MCPcopy Create free account

hub / github.com/arielshaulov/FlowMo / functions

Functions271 in github.com/arielshaulov/FlowMo

⨍ Functions 271◇ Types & classes 63
↓ 10 callersMethod_sigma_to_alpha_sigma_t
(self, sigma)
wan/utils/fm_solvers.py:333
↓ 8 callersMethod__init__
(self, vocab_size, dim, dim_attn, dim_ffn,
wan/modules/t5.py:374
↓ 8 callersMethod__init__
(self, z_dim=16, vae_pth='cache/vae_step_411000.pth', dtype
wan/modules/vae.py:621
↓ 8 callersMethod_sigma_to_alpha_sigma_t
(self, sigma)
wan/utils/fm_solvers_unipc.py:272
↓ 8 callersFunctionget_optimal_device
Select the GPU with the most available memory to balance workload. Returns the device with most available memory.
wan/modules/model.py:18
↓ 8 callersMethodset_timesteps
Sets the discrete timesteps used for the diffusion chain (to be run before inference). Args: num_inference_steps (`int`):
wan/utils/fm_solvers.py:226
↓ 7 callersMethod__init__
(self, dim, out_dim, patch_size, eps=1e-6)
wan/modules/model.py:485
↓ 7 callersMethodencode
videos: A list of videos each with shape [C, T, H, W].
wan/modules/vae.py:647
↓ 7 callersFunctionflash_attention
q: [B, Lq, Nq, C1]. k: [B, Lk, Nk, C1]. v: [B, Lk, Nk, C2]. Nq must be divisible by Nk. q_lens
wan/modules/attention.py:24
↓ 7 callersMethodstep
Predict the sample from the previous timestep by reversing the SDE. This function propagates the sample with the multistep DPMSolver.
wan/utils/fm_solvers.py:706
↓ 6 callersMethod__init__
(self, dim, mid_dim)
wan/modules/clip.py:96
↓ 6 callersFunctionhalf
(x)
wan/modules/attention.py:59
↓ 5 callersMethoddecode
(self, zs)
wan/modules/vae.py:657
↓ 5 callersFunctionfp16_clamp
(x)
wan/modules/t5.py:20
↓ 5 callersFunctionget_sampling_sigmas
(sampling_steps, shift)
wan/utils/fm_solvers.py:22
↓ 5 callersFunctionretrieve_timesteps
( scheduler, num_inference_steps=None, device=None, timesteps=None, sigmas=None, **kwa
wan/utils/fm_solvers.py:29
↓ 4 callersFunctionceil_by_factor
Returns the smallest integer greater than or equal to 'number' that is divisible by 'factor'.
wan/utils/qwen_vl_utils.py:44
↓ 4 callersMethodclear_cache
(self)
wan/modules/vae.py:582
↓ 4 callersFunctionround_by_factor
Returns the closest integer to 'number' that is divisible by 'factor'.
wan/utils/qwen_vl_utils.py:39
↓ 4 callersFunctionsmart_resize
Rescales the image so that the following conditions are met: 1. Both dimensions (height and width) are divisible by 'factor'. 2. The to
wan/utils/qwen_vl_utils.py:54
↓ 3 callersFunctionbasic_clean
(text)
wan/modules/tokenizers.py:12
↓ 3 callersFunctionfloor_by_factor
Returns the largest integer less than or equal to 'number' that is divisible by 'factor'.
wan/utils/qwen_vl_utils.py:49
↓ 3 callersMethodgenerate
r""" Generates video frames from text prompt using diffusion process. Args: input_prompt (`str`): Text pr
wan/vace.py:261
↓ 3 callersFunctionhalf
(x)
wan/distributed/xdit_context_parallel.py:196
↓ 3 callersFunctionrope_params
(max_seq_len, dim, theta=10000)
wan/modules/model.py:62
↓ 3 callersFunctionsinusoidal_embedding_1d
(dim, position)
wan/modules/model.py:47
↓ 3 callersMethodtransfer_data_to_cuda
(self, data, device)
wan/vace.py:491
↓ 3 callersMethodunpatchify
Device-consistent unpatchify implementation
wan/modules/model.py:922
↓ 3 callersMethodvisual
(self, videos)
wan/modules/clip.py:527
↓ 2 callersMethod__init__
(self, vace_layers=None, vace_in_dim=None, model_type='vace
wan/modules/vace_model.py:66
↓ 2 callersMethod__init__
(self, vocab_size=250002, max_seq_len=514, type_size=1,
wan/modules/xlm_roberta.py:81
↓ 2 callersMethod__init__
(self, model_name, is_vl=False, device=0, **kwargs)
wan/utils/prompt_extend.py:167
↓ 2 callersMethod_threshold_sample
"Dynamic thresholding: At each sampling step we set s to a certain percentile absolute pixel value in xt0 (the prediction of x_0 at t
wan/utils/fm_solvers.py:292
↓ 2 callersMethod_threshold_sample
"Dynamic thresholding: At each sampling step we set s to a certain percentile absolute pixel value in xt0 (the prediction of x_0 at t
wan/utils/fm_solvers_unipc.py:230
↓ 2 callersMethodclear
Clear a specific saved state to free memory
motion_optimizer.py:353
↓ 2 callersFunctioncount_conv3d
(model)
wan/modules/vae.py:475
↓ 2 callersMethoddecode_latent
(self, zs, ref_images=None, vae=None)
wan/vace.py:244
↓ 2 callersMethodencode
(self, x, scale)
wan/modules/vae.py:516
↓ 2 callersMethodextend
(self, prompt, system_prompt, seed=-1, *args, **kwargs)
wan/utils/prompt_extend.py:181
↓ 2 callersFunctionfetch_image
(ele: dict[str, str | Image.Image], size_factor: int = IMAGE_FACTOR)
wan/utils/qwen_vl_utils.py:85
↓ 2 callersMethodforward
(self, x)
wan/modules/clip.py:106
↓ 2 callersMethodforward
(self, x)
wan/modules/vae.py:510
↓ 2 callersMethodforward
r""" Forward pass through the diffusion model Args: x (List[Tensor]): List of input video tensors, each w
wan/modules/vace_model.py:138
↓ 2 callersMethodforward_vace
( self, x, vace_context, seq_len, kwargs )
wan/modules/vace_model.py:113
↓ 2 callersMethodindex_for_timestep
(self, timestep, schedule_timesteps=None)
wan/utils/fm_solvers.py:679
↓ 2 callersMethodindex_for_timestep
(self, timestep, schedule_timesteps=None)
wan/utils/fm_solvers_unipc.py:630
↓ 2 callersMethodload_image_batch
(self, *data_key_batch, normalize=True, seq_len=None, **kwargs)
wan/utils/vace_processor.py:69
↓ 2 callersMethodload_video_batch
(self, *data_key_batch, crop_box=None, seed=2024, **kwargs)
wan/utils/vace_processor.py:226
↓ 2 callersFunctionqkv_fn
(x)
wan/distributed/xdit_context_parallel.py:200
↓ 2 callersFunctionrope_apply
Optimized version of rope_apply function with improved device management.
wan/modules/model.py:73
↓ 2 callersFunctionrope_apply
x: [B, L, N, C]. grid_sizes: [B, 3]. freqs: [M, C // 2].
wan/distributed/xdit_context_parallel.py:26
↓ 2 callersMethodset_seq_len
(self, seq_len)
wan/utils/vace_processor.py:103
↓ 2 callersFunctionsmart_nframes
calculate the number of frames for video used for model inputs. Args: ele (dict): a dict contains the configuration of video.
wan/utils/qwen_vl_utils.py:133
↓ 2 callersMethodvace_encode_frames
(self, frames, ref_images, masks=None, vae=None)
wan/vace.py:126
↓ 2 callersMethodvace_encode_masks
(self, masks, ref_images=None, vae_stride=None)
wan/vace.py:156
↓ 2 callersMethodvace_latent
(self, z, m)
wan/vace.py:190
↓ 2 callersFunctionwhitespace_clean
(text)
wan/modules/tokenizers.py:18
↓ 1 callersMethod_clean
(self, text)
wan/modules/tokenizers.py:75
↓ 1 callersFunction_clip
(pretrained=False, pretrained_name=None, model_cls=XLMRobertaCLIP, return_transf
wan/modules/clip.py:434
↓ 1 callersMethod_cross_attn_ffn
Device-consistent cross-attention and FFN processing
wan/modules/model.py:442
↓ 1 callersMethod_get_frameid_bbox
(self, fps, frame_timestamps, h, w, crop_box, rng)
wan/utils/vace_processor.py:214
↓ 1 callersMethod_get_frameid_bbox_adjust_last
(self, fps, frame_timestamps, h, w, crop_box, rng)
wan/utils/vace_processor.py:181
↓ 1 callersMethod_get_frameid_bbox_default
(self, fps, frame_timestamps, h, w, crop_box, rng)
wan/utils/vace_processor.py:149
↓ 1 callersMethod_image_preprocess
(self, img, oh, ow, normalize=True, **kwargs)
wan/utils/vace_processor.py:60
↓ 1 callersFunction_init_logging
(rank)
generate.py:298
↓ 1 callersMethod_init_step_index
Initialize the step_index counter for the scheduler.
wan/utils/fm_solvers.py:693
↓ 1 callersMethod_init_step_index
Initialize the step_index counter for the scheduler.
wan/utils/fm_solvers_unipc.py:645
↓ 1 callersMethod_load_image
(self, img_path)
wan/utils/vace_processor.py:28
↓ 1 callersMethod_norm
(self, x)
wan/modules/model.py:151
↓ 1 callersFunction_parse_args
()
generate.py:119
↓ 1 callersMethod_pillow_convert
(self, image, cvt_type='RGB')
wan/utils/vace_processor.py:14
↓ 1 callersMethod_relative_position_bucket
(self, rel_pos)
wan/modules/t5.py:245
↓ 1 callersMethod_resize_crop
Resize, center crop, convert to tensor, and normalize.
wan/utils/vace_processor.py:35
↓ 1 callersMethod_restore_model_state
Restore model parameter states
motion_optimizer.py:170
↓ 1 callersMethod_save_model_state
Save model parameter states to restore later
motion_optimizer.py:162
↓ 1 callersFunction_t5
(name, encoder_only=False, decoder_only=False, return_tokenizer=False, tokeniz
wan/modules/t5.py:415
↓ 1 callersFunction_validate_args
(args)
generate.py:82
↓ 1 callersMethod_video_preprocess
(self, video, oh, ow)
wan/utils/vace_processor.py:146
↓ 1 callersFunction_video_vae
Autoencoder3d adapted from Stable Diffusion 1.x, 2.x and XL.
wan/modules/vae.py:592
↓ 1 callersFunctioncache_image
(tensor, save_file, nrow=8, normalize=True, va
wan/utils/utils.py:64
↓ 1 callersFunctioncache_video
(tensor, save_file=None, fps=30, suffix='.mp4',
wan/utils/utils.py:23
↓ 1 callersMethodcalculate_motion_variance
Calculate motion variance in a differentiable way from noise prediction. Handles 3D tensors with shape [C, F, H*W].
motion_optimizer.py:95
↓ 1 callersFunctioncanonicalize
(text, keep_punctuation_exact_string=None)
wan/modules/tokenizers.py:24
↓ 1 callersMethodclear_all
Clear all saved states
motion_optimizer.py:358
↓ 1 callersFunctionclip_xlm_roberta_vit_h_14
( pretrained=False, pretrained_name='open-clip-xlm-roberta-large-vit-huge-14', **kwarg
wan/modules/clip.py:471
↓ 1 callersMethodconvert_model_output
Convert the model output to the corresponding type the DPMSolver/DPMSolver++ algorithm needs. DPM-Solver is designed to discretize an
wan/utils/fm_solvers.py:341
↓ 1 callersMethodconvert_model_output
r""" Convert the model output to the corresponding type the UniPC algorithm needs. Args: model_output (`torch.Tensor`):
wan/utils/fm_solvers_unipc.py:279
↓ 1 callersMethoddecide_system_prompt
(self, tar_lang="zh", multi_images_input=False)
wan/utils/prompt_extend.py:184
↓ 1 callersMethoddecode
(self, z, scale)
wan/modules/vae.py:544
↓ 1 callersMethoddisable
Restore original forward method
motion_optimizer.py:57
↓ 1 callersMethoddpm_solver_first_order_update
One step for the first-order DPMSolver (equivalent to DDIM). Args: model_output (`torch.Tensor`): The dir
wan/utils/fm_solvers.py:415
↓ 1 callersMethoddynamic_load
(self)
wan/vace.py:470
↓ 1 callersMethodenable
Enable gradient checkpointing by patching the forward method
motion_optimizer.py:20
↓ 1 callersMethodextend_with_img
(self, prompt, system_prompt, image=No
wan/utils/prompt_extend.py:172
↓ 1 callersFunctionextract_vision_info
( conversations: list[dict] | list[list[dict]])
wan/utils/qwen_vl_utils.py:328
↓ 1 callersFunctionfetch_video
( ele: dict, image_factor: int = IMAGE_FACTOR)
wan/utils/qwen_vl_utils.py:274
↓ 1 callersMethodforward
(self, x)
wan/modules/t5.py:48
↓ 1 callersFunctiongenerate
(args)
generate.py:310
↓ 1 callersMethodgenerate
r""" Generates video frames from input first-last frame and text prompt using diffusion process. Args: input_prompt (`str
wan/first_last_frame2video.py:129
↓ 1 callersMethodgenerate
r""" Generates video frames from input image and text prompt using diffusion process. Args: input_prompt (`str`):
wan/image2video.py:129
next →1–100 of 271, ranked by callers