MCPcopy Create free account

hub / github.com/Lightricks/ComfyUI-LTXVideo / functions

Functions429 in github.com/Lightricks/ComfyUI-LTXVideo

⨍ Functions 429◇ Types & classes 117
↓ 74 callersMethodget
(self, key, default=None)
tricks/utils/attn_bank.py:15
↓ 13 callersMethodexecute
( cls, noise, guider, sampler, sigmas, latent_image, v
easy_samplers.py:941
↓ 12 callersMethodgenerate
( self, vae, positive, negative, latent, guiding_latent, latent_idx, strength )
latents.py:434
↓ 12 callersFunctionsyncWidgets
(node)
web/js/sparse_track_editor.js:568
↓ 9 callersMethodselect_latents
Selects a range of frames from the video latent. Args: samples (dict): Video latent dictionary start_index (
latents.py:49
↓ 9 callersMethodset_conds
(self, positive, negative)
stg.py:224
↓ 8 callersFunctionappend_guide_attention_entry
Append a new guide attention entry to conditioning metadata. Called by guide-adding nodes after appending tokens. Args: conditioning
iclora_attention.py:22
↓ 8 callersMethodexecute
(cls, positive, negative, audio_latent)
iclora.py:589
↓ 6 callersMethodsample
( self, model, vae, width, height, num_frames, guider,
easy_samplers.py:128
↓ 6 callersMethodunpack_latents
(self, x: torch.Tensor)
guiders/multimodal_guider.py:87
↓ 5 callersMethodapply
(self, model, power, only_first_frame)
dynamic_conditioning.py:23
↓ 5 callersMethodencode
( cls, vae, latent_width, latent_height, images, scale_factors
iclora.py:105
↓ 5 callersMethodupdate
(self, update_value: torch.Tensor)
stg.py:41
↓ 4 callersMethod_extract_latent_spatial_tile
Extract spatial tile from a latent dictionary.
looping_sampler.py:241
↓ 4 callersFunctionblur_internal
(image, blur_radius)
guide.py:9
↓ 4 callersMethoddilate_latent
( self, latent: dict, horizontal_scale: int, vertical_scale: int )
latents.py:357
↓ 4 callersFunctionimgToCanvas
(node, p)
web/js/sparse_track_editor.js:296
↓ 4 callersFunctionmouseToCanvas
(node, e)
web/js/sparse_track_editor.js:306
↓ 4 callersMethodpatch_model
(cls, model: ModelPatcher, stg_flag: STGFlag)
stg.py:208
↓ 3 callersMethod__init__
(self, aggregate_embed: nn.Module, is_av: bool = False)
text_embeddings_connectors.py:168
↓ 3 callersMethod__init__
(self, tokenizer_path: str, max_length: int = 1024)
gemma_encoder.py:52
↓ 3 callersFunction_cat_with_padding
Concatenate a tensor with a padding tensor of the given value.
gemma_encoder.py:499
↓ 3 callersFunction_get_raw_conds_from_guider
(guider)
easy_samplers.py:19
↓ 3 callersFunction_pad_for_laplacian
Pad an image tensor so H and W are powers of two (kornia requirement).
pyramid_blending.py:23
↓ 3 callersMethoddecode
( self, vae, latents, horizontal_tiles, vertical_tiles, overla
tiled_vae_decode.py:37
↓ 3 callersMethoddo_skip
(self, step: int)
guiders/parameters.py:67
↓ 3 callersFunctionfind_matching_dir
Recursively search for files matching a glob pattern and return the parent directory of the first match.
gemma_encoder.py:200
↓ 3 callersFunctiongetScale
(node)
web/js/sparse_track_editor.js:288
↓ 3 callersFunctionhitTestPoint
(node, canvasPos)
web/js/sparse_track_editor.js:318
↓ 3 callersMethodsample
( self, model, vae, noise, sampler, sigmas, guider,
looping_sampler.py:692
↓ 2 callersMethod__init__
( self, model: ModelPatcher, cfg, stg_scale, rescale_scale: float = None )
stg.py:190
↓ 2 callersFunction_enhance
Common enhancement logic for both T2V and I2V modes.
gemma_encoder.py:402
↓ 2 callersFunction_filter_sd
Extract keys with *prefix* and strip the prefix.
text_embeddings_connectors.py:45
↓ 2 callersFunction_generate_and_decode_prompts
( prompt_enhancer_model, prompt_enhancer_tokenizer, model_inputs, max_new_tokens: int )
prompt_enhancer_utils.py:239
↓ 2 callersFunction_load_aggregate_embed
Load an aggregate_embed Linear from the state dict. Args: sd: Full checkpoint state dict. modality: ``"video"`` or ``"audio"``.
text_embeddings_connectors.py:50
↓ 2 callersFunction_load_system_prompt
Load system prompt from file at module level.
gemma_encoder.py:28
↓ 2 callersFunction_locate_model_within_model
(super_model, model_name)
gemma_encoder.py:550
↓ 2 callersMethod_parse_comma_separated_string
Parse a comma-separated configuration string into a list with type conversion.
looping_sampler.py:580
↓ 2 callersFunction_rescale_norm
Rescale normalization: x * sqrt(target_dim / source_dim).
text_embeddings_connectors.py:40
↓ 2 callersFunction_to_binary_mask
Convert additive mask to binary [B, seq, 1] and zero out padding.
text_embeddings_connectors.py:33
↓ 2 callersFunctionapg
( noise_pred_pos: torch.Tensor, noise_pred_neg: torch.Tensor, cfg_scale: float, eta: float = 1
stg.py:55
↓ 2 callersMethodbatch_normalize
(self, latents, reference, factor, per_frame=False)
latent_norm.py:42
↓ 2 callersMethodcalculate
( self, noise_pred_pos, noise_pred_neg, noise_pred_pertubed, noise_pred_modality )
guiders/parameters.py:41
↓ 2 callersFunctioncamel_case_to_spaces
(text: str)
nodes_registry.py:149
↓ 2 callersFunctioncanvasToImg
(node, p)
web/js/sparse_track_editor.js:301
↓ 2 callersFunctiondist
(a, b)
web/js/sparse_track_editor.js:73
↓ 2 callersMethoddo_cross_attn
(self, step: int)
guiders/parameters.py:73
↓ 2 callersFunctiongenerate_trend_values
(steps, start_time, end_time, eta, eta_trend)
tricks/nodes/rectified_sampler_nodes.py:6
↓ 2 callersMethodget_model_size
(model)
prompt_enhancer_nodes.py:55
↓ 2 callersMethodget_subbatch
(self, samples)
easy_samplers.py:866
↓ 2 callersFunctionhasConnectedInput
(node, widgetName)
web/js/guider_parameters_advanced.js:178
↓ 2 callersFunctionhideWidget
(w)
web/js/sparse_track_editor.js:77
↓ 2 callersFunctioninterpolateSpline
(controlPoints, numSamples)
web/js/sparse_track_editor.js:26
↓ 2 callersMethodload
(self, llm_name, image_captioner_name)
prompt_enhancer_nodes.py:141
↓ 2 callersFunctionload_embeddings_connector
( sd, connector_prefix, connector_config, dtype=torch.bfloat16, rope_type=LTXRopeType.INTE
embeddings_connector.py:349
↓ 2 callersFunctionltxv_gemma_tokenizer
(tokenizer_path, max_length=256)
gemma_encoder.py:163
↓ 2 callersMethodmodel_path_download_if_needed
(self, model_name)
prompt_enhancer_nodes.py:95
↓ 2 callersFunctionnormalize_mask
Normalize a ComfyUI MASK to (1, 1, F, H, W) for downstream processing. ComfyUI MASK type is typically (B, H, W) or (H, W).
iclora_attention.py:56
↓ 2 callersMethodpack_latents
(self, vx: torch.Tensor, ax: torch.Tensor)
guiders/multimodal_guider.py:97
↓ 2 callersMethodpatch_smooth_norm
( model: ModelPatcher, factors: str, normalize_fn: Callable )
latent_norm.py:224
↓ 2 callersMethodrequired
(s)
latent_norm.py:207
↓ 2 callersFunctionstg
( noise_pred_pos, noise_pred_neg, noise_pred_pertubed, cfg_scale, stg_scale, rescale_s
stg.py:16
↓ 2 callersFunctiontoggleModalityWidgets
(node, prefix, enabled)
web/js/guider_parameters_advanced.js:48
↓ 2 callersFunctiontoggleWidget
(node, widget, show)
web/js/guider_parameters_advanced.js:23
↓ 1 callersMethodINPUT_TYPES
(s)
low_vram_loaders.py:62
↓ 1 callersMethod__init__
( self, in_channels=128, cross_attention_dim=2048, attention_head_dim=128,
embeddings_connector.py:120
↓ 1 callersFunction_age_color_batch
Vectorised age-ratio -> RGB [0..1] mapping on GPU. Gradient: blue -> green -> yellow -> red.
sparse_tracks.py:164
↓ 1 callersFunction_apply_low_res_mask_dilation
( mask: torch.Tensor, spatial_radius: int, long_side: int = _MASK_LOW_RES_LONG_SIDE, )
pyramid_blending.py:71
↓ 1 callersMethod_apply_masks
Applies resized masks to all frames.
latents.py:321
↓ 1 callersMethod_calculate_keyframe_per_tile_indices
Calculate which temporal tile each keyframe falls into. Returns a list of tuples (temporal_tile_index, in_tile_index) for each keyfr
looping_sampler.py:620
↓ 1 callersMethod_calculate_tile_seed
Calculate the seed value for a specific temporal and spatial tile.
looping_sampler.py:557
↓ 1 callersFunction_catmull_rom
(p0: dict, p1: dict, p2: dict, p3: dict, t: float)
sparse_tracks.py:13
↓ 1 callersMethod_create_spatial_weights
Create blending weights for spatial tiles.
looping_sampler.py:525
↓ 1 callersMethod_extract_spatial_tile
Extract spatial tiles from all inputs for a given spatial region.
looping_sampler.py:254
↓ 1 callersFunction_feta_score
(query_image, key_image, head_dim, num_frames, enhance_weight)
tricks/utils/feta_enhance_utils.py:5
↓ 1 callersFunction_format_description
Format nodes display name to a standard format
nodes_registry.py:123
↓ 1 callersFunction_gaussian_pyramid
( images: torch.Tensor, max_level: int, border_type: str = "reflect", align_corners: bool = Fa
pyramid_blending.py:35
↓ 1 callersFunction_generate_i2v_prompt
( image_caption_model, image_caption_processor, prompt_enhancer_model, prompt_enhancer_tokeniz
prompt_enhancer_utils.py:165
↓ 1 callersFunction_generate_image_captions
( image_caption_model, image_caption_processor, images: List[Image.Image], system_prompt: str
prompt_enhancer_utils.py:202
↓ 1 callersFunction_generate_t2v_prompt
( prompt_enhancer_model, prompt_enhancer_tokenizer, prompts: List[str], max_new_tokens: int,
prompt_enhancer_utils.py:135
↓ 1 callersFunction_get_first_frames_from_conditioning_item
( conditioning_item: Tuple[torch.Tensor, int, float] )
prompt_enhancer_utils.py:124
↓ 1 callersFunction_get_guide_attention_entries
Read the current guide_attention_entries list from conditioning.
iclora_attention.py:4
↓ 1 callersMethod_get_per_tile_value
Get a value from a per-tile configuration list, using the last value if the list is shorter.
looping_sampler.py:576
↓ 1 callersFunction_get_random_scene_type
Randomly select a scene type to add to the prompt.
prompt_enhancer_utils.py:225
↓ 1 callersFunction_hdr_decompress
Decompress VAE-decoded image from [0,1] to linear HDR [0, inf). ComfyUI's VAE decode returns images in [0, 1] via ``(raw + 1) / 2``. This fun
hdr.py:63
↓ 1 callersFunction_interpolate_spline
Catmull-Rom spline interpolation matching the JS frontend logic.
sparse_tracks.py:34
↓ 1 callersFunction_is_v3_node
Check if the node class is a v3 node (has define_schema method).
nodes_registry.py:40
↓ 1 callersFunction_linear_to_srgb
Convert linear [0, 1] to sRGB [0, 1].
hdr.py:73
↓ 1 callersFunction_load_single_aggregate_embed
Load the single aggregate_embed (V1 models). Returns: nn.Linear or None if key not found.
text_embeddings_connectors.py:259
↓ 1 callersFunction_load_single_aggregate_embed_from_file
Load aggregate_embed from a standalone checkpoint file (legacy fallback). Args: path: Path to safetensors file containing aggregate_embed
text_embeddings_connectors.py:278
↓ 1 callersFunction_locate_unique_parameter_owner_by_leaf
( root: torch.nn.Module, leaf_param_name: str, must_have_in_path: Optional[str] = None, )
gemma_encoder.py:560
↓ 1 callersMethod_merge_noise_masks
Merges noise masks from both latents with proper handling.
latents.py:193
↓ 1 callersFunction_pad_inputs_for_attention_alignment
Pad sequence length to multiple of alignment for Flash Attention compatibility. Flash Attention within SDPA requires sequence lengths aligned to
gemma_encoder.py:519
↓ 1 callersFunction_parse_tracks
Parse tracks from a JSON string, handling nested/wrapped formats.
sparse_tracks.py:142
↓ 1 callersFunction_patchify_audio_latent
Reshape audio latent (b, c, t, f) → ref_audio token dict (b, t, c*f).
iclora.py:526
↓ 1 callersMethod_prepare_guider_for_chunk
Prepare the guider for a specific chunk, handling optional positive conditionings.
looping_sampler.py:597
↓ 1 callersMethod_process_temporal_chunks
Process all temporal chunks for a single spatial tile.
looping_sampler.py:310
↓ 1 callersFunction_pyramid_blend
( image1: torch.Tensor, image2: torch.Tensor, mask: torch.Tensor, max_level: int = 7, devi
pyramid_blending.py:117
↓ 1 callersFunction_pyramid_blend_chunk
Blend a single chunk (already padded, already on device).
pyramid_blending.py:97
↓ 1 callersFunction_rasterise_circles
Stamp filled circles onto *frame* fully on-device. Uses ``scatter_reduce_`` with ``'amax'`` to resolve overlaps in painter's order (circles a
sparse_tracks.py:330
next →1–100 of 429, ranked by callers