MCPcopy Create free account

hub / github.com/FunAudioLLM/ThinkSound / functions

Functions638 in github.com/FunAudioLLM/ThinkSound

⨍ Functions 638◇ Types & classes 175
Method__init__
(self, n_q: int, empty_initial: int = 0)
ThinkSound/models/codebook_patterns.py:365
Method__init__
(self, n_q: int, flattening: tp.Optional[tp.List[int]] = None, delays: tp.Optional[tp.List[in
ThinkSound/models/codebook_patterns.py:420
Method__init__
(self, n_q: int, delays: tp.Optional[tp.List[int]] = None)
ThinkSound/models/codebook_patterns.py:504
Method__init__
(self, model, scale=1.0, model_half=False, iterate_batch=False, chunked=False)
ThinkSound/models/pretransforms.py:29
Method__init__
(self, channels, levels, wavelet)
ThinkSound/models/pretransforms.py:93
Method__init__
(self, attenuation=100, num_bands=16)
ThinkSound/models/pretransforms.py:112
Method__init__
(self, model_type="44khz", model_bitrate="8kbps", scale=1.0, quantize_on_decode: bool = True, chunked=True)
ThinkSound/models/pretransforms.py:134
Method__init__
(self, model_type="facebook/encodec_32khz", scale=1.0, quantize_on_decode: bool = True)
ThinkSound/models/pretransforms.py:195
Method__init__
(self, dim: int, nheads: int)
ThinkSound/models/transformer_layers.py:49
Method__init__
(self, dim: int, nheads: int)
ThinkSound/models/transformer_layers.py:91
Method__init__
(self, dim: int, nhead: int, mlp_ratio: float = 4.0,
ThinkSound/models/transformer_layers.py:137
Method__init__
(self, dim, out_dim)
ThinkSound/models/transformer_layers.py:261
Method__init__
(self, dim, max_seq_len)
ThinkSound/models/transformer.py:45
Method__init__
(self, dim, theta = 10000)
ThinkSound/models/transformer.py:66
Method__init__
( self, dim, use_xpos = False, scale_base = 512, interpolation_factor
ThinkSound/models/transformer.py:90
Method__init__
bias-less layernorm has been shown to be more stable. most newer models have moved towards rmsnorm, also bias-less
ThinkSound/models/transformer.py:174
Method__init__
( self, dim, dim_out = None, mult = 4, no_bias = False, glu =
ThinkSound/models/transformer.py:222
Method__init__
( self, dim, dim_heads = 64, dim_context = None, causal = False,
ThinkSound/models/transformer.py:272
Method__init__
( self, dim, norm_kwargs = {}, )
ThinkSound/models/transformer.py:555
Method__init__
( self, dim, dim_heads = 64, cross_attend = False,
ThinkSound/models/transformer.py:591
Method__init__
( self, dim, depth, *, dim_in = None, dim_out = None,
ThinkSound/models/transformer.py:703
Method__init__
( self, dim: int, output_dim: int, project_out: bool = False
ThinkSound/models/conditioners.py:21
Method__init__
(self, output_dim: int, hiera_ckpt_path, project_out: boo
ThinkSound/models/conditioners.py:38
Method__init__
(self, dim, output_dim)
ThinkSound/models/conditioners.py:103
Method__init__
(self, dim, output_dim, global_dim=1536)
ThinkSound/models/conditioners.py:130
Method__init__
(self, dim, output_dim)
ThinkSound/models/conditioners.py:187
Method__init__
(self, dim, output_dim)
ThinkSound/models/conditioners.py:215
Method__init__
( self, dim: int, output_dim: int, clip_model_name: str = "met
ThinkSound/models/conditioners.py:247
Method__init__
(self, output_dim: int, min_val: int=0, max_val: int=512
ThinkSound/models/conditioners.py:299
Method__init__
(self, output_dim: int, min_val: float=0, max_val: float=1
ThinkSound/models/conditioners.py:325
Method__init__
(self, output_dim: int, clap_ckpt_path, use_text_features
ThinkSound/models/conditioners.py:357
Method__init__
(self, output_dim: int, clap_ckpt_path, audio_model_type=
ThinkSound/models/conditioners.py:440
Method__init__
( self, output_dim: int, t5_model_name: str = "t5-base", max_l
ThinkSound/models/conditioners.py:524
Method__init__
( self, output_dim: int, max_length: str = 77, enable_grad: bo
ThinkSound/models/conditioners.py:604
Method__init__
( self, output_dim: int, max_length: str = 77, enable_grad: bo
ThinkSound/models/conditioners.py:695
Method__init__
( self, output_dim: int, max_length: int = 1024, project_out:
ThinkSound/models/conditioners.py:756
Method__init__
( self, tokenizer_name: str, # Name of a tokenizer from the Hugging Face transformers
ThinkSound/models/conditioners.py:812
Method__init__
(self, pretransform: Pretransform, output_dim: int)
ThinkSound/models/conditioners.py:867
Method__init__
(self, conditioners: tp.Dict[str, Conditioner], default_keys: tp.Dict[str, str] = {})
ThinkSound/models/conditioners.py:897
Method__init__
(self, dim, frequency_embedding_size, max_period)
ThinkSound/models/embeddings.py:48
Method__init__
(self, num_quantizers, codebook_size, tokens_id)
ThinkSound/models/bottleneck.py:23
Method__init__
(self)
ThinkSound/models/bottleneck.py:34
Method__init__
(self)
ThinkSound/models/bottleneck.py:62
Method__init__
(self, noise_augment_dim: int = 0, bypass_mmd: bool = False)
ThinkSound/models/bottleneck.py:98
Method__init__
(self)
ThinkSound/models/bottleneck.py:130
Method__init__
(self, **quantizer_kwargs)
ThinkSound/models/bottleneck.py:147
Method__init__
(self, **quantizer_kwargs)
ThinkSound/models/bottleneck.py:176
Method__init__
(self, quantize_on_decode=False, noise_augment_dim=0, **quantizer_kwargs)
ThinkSound/models/bottleneck.py:209
Method__init__
(self, quantize_on_decode=False, **quantizer_kwargs)
ThinkSound/models/bottleneck.py:262
Method__init__
(self, noise_augment_dim=0, **kwargs)
ThinkSound/models/bottleneck.py:314
Method__init__
(self, io_channels=32, patch_size=1, embed_dim=768, cond_token_dim=0,
ThinkSound/models/dit.py:14
Method__init__
( self, root: Union[str, Path], *, tsv_path: Union[str, Path] = 'dataset/vggso
data_utils/v2a_utils/vggsound_text.py:25
Method__init__
( self, root: Union[str, Path], *, tsv_path: Union[str, Path] = 'dataset/vggso
data_utils/v2a_utils/audioset_224.py:76
Method__init__
( self, root: Union[str, Path], *, tsv_path: Union[str, Path] = 'dataset/vggso
data_utils/v2a_utils/audioset_video_224.py:76
Method__init__
( self, *, vae_ckpt: Optional[str] = None, vae_config: Optional[str] = None,
data_utils/v2a_utils/feature_utils_224.py:54
Method__init__
( self, root: Union[str, Path], *, tsv_path: Union[str, Path] = 'dataset/vggso
data_utils/v2a_utils/vggsound.py:25
Method__init__
( self, root: Union[str, Path], *, tsv_path: Union[str, Path] = 'dataset/vggso
data_utils/v2a_utils/vggsound_224_no_sync.py:76
Method__init__
( self, *, vae_ckpt: Optional[str] = None, vae_config: Optional[str] = None,
data_utils/v2a_utils/feature_utils_224_audio.py:56
Method__init__
( self, root: Union[str, Path], *, tsv_path: Union[str, Path] = 'dataset/vggso
data_utils/v2a_utils/vggsound_224_no_audio.py:76
Method__init__
( self, root: Union[str, Path], *, tsv_path: Union[str, Path] = 'dataset/vggso
data_utils/v2a_utils/audio_text_dataset.py:29
Method__init__
( self, root: Union[str, Path], *, tsv_path: Union[str, Path] = 'dataset/vggso
data_utils/v2a_utils/vggsound_224.py:76
Method__init__
(self, cfg)
data_utils/ext/synchformer/video_model_builder.py:19
Method__init__
(self, dim, num_heads=8, qkv_bias=False, attn_drop=0., proj_drop=0.)
data_utils/ext/synchformer/vit_helper.py:38
Method__init__
(self, dim=768, num_heads=12, attn_type='divided',
data_utils/ext/synchformer/vit_helper.py:121
Method__init__
(self, img_size=224, patch_size=16, in_chans=3, embed_dim=768)
data_utils/ext/synchformer/vit_helper.py:218
Method__init__
(self, img_size=224, temporal_resolution=4, in_chans=3,
data_utils/ext/synchformer/vit_helper.py:238
Method__init__
(self, n_input, n_classes, n_hidden=512, p=0.1)
data_utils/ext/synchformer/vit_helper.py:270
Method__init__
(self)
data_utils/ext/synchformer/synchformer.py:12
Method__init__
(self, add_pos_emb: bool = False, pos_emb_drop: float = None,
data_utils/ext/synchformer/motionformer.py:269
Method__init__
(self, *args, **kwargs)
data_utils/ext/synchformer/motionformer.py:343
Method__init__
(self, *args, **kwargs)
data_utils/ext/synchformer/motionformer.py:372
Method__init__
patterns are e.g. "bs t d -> bs d"
data_utils/ext/synchformer/motionformer.py:388
Method__len__
(self)
ThinkSound/data/dataset.py:207
Method__len__
(self)
ThinkSound/data/dataset.py:334
Method__len__
(self)
ThinkSound/data/dataset.py:425
Method__len__
(self)
ThinkSound/data/dataset.py:538
Method__len__
(self)
ThinkSound/data/dataset.py:580
Method__len__
(self)
data_utils/v2a_utils/vggsound_text.py:95
Method__len__
(self)
data_utils/v2a_utils/audioset_224.py:301
Method__len__
(self)
data_utils/v2a_utils/audioset_video_224.py:254
Method__len__
(self)
data_utils/v2a_utils/vggsound.py:245
Method__len__
(self)
data_utils/v2a_utils/vggsound_224_no_sync.py:209
Method__len__
(self)
data_utils/v2a_utils/vggsound_224_no_audio.py:261
Method__len__
(self)
data_utils/v2a_utils/audio_text_dataset.py:159
Method__len__
(self)
data_utils/v2a_utils/vggsound_224.py:311
Method__post_init__
(self)
ThinkSound/models/codebook_patterns.py:47
Method__repr__
(self)
ThinkSound/training/diffusion.py:36
Method__repr__
(self)
ThinkSound/models/diffusion.py:26
Method_basic_init
(module)
ThinkSound/models/mmdit.py:225
Function_conv_filter
convert patch embedding weight from manual patchify + linear proj to conv
data_utils/ext/synchformer/vit_helper.py:292
Method_init_weights
(self, m)
data_utils/ext/synchformer/video_model_builder.py:148
Method_init_weights
(self, m)
data_utils/ext/synchformer/motionformer.py:326
Method_num_inner_steps
Number of inner steps to unroll between timesteps in order to flatten the codebooks.
ThinkSound/models/codebook_patterns.py:454
Methodadd_prefix
(s)
ThinkSound/data/dataset.py:175
Methodadd_prefix
(s)
ThinkSound/data/dataset.py:302
Methodadd_prefix
(s)
ThinkSound/data/dataset.py:390
Methodadd_prefix
(s)
ThinkSound/data/dataset.py:486
Functionalpha_sigma_to_t
Returns a timestep, given the scaling factors for the clean image and for the noise.
ThinkSound/inference/sampling.py:13
Functionaudio_decoder
(key, value)
ThinkSound/data/dataset.py:957
Methodbuild_pattern_sequence
Build sequence corresponding to the pattern from the input tensor z. The sequence is built using up to sequence_steps if specified, and non-pa
ThinkSound/models/codebook_patterns.py:151
← previousnext →301–400 of 638, ranked by callers