diffsynth/configs/model_configs.py

# Model loading hashes for the Wan/EverAnimate public release.

MODEL_CONFIGS = [{'extra_kwargs': {'dim': 5120,
                   'eps': 1e-06,
                   'ffn_dim': 13824,
                   'freq_dim': 256,
                   'has_image_input': False,
                   'in_dim': 16,
                   'num_heads': 40,
                   'num_layers': 40,
                   'out_dim': 16,
                   'patch_size': [1, 2, 2],
                   'text_dim': 4096},
  'model_class': 'diffsynth.models.wan_video_dit.WanModel',
  'model_hash': '5ec04e02b42d2580483ad69f4e76346a',
  'model_name': 'wan_video_dit',
  'state_dict_converter': 'diffsynth.utils.state_dict_converters.wan_video_dit.WanVideoDiTStateDictConverter'},
 {'model_class': 'diffsynth.models.wan_video_text_encoder.WanTextEncoder',
  'model_hash': '9c8818c2cbea55eca56c7b447df170da',
  'model_name': 'wan_video_text_encoder'},
 {'model_class': 'diffsynth.models.wan_video_vae.WanVideoVAE',
  'model_hash': 'ccc42284ea13e1ad04693284c7a09be6',
  'model_name': 'wan_video_vae',
  'state_dict_converter': 'diffsynth.utils.state_dict_converters.wan_video_vae.WanVideoVAEStateDictConverter'},
 {'model_class': 'diffsynth.models.longcat_video_dit.LongCatVideoTransformer3DModel',
  'model_hash': '8b27900f680d7251ce44e2dc8ae1ffef',
  'model_name': 'wan_video_dit'},
 {'extra_kwargs': {'dim': 5120,
                   'eps': 1e-06,
                   'ffn_dim': 13824,
                   'freq_dim': 256,
                   'has_image_input': True,
                   'in_dim': 36,
                   'num_heads': 40,
                   'num_layers': 40,
                   'out_dim': 16,
                   'patch_size': [1, 2, 2],
                   'text_dim': 4096},
  'model_class': 'diffsynth.models.wan_video_dit.WanModel',
  'model_hash': '5f90e66a0672219f12d9a626c8c21f61',
  'model_name': 'wan_video_dit',
  'state_dict_converter': 'diffsynth.utils.state_dict_converters.wan_video_dit.WanVideoDiTFromDiffusers'},
 {'model_class': 'diffsynth.models.wan_video_mot.MotWanModel',
  'model_hash': '5f90e66a0672219f12d9a626c8c21f61',
  'model_name': 'wan_video_vap',
  'state_dict_converter': 'diffsynth.utils.state_dict_converters.wan_video_mot.WanVideoMotStateDictConverter'},
 {'model_class': 'diffsynth.models.wan_video_image_encoder.WanImageEncoder',
  'model_hash': '5941c53e207d62f20f9025686193c40b',
  'model_name': 'wan_video_image_encoder',
  'state_dict_converter': 'diffsynth.utils.state_dict_converters.wan_video_image_encoder.WanImageEncoderStateDictConverter'},
 {'model_class': 'diffsynth.models.wan_video_motion_controller.WanMotionControllerModel',
  'model_hash': 'dbd5ec76bbf977983f972c151d545389',
  'model_name': 'wan_video_motion_controller'},
 {'extra_kwargs': {'dim': 1536,
                   'eps': 1e-06,
                   'ffn_dim': 8960,
                   'freq_dim': 256,
                   'has_image_input': False,
                   'in_dim': 16,
                   'num_heads': 12,
                   'num_layers': 30,
                   'out_dim': 16,
                   'patch_size': [1, 2, 2],
                   'text_dim': 4096},
  'model_class': 'diffsynth.models.wan_video_dit.WanModel',
  'model_hash': '9269f8db9040a9d860eaca435be61814',
  'model_name': 'wan_video_dit'},
 {'extra_kwargs': {'dim': 5120,
                   'eps': 1e-06,
                   'ffn_dim': 13824,
                   'freq_dim': 256,
                   'has_image_input': True,
                   'has_image_pos_emb': True,
                   'in_dim': 36,
                   'num_heads': 40,
                   'num_layers': 40,
                   'out_dim': 16,
                   'patch_size': [1, 2, 2],
                   'text_dim': 4096},
  'model_class': 'diffsynth.models.wan_video_dit.WanModel',
  'model_hash': '3ef3b1f8e1dab83d5b71fd7b617f859f',
  'model_name': 'wan_video_dit'},
 {'extra_kwargs': {'dim': 1536,
                   'eps': 1e-06,
                   'ffn_dim': 8960,
                   'freq_dim': 256,
                   'has_image_input': True,
                   'in_dim': 48,
                   'num_heads': 12,
                   'num_layers': 30,
                   'out_dim': 16,
                   'patch_size': [1, 2, 2],
                   'text_dim': 4096},
  'model_class': 'diffsynth.models.wan_video_dit.WanModel',
  'model_hash': '349723183fc063b2bfc10bb2835cf677',
  'model_name': 'wan_video_dit'},
 {'extra_kwargs': {'dim': 1536,
                   'eps': 1e-06,
                   'ffn_dim': 8960,
                   'freq_dim': 256,
                   'has_image_input': True,
                   'in_dim': 36,
                   'num_heads': 12,
                   'num_layers': 30,
                   'out_dim': 16,
                   'patch_size': [1, 2, 2],
                   'text_dim': 4096},
  'model_class': 'diffsynth.models.wan_video_dit.WanModel',
  'model_hash': '6d6ccde6845b95ad9114ab993d917893',
  'model_name': 'wan_video_dit'},
 {'extra_kwargs': {'dim': 5120,
                   'eps': 1e-06,
                   'ffn_dim': 13824,
                   'freq_dim': 256,
                   'has_image_input': True,
                   'in_dim': 48,
                   'num_heads': 40,
                   'num_layers': 40,
                   'out_dim': 16,
                   'patch_size': [1, 2, 2],
                   'text_dim': 4096},
  'model_class': 'diffsynth.models.wan_video_dit.WanModel',
  'model_hash': 'efa44cddf936c70abd0ea28b6cbe946c',
  'model_name': 'wan_video_dit'},
 {'extra_kwargs': {'dim': 5120,
                   'eps': 1e-06,
                   'ffn_dim': 13824,
                   'freq_dim': 256,
                   'has_image_input': True,
                   'in_dim': 36,
                   'num_heads': 40,
                   'num_layers': 40,
                   'out_dim': 16,
                   'patch_size': [1, 2, 2],
                   'text_dim': 4096},
  'model_class': 'diffsynth.models.wan_video_dit.WanModel',
  'model_hash': '6bfcfb3b342cb286ce886889d519a77e',
  'model_name': 'wan_video_dit'},
 {'extra_kwargs': {'add_control_adapter': True,
                   'dim': 1536,
                   'eps': 1e-06,
                   'ffn_dim': 8960,
                   'freq_dim': 256,
                   'has_image_input': True,
                   'has_ref_conv': False,
                   'in_dim': 32,
                   'in_dim_control_adapter': 24,
                   'num_heads': 12,
                   'num_layers': 30,
                   'out_dim': 16,
                   'patch_size': [1, 2, 2],
                   'text_dim': 4096},
  'model_class': 'diffsynth.models.wan_video_dit.WanModel',
  'model_hash': 'ac6a5aa74f4a0aab6f64eb9a72f19901',
  'model_name': 'wan_video_dit'},
 {'extra_kwargs': {'dim': 1536,
                   'eps': 1e-06,
                   'ffn_dim': 8960,
                   'freq_dim': 256,
                   'has_image_input': True,
                   'has_ref_conv': True,
                   'in_dim': 48,
                   'num_heads': 12,
                   'num_layers': 30,
                   'out_dim': 16,
                   'patch_size': [1, 2, 2],
                   'text_dim': 4096},
  'model_class': 'diffsynth.models.wan_video_dit.WanModel',
  'model_hash': '70ddad9d3a133785da5ea371aae09504',
  'model_name': 'wan_video_dit'},
 {'extra_kwargs': {'add_control_adapter': True,
                   'dim': 5120,
                   'eps': 1e-06,
                   'ffn_dim': 13824,
                   'freq_dim': 256,
                   'has_image_input': True,
                   'has_ref_conv': False,
                   'in_dim': 32,
                   'in_dim_control_adapter': 24,
                   'num_heads': 40,
                   'num_layers': 40,
                   'out_dim': 16,
                   'patch_size': [1, 2, 2],
                   'text_dim': 4096},
  'model_class': 'diffsynth.models.wan_video_dit.WanModel',
  'model_hash': 'b61c605c2adbd23124d152ed28e049ae',
  'model_name': 'wan_video_dit'},
 {'extra_kwargs': {'dim': 5120,
                   'eps': 1e-06,
                   'ffn_dim': 13824,
                   'freq_dim': 256,
                   'has_image_input': True,
                   'has_ref_conv': True,
                   'in_dim': 48,
                   'num_heads': 40,
                   'num_layers': 40,
                   'out_dim': 16,
                   'patch_size': [1, 2, 2],
                   'text_dim': 4096},
  'model_class': 'diffsynth.models.wan_video_dit.WanModel',
  'model_hash': '26bde73488a92e64cc20b0a7485b9e5b',
  'model_name': 'wan_video_dit'},
 {'extra_kwargs': {'dim': 5120,
                   'eps': 1e-06,
                   'ffn_dim': 13824,
                   'freq_dim': 256,
                   'has_image_input': False,
                   'in_dim': 16,
                   'num_heads': 40,
                   'num_layers': 40,
                   'out_dim': 16,
                   'patch_size': [1, 2, 2],
                   'text_dim': 4096},
  'model_class': 'diffsynth.models.wan_video_dit.WanModel',
  'model_hash': 'aafcfd9672c3a2456dc46e1cb6e52c70',
  'model_name': 'wan_video_dit'},
 {'extra_kwargs': {'dim': 1536,
                   'eps': 1e-06,
                   'ffn_dim': 8960,
                   'freq_dim': 256,
                   'has_image_input': False,
                   'in_dim': 16,
                   'num_heads': 12,
                   'num_layers': 30,
                   'out_dim': 16,
                   'patch_size': [1, 2, 2],
                   'text_dim': 4096},
  'model_class': 'diffsynth.models.wan_video_dit.WanModel',
  'model_hash': 'a61453409b67cd3246cf0c3bebad47ba',
  'model_name': 'wan_video_dit',
  'state_dict_converter': 'diffsynth.utils.state_dict_converters.wan_video_dit.WanVideoDiTStateDictConverter'},
 {'model_class': 'diffsynth.models.wan_video_vace.VaceWanModel',
  'model_hash': 'a61453409b67cd3246cf0c3bebad47ba',
  'model_name': 'wan_video_vace',
  'state_dict_converter': 'diffsynth.utils.state_dict_converters.wan_video_vace.VaceWanModelDictConverter'},
 {'extra_kwargs': {'dim': 5120,
                   'eps': 1e-06,
                   'ffn_dim': 13824,
                   'freq_dim': 256,
                   'has_image_input': False,
                   'in_dim': 16,
                   'num_heads': 40,
                   'num_layers': 40,
                   'out_dim': 16,
                   'patch_size': [1, 2, 2],
                   'text_dim': 4096},
  'model_class': 'diffsynth.models.wan_video_dit.WanModel',
  'model_hash': '7a513e1f257a861512b1afd387a8ecd9',
  'model_name': 'wan_video_dit',
  'state_dict_converter': 'diffsynth.utils.state_dict_converters.wan_video_dit.WanVideoDiTStateDictConverter'},
 {'extra_kwargs': {'dim': 5120,
                   'eps': 1e-06,
                   'ffn_dim': 13824,
                   'has_image_input': False,
                   'num_heads': 40,
                   'patch_size': (1, 2, 2),
                   'vace_in_dim': 96,
                   'vace_layers': (0, 5, 10, 15, 20, 25, 30, 35)},
  'model_class': 'diffsynth.models.wan_video_vace.VaceWanModel',
  'model_hash': '7a513e1f257a861512b1afd387a8ecd9',
  'model_name': 'wan_video_vace',
  'state_dict_converter': 'diffsynth.utils.state_dict_converters.wan_video_vace.VaceWanModelDictConverter'},
 {'extra_kwargs': {'dim': 5120,
                   'eps': 1e-06,
                   'ffn_dim': 13824,
                   'freq_dim': 256,
                   'has_image_input': True,
                   'in_dim': 36,
                   'num_heads': 40,
                   'num_layers': 40,
                   'out_dim': 16,
                   'patch_size': [1, 2, 2],
                   'text_dim': 4096},
  'model_class': 'diffsynth.models.wan_video_dit.WanModel',
  'model_hash': '31fa352acb8a1b1d33cd8764273d80a2',
  'model_name': 'wan_video_dit',
  'state_dict_converter': 'diffsynth.utils.state_dict_converters.wan_video_dit.WanVideoDiTStateDictConverter'},
 {'model_class': 'diffsynth.models.wan_video_animate_adapter.WanAnimateAdapter',
  'model_hash': '31fa352acb8a1b1d33cd8764273d80a2',
  'model_name': 'wan_video_animate_adapter',
  'state_dict_converter': 'diffsynth.utils.state_dict_converters.wan_video_animate_adapter.WanAnimateAdapterStateDictConverter'},
 {'extra_kwargs': {'add_control_adapter': True,
                   'dim': 5120,
                   'eps': 1e-06,
                   'ffn_dim': 13824,
                   'freq_dim': 256,
                   'has_image_input': False,
                   'has_ref_conv': False,
                   'in_dim': 36,
                   'in_dim_control_adapter': 24,
                   'num_heads': 40,
                   'num_layers': 40,
                   'out_dim': 16,
                   'patch_size': [1, 2, 2],
                   'require_clip_embedding': False,
                   'text_dim': 4096},
  'model_class': 'diffsynth.models.wan_video_dit.WanModel',
  'model_hash': '47dbeab5e560db3180adf51dc0232fb1',
  'model_name': 'wan_video_dit'},
 {'extra_kwargs': {'dim': 5120,
                   'eps': 1e-06,
                   'ffn_dim': 13824,
                   'freq_dim': 256,
                   'has_image_input': False,
                   'has_ref_conv': True,
                   'in_dim': 52,
                   'num_heads': 40,
                   'num_layers': 40,
                   'out_dim': 16,
                   'patch_size': [1, 2, 2],
                   'require_clip_embedding': False,
                   'text_dim': 4096},
  'model_class': 'diffsynth.models.wan_video_dit.WanModel',
  'model_hash': '2267d489f0ceb9f21836532952852ee5',
  'model_name': 'wan_video_dit'},
 {'extra_kwargs': {'dim': 5120,
                   'eps': 1e-06,
                   'ffn_dim': 13824,
                   'freq_dim': 256,
                   'has_image_input': False,
                   'in_dim': 36,
                   'num_heads': 40,
                   'num_layers': 40,
                   'out_dim': 16,
                   'patch_size': [1, 2, 2],
                   'require_clip_embedding': False,
                   'text_dim': 4096},
  'model_class': 'diffsynth.models.wan_video_dit.WanModel',
  'model_hash': '5b013604280dd715f8457c6ed6d6a626',
  'model_name': 'wan_video_dit'},
 {'extra_kwargs': {'audio_dim': 1024,
                   'cond_dim': 16,
                   'dim': 5120,
                   'eps': 1e-06,
                   'ffn_dim': 13824,
                   'freq_dim': 256,
                   'in_dim': 16,
                   'num_audio_token': 4,
                   'num_heads': 40,
                   'num_layers': 40,
                   'out_dim': 16,
                   'patch_size': (1, 2, 2),
                   'text_dim': 4096},
  'model_class': 'diffsynth.models.wan_video_dit_s2v.WanS2VModel',
  'model_hash': '966cffdcc52f9c46c391768b27637614',
  'model_name': 'wan_video_dit'},
 {'extra_kwargs': {'dim': 3072,
                   'eps': 1e-06,
                   'ffn_dim': 14336,
                   'freq_dim': 256,
                   'fuse_vae_embedding_in_latents': True,
                   'has_image_input': False,
                   'in_dim': 48,
                   'num_heads': 24,
                   'num_layers': 30,
                   'out_dim': 48,
                   'patch_size': [1, 2, 2],
                   'require_clip_embedding': False,
                   'require_vae_embedding': False,
                   'seperated_timestep': True,
                   'text_dim': 4096},
  'model_class': 'diffsynth.models.wan_video_dit.WanModel',
  'model_hash': '1f5ab7703c6fc803fdded85ff040c316',
  'model_name': 'wan_video_dit'},
 {'model_class': 'diffsynth.models.wan_video_vae.WanVideoVAE38',
  'model_hash': 'e1de6c02cdac79f8b739f4d3698cd216',
  'model_name': 'wan_video_vae',
  'state_dict_converter': 'diffsynth.utils.state_dict_converters.wan_video_vae.WanVideoVAEStateDictConverter'},
 {'model_class': 'diffsynth.models.wav2vec.WanS2VAudioEncoder',
  'model_hash': '06be60f3a4526586d8431cd038a71486',
  'model_name': 'wans2v_audio_encoder',
  'state_dict_converter': 'diffsynth.utils.state_dict_converters.wans2v_audio_encoder.WanS2VAudioEncoderStateDictConverter'}]