expansion=4.0,
            upscale_factor=16,
            checkpoint=HubFile.Config(
                repo_id='MiniMaxAI/VTP-Large-f16d64',
                filename='model.safetensors',
                revision='d1726ca9abd5ed0b3ad938c508c635d8b1a51a31',
                sha256='de5df2006083a9536c4d3ea36c6ae2181ec604e06550f1b2d7cece3b16aac32f'
            )
        ),
        image_size=256,
        pixel_mean=(0.485, 0.456, 0.406),
        pixel_std=(0.229, 0.224, 0.225),
        dtype_autocast=None,
        latent_norm=ChannelLatentStats.Config(
            stats=UrlFile.Config(
                url='https://raw.githubusercontent.com/MiniMax-AI/VTP/5ce1eb67010fff3c1eed483352483be6a1838556/generation/latent_stats/vtp_l/latents_stats.pt',
                sha256='0ad1fadbfb8959f3a93147a4db668b6d704ae36ff320bfad53cd3d568de7b0fd'
            ),
            multiplier=1.0
        )
    )
