Functions — github.com/antgroup/echomimic

hub / github.com/antgroup/echomimic / functions

Functions352 in github.com/antgroup/echomimic

⨍ Functions 352◇ Types & classes 83

Methodfrom_pretrained_2d

( cls, pretrained_model_path: PathLike, motion_module_path: PathLike, subfolde

src/models/unet_3d_echo.py:591

Functiongenerate_video

(uploaded_img, uploaded_audio, width, height, length, seed, facemask_dilation_ratio, facecrop_dilation_ratio,

webgui.py:231

Functionget_down_block

( down_block_type, num_layers, in_channels, out_channels, temb_channels, add_downsampl

src/models/unet_3d_blocks.py:15

Functionget_fps

(video_path)

src/utils/util.py:123

Methodget_sliced_feature_sparse

Get sliced features based on a given index :param feature_array: :param start_idx: the start index of the feature :p

src/models/whisper/audio2feature.py:48

Functionget_total_steps

( scheduler, timesteps: List[int], num_steps: Optional[int] = None, num_frames: int = ...,

src/pipelines/context.py:52

Functionget_up_block

( up_block_type, num_layers, in_channels, out_channels, prev_output_channel, temb_chan

src/models/unet_3d_blocks.py:94

Methodhacked_basic_transformer_inner_forward

( self, hidden_states: torch.FloatTensor, attention_mask: Optional[torch.F

src/models/mutual_self_attention.py:93

Functionimport_filename

src/models/whisper/whisper/model.py:280

Methodis_multilingual

(self)

src/models/whisper/whisper/model.py:253

Methodlanguage_token

Returns the token id corresponding to the value of the `language` field

src/models/whisper/whisper/tokenizer.py:196

Functionlinear

(v1, v2, t)

src/pipelines/utils.py:15

Methodlogits

(self, tokens: Tensor, audio_features: Tensor, include_embeddings=False)

src/models/whisper/whisper/decoding.py:139

Methodlogits

(self, tokens: torch.Tensor, audio_features: torch.Tensor)

src/models/whisper/whisper/model.py:242

Methodno_speech

(self)

src/models/whisper/whisper/tokenizer.py:181

Methodno_timestamps

(self)

src/models/whisper/whisper/tokenizer.py:186

Methodnon_speech_tokens

Returns the list of tokens to suppress in order to avoid any speaker tags or non-speech annotations, to prevent sampling texts that a

src/models/whisper/whisper/tokenizer.py:237

Functionold_motion_sync_old

(sequence_driver_det, reference_det)

src/utils/motion_utils.py:88

Functionoptional_float

(string)

src/models/whisper/whisper/utils.py:22

Functionoptional_int

(string)

src/models/whisper/whisper/utils.py:18

Methodpackets_callback

(output_packets: Mapping[str, packet_module.Packet])

src/utils/face_landmark.py:3116

Methodprepare_latents_bp

( self, batch_size, num_channels_latents, width, height, video

src/pipelines/pipeline_echo_mimic_pose.py:144

Methodprepare_latents_bp

( self, batch_size, num_channels_latents, width, height, video

src/pipelines/pipeline_echo_mimic_pose_acc.py:145

Methodrank

(self, tokens: List[List[Tensor]], sum_logprobs: List[List[float]])

src/models/whisper/whisper/decoding.py:182

Functionread_frames

(video_path)

src/utils/util.py:106

Methodrearrange_kv_cache

(self, source_indices)

src/models/whisper/whisper/decoding.py:158

Functionremove_symbols

Replace any other markers, symbols, punctuations with a space, keeping diacritics

src/models/whisper/whisper/normalizers/basic.py:46

Methodreset

(self)

src/models/whisper/whisper/decoding.py:292

Methodsave_to_cache

(module, _, output)

src/models/whisper/whisper/model.py:273

Functionsave_videos_from_pils

(pils, path, fps=24)

src/utils/img_utils.py:35

Methodscores

(logprobs, lengths)

src/models/whisper/whisper/decoding.py:183

Functionseed_everything

(seed)

src/utils/util.py:16

Methodset_attn_processor

r""" Sets the attention processor to use to compute attention. Parameters: processor (`dict` of `AttentionProcessor` or o

src/models/unet_3d.py:359

Methodset_attn_processor

r""" Sets the attention processor to use to compute attention. Parameters: processor (`dict` of `AttentionProcessor` or o

src/models/unet_3d_echo.py:367

Methodset_chunk_feed_forward

(self, chunk_size: Optional[int], dim: int = 0)

src/models/attention.py:145

Methodset_default_attn_processor

Disables custom attention processors and sets the default attention implementation.

src/models/unet_2d_condition.py:741

Functionset_tensor_interpolation_method

(is_slerp)

src/pipelines/utils.py:10

Methodset_use_memory_efficient_attention_xformers

( self, use_memory_efficient_attention_xformers: bool, attention_op: Optional[Callable

src/models/motion_module.py:309

Functionslerp

( v0: torch.Tensor, v1: torch.Tensor, t: float, DOT_THRESHOLD: float = 0.9995 )

src/pipelines/utils.py:19

Methodsmooth_f_axis

对5维tensor的F轴进行平滑，首尾帧保持原始值，其他帧应用平滑操作。参数: tensor -- 输入的5D Tensor，形状为(B, C, F, W, H) smoothing_coef -- 平滑前后帧的系数，默认为0.3

src/pipelines/pipeline_echo_mimic.py:553

Methodsmooth_f_axis

src/pipelines/pipeline_echo_mimic_acc.py:560

Methodsot

(self)

src/models/whisper/whisper/tokenizer.py:166

Methodsot_lm

(self)

src/models/whisper/whisper/tokenizer.py:171

Methodsot_prev

(self)

src/models/whisper/whisper/tokenizer.py:176

Methodsot_sequence_including_notimestamps

(self)

src/models/whisper/whisper/tokenizer.py:232

Functionstr2bool

(string)

src/models/whisper/whisper/utils.py:10

Methodtimestamp_begin

(self)

src/models/whisper/whisper/tokenizer.py:191

Methodto_fraction

(s: str)

src/models/whisper/whisper/normalizers/english.py:165

Functionuniform

( step: int = ..., num_steps: Optional[int] = None, num_frames: int = ..., context_size: Optio

src/pipelines/context.py:15

Methodupdate

(self, tokens: Tensor, logits: Tensor, sum_logprobs: Tensor)

src/models/whisper/whisper/decoding.py:258

Methodupdate

(self, tokens: Tensor, logits: Tensor, sum_logprobs: Tensor)

src/models/whisper/whisper/decoding.py:295

← previous301–352 of 352, ranked by callers