MCPcopy Create free account

hub / github.com/MotrixLab/ViMoGen / functions

Functions457 in github.com/MotrixLab/ViMoGen

Methodget_fsdp_wrap_module_list
(self, )
models/transformer/wan/modules/tm2m_model.py:647
Functionget_image_size
(resolution, ar_ratio)
datasets/aspect.py:467
Functionget_local_rank
()
parallel/utils.py:35
Functionget_local_world_size
()
parallel/utils.py:39
Methodget_min_snr_weight
the original min snr weighting in paper https://arxiv.org/abs/2303.09556.
trainer/scheduler.py:17
Methodget_min_snr_weight_cogvideo
modified loss weighting, so that the weighting at zero snr is not zero the original min snr weighting in paper https://arxiv.org/abs/2
trainer/scheduler.py:45
Methodget_min_snr_weight_modified
modified loss weighting, so that the weighting at zero snr is not zero the original min snr weighting in paper https://arxiv.org/abs/2
trainer/scheduler.py:29
Functionget_node_rank
()
parallel/utils.py:56
Functionget_num_frames
(num_frames)
datasets/aspect.py:492
Methodget_prob
(self, bucket_id)
datasets/bucket.py:136
Methodget_projection_matrix
(self, width=None, height=None)
mbench/render.py:39
Functionget_prompt_from_filename
Extract prompt from video filename by removing extensions and special characters Args: filename: Video filename Ret
mbench/utils.py:8
Functionget_rank
()
parallel/utils.py:43
Functionget_scheduler
(beta_type: str = None, use_dpm_solver: bool = False)
trainer/scheduler.py:93
Functionget_side_list_64_base
()
datasets/bucket_config.py:76
Functionget_smooth_dynamic_cfg_scale_list
(cfg_scale, num_inference_steps)
trainer/scheduler.py:119
Functionguess_cogvideo_config_from_state_dict
(state_dict: dict)
models/transformer/utils.py:128
Functionguess_cogvideo_config_from_state_dict
(state_dict: dict)
models/transformer/wan/modules/utils.py:128
Functionhash_state_dict_keys
(state_dict, with_shape=True)
models/transformer/utils.py:205
Functionhelp
(args)
mbench/cli/mbench.py:18
Functioninit_weights
(m)
models/transformer/wan/modules/t5.py:27
Functionload_safetensors
(in_path: str)
models/transformer/wan/modules/utils.py:102
Functionload_single_motion
(entry)
scripts/prepare_training_data.py:30
Methodload_state_dict
(self, state_dict: dict)
datasets/sampler.py:56
Methodload_state_dict
(self, state_dict: dict)
datasets/sampler.py:317
Methodload_state_dict
(self, state_dict: dict)
datasets/sampler.py:358
Methodloss_weight
(self, sigmas, weight_type: str = 'SNR')
trainer/sd3_scheduler.py:155
Functionlr_lambda
(current_step)
trainer/base_trainer.py:414
Functionmain
()
mbench/cli/mbench.py:7
Methodmerge_bucket
(bucket_dict: OrderedDict, min_length: int)
datasets/multi_resolution_sampler.py:41
Functionon_main_process
Decorator to selectively run the decorated function on the main process only based on the `main_process_only` attribute in a class. Checks at
parallel/utils.py:15
Functionprepare_extra_step_kwargs
(scheduler, generator, eta=0.0)
models/transformer/utils.py:73
Functionprepare_extra_step_kwargs
(scheduler, generator, eta=0.0)
models/transformer/wan/modules/utils.py:73
Methodqkv_fn
(x)
models/transformer/wan/modules/tm2m_model.py:288
Methodqkv_fn
(x)
models/transformer/wan/modules/t2m_model.py:287
Functionquaternion_to_rot6d
Convert 4d quaternion to 6d rotation representation. Shape: - Input: :Torch:`(N, 4)` - Output: :Torch:`(N, 6)`
motion_rep/rotation_transform.py:102
Functionrandn_tensor
A helper function to create random tensors on the desired `device` with the desired `dtype`. When passing a list of generators, you can seed
models/transformer/wan/modules/utils.py:8
Functionregister_subparsers
(subparser)
mbench/cli/evaluate.py:7
Methodreset
(self)
datasets/sampler.py:50
Methodreset
(self)
datasets/sampler.py:352
Methodretrieve_inference_timesteps_and_sigma
( self, num_inference_steps: int, frames: int, height: int, width: int
trainer/sd3_scheduler.py:141
Methodreturn_to_timestep
(self, timestep, sample, sample_stablized)
trainer/scheduler.py:199
Functionrot6d_to_quaternion
Convert 6d rotation representation to 4d quaternion. Shape: - Input: :Torch:`(N, 4)` - Output: :Torch:`(N, 6)`
motion_rep/rotation_transform.py:112
Methodsample_t
(self, batch_size: int, training_iter: int = None)
trainer/scheduler.py:81
Methodsample_t_and_sigma
( self, batch_size: int, frames: int, height: int, width: int,
trainer/sd3_scheduler.py:30
Functionseed_worker
(worker_id)
datasets/dataloader.py:18
Functionsmpl_to_joints_and_verts
Convert SMPL pose parameters to joint positions and vertices. smpl_model: SMPL model object smpl_params: Dictionary of SMPL parameters (e
mbench/pose_quality.py:38
Methodstate_dict
(self, step)
datasets/sampler.py:53
Methodstate_dict
(self, num_steps: int)
datasets/sampler.py:306
Methodstate_dict
(self, step)
datasets/sampler.py:355
Methodstate_dict_converter
()
models/transformer/wan/modules/tm2m_model.py:785
Methodstate_dict_converter
()
models/transformer/wan/modules/t2m_model.py:705
Methodstep
(self, model_output, timestep, sample, to_final=False)
trainer/scheduler.py:186
Functionstringify_cmd
(cmd_ls)
mbench/cli/evaluate.py:48
Methodtracker
(self)
parallel/logging.py:91
Functionvalidate_video_path
Validate that a video file exists and has a supported extension Args: video_path: Path to video file Returns:
mbench/utils.py:62
Methodwarning_once
This method is identical to `logger.warning()`, but will emit the warning with the same message only once. Note: The cache is for the
parallel/logging.py:57
← previous401–457 of 457, ranked by callers