Functions — github.com/antgroup/echomimic

↓ 1 callersFunctionmain

()

infer_preview.py:162

↓ 1 callersMethodmultistep_dpm_solver_second_order_update

One step for the second-order multistep DPMSolver. Args: model_output_list (`List[torch.Tensor]`): The di

src/fm_solvers.py:486

↓ 1 callersMethodmultistep_dpm_solver_third_order_update

One step for the third-order multistep DPMSolver. Args: model_output_list (`List[torch.Tensor]`): The dir

src/fm_solvers.py:596

↓ 1 callersMethodmultistep_uni_c_bh_update

One step for the UniC (B(h) version). Args: this_model_output (`torch.Tensor`): The model outputs at `x_

src/fm_solvers_unipc.py:486

↓ 1 callersMethodmultistep_uni_p_bh_update

One step for the UniP (B(h) version). Alternatively, `self.solver_p` is used if is specified. Args: model_output (`torch

src/fm_solvers_unipc.py:350

↓ 1 callersFunctionparse_args

()

infer_flash.py:41

↓ 1 callersFunctionpos_interpolate

(pos, seq_len)

src/wan_image_encoder.py:24

↓ 1 callersMethodprepare_extra_step_kwargs

(self, generator, eta)

src/pipeline_wan_fun_inpaint_audio.py:388

↓ 1 callersMethodprepare_extra_step_kwargs

(self, generator, eta)

src/pipeline_wan_fun_inpaint_audio_2512.py:386

↓ 1 callersMethodprepare_latents

( self, batch_size, num_channels_latents, num_frames, height, width, dtype, device, generator, latents

src/pipeline_wan_fun_inpaint_audio.py:318

↓ 1 callersMethodprepare_latents

( self, batch_size, num_channels_latents, num_frames, height, width, dtype, device, generator, latents

src/pipeline_wan_fun_inpaint_audio_2512.py:322

↓ 1 callersMethodprepare_mask_latents

( self, mask, masked_image, batch_size, height, width, dtype, device, generator, do_classifier_free_gu

src/pipeline_wan_fun_inpaint_audio.py:345

↓ 1 callersMethodprepare_mask_latents

( self, mask, masked_image, batch_size, height, width, dtype, device, generator, do_classifier_free_gu

src/pipeline_wan_fun_inpaint_audio_2512.py:349

↓ 1 callersMethodreparameterize

(self, mu, log_var)

src/wan_vae.py:580

↓ 1 callersFunctionresize_mask

(mask, latent, process_first_frame_only=True)

src/pipeline_wan_fun_inpaint_audio.py:104

↓ 1 callersFunctionresize_mask

(mask, latent, process_first_frame_only=True)

src/pipeline_wan_fun_inpaint_audio_2512.py:107

↓ 1 callersFunctionsinusoidal_embedding_1d

(dim, position)

src/wan_transformer3d_audio_2512.py:481

↓ 1 callersFunctionsinusoidal_embedding_1d

(dim, position)

src/wan_transformer3d_audio.py:421

↓ 1 callersFunctionsplit_audio_sequence

Map the audio feature sequence to corresponding latent frame slices. Args: audio_proj_length (int): The total length of

src/wan_transformer3d_audio.py:214

↓ 1 callersFunctionsplit_tensor_with_padding

Split the input tensor into subsequences based on index ranges, and apply right-side zero-padding if the range exceeds the input boundaries.

src/wan_transformer3d_audio.py:255

↓ 1 callersMethodstep

Predict the sample from the previous timestep by reversing the SDE. This function propagates the sample with the multistep UniPC.

src/fm_solvers_unipc.py:655

↓ 1 callersMethodtime_shift

(self, mu: float, sigma: float, t: torch.Tensor)

src/fm_solvers.py:337

↓ 1 callersMethodtime_shift

(self, mu: float, sigma: float, t: torch.Tensor)

src/fm_solvers_unipc.py:276

↓ 1 callersMethodunpatchify

r""" Reconstruct video tensors from patch embeddings. Args: x (List[Tensor]): List of patchified features

src/wan_transformer3d_audio_2512.py:1454

↓ 1 callersMethodunpatchify

r""" Reconstruct video tensors from patch embeddings. Args: x (List[Tensor]): List of patchified features

src/wan_transformer3d_audio.py:1435

Method__call__

Function invoked when calling the pipeline for generation. Args: Examples: Returns:

src/pipeline_wan_fun_inpaint_audio.py:479

Method__call__

Function invoked when calling the pipeline for generation. Args: Examples: Returns:

src/pipeline_wan_fun_inpaint_audio_2512.py:469

Method__init__

(self)

app_mm.py:69

Method__init__

(self)

infer_preview.py:45

Method__init__

(self)

app.py:56

Method__init__

( self, seq_len=5, seq_len_vf=12, blocks=12, channels=768,

src/wan_transformer3d_audio_2512.py:177

Method__init__

(self, dim, eps=1e-5)

src/wan_transformer3d_audio_2512.py:615

Method__init__

(self, dim, eps=1e-6, elementwise_affine=False)

src/wan_transformer3d_audio_2512.py:634

Method__init__

(self, dim, num_heads, window_size=(-1, -1),

src/wan_transformer3d_audio_2512.py:647

Method__init__

(self, dim, num_heads, window_size=(-1, -1),

src/wan_transformer3d_audio_2512.py:735

Method__init__

(self, dim, num_heads, window_size=(-1, -1),

src/wan_transformer3d_audio_2512.py:791

Method__init__

(self, cross_attn_type, dim, ffn_dim, num_

src/wan_transformer3d_audio_2512.py:888

Method__init__

(self, in_dim, out_dim)

src/wan_transformer3d_audio_2512.py:1000

Method__init__

r""" Initialize the diffusion model backbone. Args: model_type (`str`, *optional*, defaults to 't2v'): Mo

src/wan_transformer3d_audio_2512.py:1022

Method__init__

( self, tokenizer: AutoTokenizer, text_encoder: WanT5EncoderModel, vae: Autoen

src/pipeline_wan_fun_inpaint_audio.py:172

Method__init__

(self, dim)

src/wan_camera_adapter.py:43

Method__init__

(self, dim, num_heads, dropout=0.1, eps=1e-5)

src/wan_xlm_roberta.py:12

Method__init__

(self, dim, num_heads, post_norm, dropout=0.1, eps=1e-5)

src/wan_xlm_roberta.py:51

Method__init__

(self, audio_in_dim=1024, cross_attention_dim=1024)

src/wan_transformer3d_audio.py:154

Method__init__

(self, dim, eps=1e-5)

src/wan_transformer3d_audio.py:555

Method__init__

(self, dim, eps=1e-6, elementwise_affine=False)

src/wan_transformer3d_audio.py:574

Method__init__

(self, dim, num_heads, window_size=(-1, -1),

src/wan_transformer3d_audio.py:587

Method__init__

(self, dim, num_heads, window_size=(-1, -1),

src/wan_transformer3d_audio.py:676

Method__init__

(self, dim, num_heads, window_size=(-1, -1),

src/wan_transformer3d_audio.py:733

Method__init__

(self, cross_attn_type, dim, ffn_dim, num_

src/wan_transformer3d_audio.py:838

Method__init__

(self, in_dim, out_dim)

src/wan_transformer3d_audio.py:950

Method__init__

r""" Initialize the diffusion model backbone. Args: model_type (`str`, *optional*, defaults to 't2v'): Mo

src/wan_transformer3d_audio.py:974

Method__init__

(self, dim, dim_attn, num_heads, dropout=0.1)

src/wan_text_encoder.py:60

Method__init__

(self, dim, dim_ffn, dropout=0.1)

src/wan_text_encoder.py:114

Method__init__

(self, dim, dim_attn, dim_ffn, num_heads,

src/wan_text_encoder.py:134

Method__init__

(self, dim, dim_attn, dim_ffn, num_heads,

src/wan_text_encoder.py:167

Method__init__

(self, num_buckets, num_heads, bidirectional, max_dist=128)

src/wan_text_encoder.py:209

Method__init__

(self, vocab, dim, dim_attn, dim_ffn,

src/wan_text_encoder.py:257

Method__init__

(self, config: Wav2Vec2Config)

src/wav2vec2.py:29

Method__init__

( self, num_train_timesteps: int = 1000, solver_order: int = 2, prediction_typ

src/fm_solvers.py:129

Method__init__

( self, tokenizer: AutoTokenizer, text_encoder: WanT5EncoderModel, vae: Autoen

src/pipeline_wan_fun_inpaint_audio_2512.py:176

Method__init__

(self, *args, **kwargs)

src/wan_vae.py:28

Method__init__

(self, dim, mode)

src/wan_vae.py:74

Method__init__

(self, in_dim, out_dim, dropout=0.0)

src/wan_vae.py:194

Method__init__

(self, dim)

src/wan_vae.py:234

Method__init__

(self, dim=128, z_dim=4, dim_mult=[1, 2, 4, 4],

src/wan_vae.py:273

Method__init__

(self, dim=128, z_dim=4, dim_mult=[1, 2, 4, 4],

src/wan_vae.py:377

Method__init__

(self, dim=128, z_dim=4, dim_mult=[1, 2, 4, 4],

src/wan_vae.py:491

Method__init__

( self, latent_channels=16, temporal_compression_ratio=4, spacial_compression_

src/wan_vae.py:626

Method__init__

(self, dim, num_heads, causal=False, attn_

src/wan_image_encoder.py:57

Method__init__

(self, dim, mlp_ratio, num_heads, post_nor

src/wan_image_encoder.py:116

Method__init__

(self, dim, mlp_ratio, num_heads, activati

src/wan_image_encoder.py:160

Method__init__

(self, image_size=224, patch_size=16, dim=768,

src/wan_image_encoder.py:213

Method__init__

(self, **kwargs)

src/wan_image_encoder.py:307

Method__init__

(self, embed_dim=1024, image_size=224, patch_size=14,

src/wan_image_encoder.py:332

Method__init__

(self)

src/wan_image_encoder.py:505

Method__init__

( self, num_train_timesteps: int = 1000, solver_order: int = 2,

src/fm_solvers_unipc.py:77

Method__init__

( self, coefficients: list[float], num_steps: int, rel_l1_thresh: float = 0.0,

src/cache_utils.py:27

Method__len__

(self)

src/fm_solvers.py:856

Method__len__

(self)

src/fm_solvers_unipc.py:799

Method_set_gradient_checkpointing

(self, module, value=False)

src/wan_transformer3d_audio_2512.py:1215

Method_set_gradient_checkpointing

(self, module, value=False)

src/wan_transformer3d_audio.py:1162

Method_sigma_to_t

(self, sigma)

src/fm_solvers.py:330

Method_sigma_to_t

(self, sigma)

src/fm_solvers_unipc.py:269

Methodadd_noise

( self, original_samples: torch.Tensor, noise: torch.Tensor, timesteps: torch.

src/fm_solvers.py:815

Methodadd_noise

( self, original_samples: torch.Tensor, noise: torch.Tensor, timesteps: torch.

src/fm_solvers_unipc.py:758

Methodattention_kwargs

(self)

src/pipeline_wan_fun_inpaint_audio.py:470

Methodattention_kwargs

(self)

src/pipeline_wan_fun_inpaint_audio_2512.py:460

Functionaudio_attention

( q, k, v, q_lens=None, k_lens=None, dropout_p=0., softmax_scale=None, q_scale

src/wan_transformer3d_audio_2512.py:384

Functionaudio_attention

( q, k, v, q_lens=None, k_lens=None, dropout_p=0., softmax_scale=None, q_scale

src/wan_transformer3d_audio.py:309

Methodaudio_guidance_scale

(self)

src/pipeline_wan_fun_inpaint_audio.py:462

Methodbegin_index

The index for the first timestep. It should be set from pipeline with `set_begin_index` method.

src/fm_solvers.py:209

Methodbegin_index

The index for the first timestep. It should be set from pipeline with `set_begin_index` method.

src/fm_solvers_unipc.py:142

Methodcreate_custom_forward

(module)

src/wan_transformer3d_audio_2512.py:1360

Methodcreate_custom_forward

(module)

src/wan_transformer3d_audio.py:1338

Methodcross_attn_ffn

(x, context, context_lens, e)

src/wan_transformer3d_audio_2512.py:954

Methodcross_attn_ffn

(x, context, context_lens, e)

src/wan_transformer3d_audio.py:904

Methodcustom_forward

(*inputs)

src/wan_transformer3d_audio_2512.py:1361

Methodcustom_forward

(*inputs)

src/wan_transformer3d_audio.py:1339

Methoddisable_riflex

(self)

src/wan_transformer3d_audio_2512.py:1197

Functions301 in github.com/antgroup/echomimic_v3