MCPcopy Create free account

hub / github.com/AlayaLab/AlayaRenderer / functions

Functions1,613 in github.com/AlayaLab/AlayaRenderer

⨍ Functions 1,613◇ Types & classes 459↳ Endpoints 1
↓ 4 callersMethodprepare_image_ids
(self, latents)
game_editing/diffsynth/models/flux_dit.py:306
↓ 4 callersMethodprocess
(self, pipe, **kwargs)
game_editing/diffsynth/diffusion/base_pipeline.py:54
↓ 4 callersMethodto_torch_shape
(self)
game_editing/diffsynth/models/ltx2_common.py:52
↓ 3 callersMethod__getitem__
(self, item)
game_editing/diffsynth/utils/data/__init__.py:122
↓ 3 callersMethod__init__
( self, offload_dtype: torch.dtype = None, offload_device: Union[str, torch.device] =
game_editing/diffsynth/core/vram/layers.py:10
↓ 3 callersMethod__init__
(self, disable_guidance_embedder=False, num_joint_blocks=5, num_single_blocks=10, num_mode=0, mode_dict={}, ad
game_editing/diffsynth/models/flux_controlnet.py:62
↓ 3 callersMethod__init__
(self, cross_attention_dim=768, id_embeddings_dim=512, num_tokens=4)
game_editing/diffsynth/models/flux_ipadapter.py:24
↓ 3 callersMethod__len__
(self)
game_editing/diffsynth/utils/data/__init__.py:109
↓ 3 callersMethod_calculate_shift_qwen_image
(image_seq_len, base_seq_len=256, max_seq_len=8192, base_shift=0.5, max_shift=0.9)
game_editing/diffsynth/diffusion/flow_match.py:44
↓ 3 callersMethod_pad_with_ids
Pad feature to SEQ_MULTI_OF, create position IDs and pad mask.
game_editing/diffsynth/models/z_image_dit.py:806
↓ 3 callersMethod_prepare_sequence
Prepare sequence: apply pad token, RoPE embed, pad to batch, create attention mask.
game_editing/diffsynth/models/z_image_dit.py:683
↓ 3 callersFunctionapply_gate
AI is creating summary for apply_gate Args: x (torch.Tensor): input tensor. gate (torch.Tensor, optional): gate tensor. Defaults
game_editing/diffsynth/models/step1x_connector.py:170
↓ 3 callersFunctionbasic_clean
(text)
game_editing/diffsynth/models/wan_video_text_encoder.py:260
↓ 3 callersMethodblend_h
(self, a: torch.Tensor, b: torch.Tensor, blend_extent: int)
game_editing/diffsynth/models/flux2_vae.py:2142
↓ 3 callersMethodblend_v
(self, a: torch.Tensor, b: torch.Tensor, blend_extent: int)
game_editing/diffsynth/models/flux2_vae.py:2136
↓ 3 callersMethoddecode
(self, z, scale)
game_editing/diffsynth/models/wan_video_vae.py:1011
↓ 3 callersFunctionenable_vram_management
(model: torch.nn.Module, module_map: dict, vram_config: dict, vram_limit=None, disk_map=None, **kwargs)
game_editing/diffsynth/core/vram/layers.py:468
↓ 3 callersMethodextract_masked_hidden
(self, hidden_states: torch.Tensor, mask: torch.Tensor)
game_editing/diffsynth/pipelines/qwen_image.py:331
↓ 3 callersMethodfetch_input_params
(self)
game_editing/diffsynth/diffusion/base_pipeline.py:33
↓ 3 callersMethodfetch_output_params
(self)
game_editing/diffsynth/diffusion/base_pipeline.py:47
↓ 3 callersMethodforward
r""" Encode video frames into normalized latent representation. Args: sample: Input video (B, C, F, H, W). F must be 1 + 8
game_editing/diffsynth/models/ltx2_video_vae.py:1426
↓ 3 callersFunctionget_activation_layer
get activation layer Args: act_type (str): the activation type Returns: torch.nn.functional: the activation layer
game_editing/diffsynth/models/step1x_connector.py:264
↓ 3 callersFunctionget_device_name
Get current device name based on device type.
game_editing/diffsynth/core/device/npu_compatible_device.py:47
↓ 3 callersMethodget_rope_index
Calculate the 3D rope index based on image and video's temporal, height and width in LLM. Explanation: Each embedding se
game_editing/diffsynth/models/nexus_gen_ar_model.py:110
↓ 3 callersMethodload_from_disk
(self, torch_dtype, device, copy_module=False)
game_editing/diffsynth/core/vram/layers.py:126
↓ 3 callersMethodload_from_disk
(self, torch_dtype, device, assign=True)
game_editing/diffsynth/core/vram/layers.py:359
↓ 3 callersFunctionmake_attn
( in_channels: int, attn_type: AttentionType = AttentionType.VANILLA, norm_type: NormType = NormTy
game_editing/diffsynth/models/ltx2_audio_vae.py:256
↓ 3 callersFunctionmodulate_fp32
(norm_func, x, shift, scale)
game_editing/diffsynth/models/longcat_video_dit.py:344
↓ 3 callersFunctionprecompute_freqs_cis
(dim: int, end: int = 1024, theta: float = 10000.0)
game_editing/diffsynth/models/wan_video_dit.py:83
↓ 3 callersFunctionprecompute_freqs_cis_3d
(dim: int, end: int = 1024, theta: float = 10000.0)
game_editing/diffsynth/models/wan_video_dit.py:75
↓ 3 callersMethodprocess_stage
(self, pipe: LTX2AudioVideoPipeline, height, width, num_frames, seed, rand_device, frame_rate=24.0)
game_editing/diffsynth/pipelines/ltx2_audio_video.py:423
↓ 3 callersFunctionreshape_tensor
(x, heads)
game_editing/diffsynth/models/flux_infiniteyou.py:17
↓ 3 callersFunctionrope_precompute
(x, grid_sizes, freqs, start=None)
game_editing/diffsynth/models/wan_video_dit_s2v.py:26
↓ 3 callersFunctionskip_model_initialization
(device=torch.device("meta"))
game_editing/diffsynth/core/vram/initialization.py:6
↓ 3 callersMethodto
(self, *args, **kwargs)
game_editing/diffsynth/diffusion/training_module.py:12
↓ 3 callersMethodun_normalize
(self, x: torch.Tensor)
game_editing/diffsynth/models/ltx2_audio_vae.py:768
↓ 3 callersMethodvae_output_to_video
(self, vae_output, pattern="B C T H W", min_value=-1, max_value=1)
game_editing/diffsynth/diffusion/base_pipeline.py:140
↓ 2 callersMethod__init__
( self, mot_layers=(0, 4, 8, 12, 16, 20, 24, 28, 32, 36), patch_size=(1, 2, 2),
game_editing/diffsynth/models/wan_video_mot.py:95
↓ 2 callersMethod__init__
(self, q_dim, num_heads, head_dim, kv_dim=None, bias_q=False, bias_kv=False, bias_out=False)
game_editing/diffsynth/models/sd_text_encoder.py:18
↓ 2 callersMethod__init__
(self, dim)
game_editing/diffsynth/models/flux_lora_patcher.py:251
↓ 2 callersMethod__init__
(self, entry)
game_editing/diffsynth/models/wan_video_camera_controller.py:80
↓ 2 callersMethod__init__
(self, q_dim, num_heads, head_dim, kv_dim=None, bias_q=False, bias_kv=False, bias_out=False)
game_editing/diffsynth/models/flux_text_encoder_clip.py:6
↓ 2 callersMethod_apply_2d
(self, x2d: torch.Tensor)
game_editing/diffsynth/models/ltx2_upsampler.py:135
↓ 2 callersMethod_decode
(self, z: torch.Tensor, return_dict: bool = True)
game_editing/diffsynth/models/flux2_vae.py:2088
↓ 2 callersMethod_encode
(self, x: torch.Tensor)
game_editing/diffsynth/models/flux2_vae.py:2045
↓ 2 callersMethod_expand_pos_freqs_if_needed
(self, video_fhw, txt_seq_lens)
game_editing/diffsynth/models/qwen_image_dit.py:94
↓ 2 callersMethod_feed_spatial_noise
( self, hidden_states: torch.Tensor, per_channel_scale: torch.Tensor, generato
game_editing/diffsynth/models/ltx2_video_vae.py:657
↓ 2 callersMethod_get_audio_latent_time_in_sec
Converts latent indices into real-time seconds while honoring causal offsets and the configured hop length. Args:
game_editing/diffsynth/models/ltx2_audio_vae.py:52
↓ 2 callersMethod_init_preprocessors
Initialize preprocessors for LTX.
game_editing/diffsynth/models/ltx2_dit.py:1225
↓ 2 callersMethod_prepare_positional_embeddings
Prepare positional embeddings.
game_editing/diffsynth/models/ltx2_dit.py:628
↓ 2 callersMethod_process_cross_attn
(self, x, cond, kv_seqlen)
game_editing/diffsynth/models/longcat_video_dit.py:285
↓ 2 callersMethod_process_output
Process output for LTXV.
game_editing/diffsynth/models/ltx2_dit.py:1380
↓ 2 callersMethodalign_res_stack_to_original_blocks
(self, res_stack, num_blocks, hidden_states)
game_editing/diffsynth/models/flux_controlnet.py:104
↓ 2 callersMethodall_in_batch
(self, perturbation_type: PerturbationType, block: int)
game_editing/diffsynth/models/ltx2_dit.py:219
↓ 2 callersMethodapply_input_images_to_latents
(self, latents, input_latents, input_indexes, input_strength, initial_latents=None, num_frames=121)
game_editing/diffsynth/pipelines/ltx2_audio_video.py:225
↓ 2 callersFunctionattention
(q, k, v, attn_mask, mode="torch")
game_editing/diffsynth/models/step1x_connector.py:12
↓ 2 callersMethodbatch_decode
This method forwards all its arguments to Qwen2TokenizerFast's [`~PreTrainedTokenizer.batch_decode`]. Please refer to the docstring o
game_editing/diffsynth/models/nexus_gen_ar_model.py:1083
↓ 2 callersMethodbuild_1d_mask
(self, length, left_bound, right_bound, border_width)
game_editing/diffsynth/models/wan_video_vae.py:1081
↓ 2 callersMethodbuild_mask
(self, data, is_bound, border_width)
game_editing/diffsynth/models/wan_video_vae.py:1090
↓ 2 callersFunctionbuild_mid_block
Build the middle block with two ResNet blocks and optional attention.
game_editing/diffsynth/models/ltx2_audio_vae.py:778
↓ 2 callersMethodcal_audio_emb
(self, audio_input, motion_frames=[73, 19])
game_editing/diffsynth/models/wan_video_dit_s2v.py:484
↓ 2 callersMethodcast_to
(self, weight, dtype, device)
game_editing/diffsynth/core/vram/layers.py:60
↓ 2 callersMethodcheck_free_vram
(self)
game_editing/diffsynth/core/vram/layers.py:65
↓ 2 callersMethodclear_lora
(self, verbose=1)
game_editing/diffsynth/diffusion/base_pipeline.py:273
↓ 2 callersFunctioncompute_trapezoidal_mask_1d
Generate a 1D trapezoidal blending mask with linear ramps. Args: length: Output length of the mask. ramp_left: Fade-in length
game_editing/diffsynth/models/ltx2_video_vae.py:944
↓ 2 callersMethodconvert_state_dict
(self, state_dict)
game_editing/diffsynth/utils/lora/flux.py:84
↓ 2 callersFunctionconvert_state_dict_keys_to_single_str
(state_dict, with_shape=True)
game_editing/diffsynth/core/loader/file.py:52
↓ 2 callersFunctionconvert_to_uint8
(frames: torch.Tensor)
game_editing/diffsynth/models/ltx2_video_vae.py:2219
↓ 2 callersFunctioncount_conv3d
(model)
game_editing/diffsynth/models/wan_video_vae.py:943
↓ 2 callersFunctioncreate_custom_forward
(module)
game_editing/diffsynth/utils/xfuser/xdit_context_parallel.py:91
↓ 2 callersFunctioncreate_custom_forward
(module)
game_editing/diffsynth/core/gradient/gradient_checkpoint.py:4
↓ 2 callersFunctioncreate_custom_forward_vap
(block, vap)
game_editing/diffsynth/pipelines/wan_video.py:1329
↓ 2 callersMethodedit_image_auto_resize
(self, edit_image)
game_editing/diffsynth/pipelines/qwen_image.py:547
↓ 2 callersMethodencode_images_using_qwenvl
(self, pipe: QwenImagePipeline, images: list[Image.Image], highres=False)
game_editing/diffsynth/pipelines/qwen_image.py:622
↓ 2 callersMethodencode_prompt
(self, pipe: WanVideoPipeline, prompt)
game_editing/diffsynth/pipelines/wan_video.py:693
↓ 2 callersMethodexport_trainable_state_dict
(self, state_dict, remove_prefix=None)
game_editing/diffsynth/diffusion/training_module.py:54
↓ 2 callersMethodflush_files
(self)
game_editing/diffsynth/core/vram/disk_map.py:46
↓ 2 callersMethodforward
imgs: [B, 3, H, W] of torch.float32. - mean: [0.48145466, 0.4578275, 0.40821073] - std: [0.26862954, 0.2613025
game_editing/diffsynth/models/wan_video_image_encoder.py:571
↓ 2 callersMethodforward
(self, x: torch.Tensor, first_chunk=False)
game_editing/diffsynth/models/wan_video_vae.py:417
↓ 2 callersMethodforward
r""" Decode latent representation into video frames. Args: sample: Latent tensor (B, 128, F', H', W'). timeste
game_editing/diffsynth/models/ltx2_video_vae.py:1896
↓ 2 callersMethodforward
Forward pass through the upsampling block. Args: x (torch.Tensor): Input tensor feat_cache (list, optional):
game_editing/diffsynth/models/qwen_image_vae.py:496
↓ 2 callersFunctionfp16_clamp
(x)
game_editing/diffsynth/models/wan_video_text_encoder.py:11
↓ 2 callersFunctionfused_leaky_relu
(input, bias, negative_slope=0.2, scale=2 ** 0.5)
game_editing/diffsynth/models/wan_video_animate_adapter.py:321
↓ 2 callersMethodget_av_ca_ada_values
( self, scale_shift_table: torch.Tensor, batch_size: int, scale_shift_timestep
game_editing/diffsynth/models/ltx2_dit.py:872
↓ 2 callersMethodget_grid_sizes
(self, grid_size_x, grid_size_ref)
game_editing/diffsynth/models/wan_video_dit_s2v.py:491
↓ 2 callersMethodget_i2v_mask
(self, lat_t, lat_h, lat_w, mask_len=1, mask_pixel_values=None, device=get_device_type())
game_editing/diffsynth/pipelines/wan_video.py:968
↓ 2 callersMethodget_image_latent
(self, pipe, input_image, height, width, tiled, tile_size_in_pixels, tile_overlap_in_pixels)
game_editing/diffsynth/pipelines/ltx2_audio_video.py:478
↓ 2 callersFunctionget_norm_layer
Get the normalization layer. Args: norm_layer (str): The type of normalization layer. Returns: norm_layer (nn.Module):
game_editing/diffsynth/models/step1x_connector.py:246
↓ 2 callersMethodget_num_image_tokens
(self, image_grid_thw)
game_editing/diffsynth/models/nexus_gen_ar_model.py:738
↓ 2 callersMethodget_patch_grid_bounds
( self, output_shape: AudioLatentShape | VideoLatentShape, device: torch.device | None
game_editing/diffsynth/models/ltx2_common.py:325
↓ 2 callersMethodget_position_ids
Generates position ids for the input embeddings grid. modified from the qwen2_vl mrope.
game_editing/diffsynth/models/nexus_gen_projector.py:327
↓ 2 callersMethodget_target_embeddings
(self, images, messages, processor, model, num_img_tokens=81)
game_editing/diffsynth/models/nexus_gen.py:116
↓ 2 callersMethodget_tensor
(self, name)
game_editing/diffsynth/core/vram/disk_map.py:21
↓ 2 callersMethodinject_motion
(self, x, rope_embs, mask_input, motion_latents, drop_motion_frames=True, add_last_motion=2)
game_editing/diffsynth/models/wan_video_dit_s2v.py:448
↓ 2 callersFunctioninteract_with_ipadapter
(hidden_states, q, ip_k, ip_v, scale=1.0)
game_editing/diffsynth/models/flux_dit.py:6
↓ 2 callersMethodload_lora
( self, module: torch.nn.Module, lora_config: Union[ModelConfig, str] = None,
game_editing/diffsynth/diffusion/base_pipeline.py:233
↓ 2 callersMethodload_lora
(self, lora_config, dtype, device)
game_editing/diffsynth/pipelines/flux_image.py:867
↓ 2 callersFunctionload_state_dict_from_bin
(file_path, torch_dtype=None, device="cpu")
game_editing/diffsynth/core/loader/file.py:36
↓ 2 callersFunctionmerge_lora
(loras: List[Dict[str, torch.Tensor]], alpha=1)
game_editing/diffsynth/utils/lora/merge.py:11
↓ 2 callersMethodoffload
(self)
game_editing/diffsynth/core/vram/layers.py:71
← previousnext →101–200 of 1,613, ranked by callers