MCPcopy Create free account

hub / github.com/PKU-YuanGroup/Helios / functions

Functions847 in github.com/PKU-YuanGroup/Helios

⨍ Functions 847◇ Types & classes 124
↓ 2 callersFunctionoptimized_scale
(positive_flat, negative_flat)
helios/pipelines/pipeline_helios_ode.py:86
↓ 2 callersFunctionoptimized_scale
(positive_flat, negative_flat)
helios/pipelines/pipeline_helios.py:87
↓ 2 callersFunctionpred_flow
(img1, img2)
eval/utils/third_party/amt/flow_generation/gen_flow.py:27
↓ 2 callersFunctionprocess_and_update_error_buffers
( args, recycle_vars, accelerator, global_step, noise_scheduler_copy, model_pred,
helios/utils/utils_recycle_batch.py:594
↓ 2 callersFunctionrandom_crop_woflow
(img0, imgt, img1, crop_size=(224, 224))
eval/utils/third_party/amt/datasets/gopro_datasets.py:20
↓ 2 callersFunctionrandom_horizontal_flip_woflow
(img0, imgt, img1, p=0.5)
eval/utils/third_party/amt/datasets/gopro_datasets.py:56
↓ 2 callersFunctionrandom_resize_woflow
(img0, imgt, img1, p=0.1)
eval/utils/third_party/amt/datasets/gopro_datasets.py:12
↓ 2 callersFunctionrandom_reverse_channel_woflow
(img0, imgt, img1, p=0.5)
eval/utils/third_party/amt/datasets/gopro_datasets.py:40
↓ 2 callersFunctionrandom_reverse_time_woflow
(img0, imgt, img1, embt, p=0.5)
eval/utils/third_party/amt/datasets/gopro_datasets.py:72
↓ 2 callersFunctionrandom_rotate_woflow
(img0, imgt, img1, p=0.05)
eval/utils/third_party/amt/datasets/gopro_datasets.py:64
↓ 2 callersFunctionrandom_vertical_flip_woflow
(img0, imgt, img1, p=0.3)
eval/utils/third_party/amt/datasets/gopro_datasets.py:48
↓ 2 callersFunctionreplace_linear_with_tiled_linear
(model, num_shards=None, patch_by_names=True, patch_by_types=True)
helios/modules/helios_kernels/tiled_linear.py:15
↓ 2 callersMethodreset
(self, name=None)
eval/utils/third_party/amt/utils/utils.py:42
↓ 2 callersMethodresize
(self, x, scale_factor)
eval/utils/third_party/amt/losses/loss.py:192
↓ 2 callersFunctionrun_experiment
运行完整实验
tools/others/benchmark/benchmark_patchification_performance.py:211
↓ 2 callersFunctionrun_generator
( args, accelerator, transformer, scheduler, noise, prompt_embeds, # For VRAM mana
helios/utils/utils_helios_post.py:1501
↓ 2 callersFunctionsample_dynamic_dmd_num_latent_sections
( min_sections: int = 3, max_sections: int = 3, dmd_dynamic_alpha: float = 1.5, dmd_dynamic_be
helios/utils/utils_helios_post.py:463
↓ 2 callersFunctionsave_model_checkpoint
( transformer, args, save_path, weight_dtype=None, unwrap_model_fn=None, get_peft_mode
helios/utils/utils_base.py:403
↓ 2 callersMethodset_epoch
(self, epoch)
helios/dataset/dataloader_history_latents_dist.py:316
↓ 2 callersMethodset_epoch
(self, epoch)
helios/dataset/dataloader_mp4_dist.py:471
↓ 2 callersMethodset_timesteps
Setting the timesteps and sigmas for each stage
helios/diffusers_version/scheduling_helios_diffusers.py:195
↓ 2 callersFunctionsetup_pipeline
设置pipeline
tools/others/benchmark/benchmark_compile_performance.py:43
↓ 2 callersFunctionsmart_nframes
calculate the number of frames for video used for model inputs. Args: ele (dict): a dict contains the configuration of video.
helios/videoalign/vision_process.py:124
↓ 2 callersMethodstep_unipc
Predict the sample from the previous timestep by reversing the SDE. This function propagates the sample with the multistep UniPC.
helios/scheduler/scheduling_helios.py:727
↓ 2 callersFunctiontensor2img
(img_t)
eval/utils/third_party/amt/utils/utils.py:90
↓ 2 callersMethodtime_shift
Apply time shifting to the sigmas. Args: mu (`float`): The mu parameter for the time shift.
helios/diffusers_version/scheduling_helios_diffusers.py:259
↓ 2 callersMethodtransform
(self, tensor)
eval/utils/third_party/amt/losses/loss.py:64
↓ 2 callersMethodtransform
(self, tensor)
eval/utils/third_party/amt/losses/loss.py:98
↓ 2 callersFunctionunwrap_model
(model)
helios/utils/utils_helios_post.py:1640
↓ 2 callersMethodupdate
(self, dict, n=1)
eval/utils/third_party/amt/utils/utils.py:36
↓ 2 callersFunctionupdate_error_buffers_distributed
Update error buffers with samples gathered from all processes. Args: gathered_noise_errors: shape [num_gpus, batch_size, ...] gath
helios/utils/utils_recycle_batch.py:561
↓ 2 callersFunctionupdate_error_buffers_local
Update error buffers with samples from local GPU only (post-warmup). Args: noise_error: shape [batch_size, ...] y_error: shape [ba
helios/utils/utils_recycle_batch.py:583
↓ 2 callersMethodupdate_global_statistics
(self, current_mean: torch.Tensor, current_var: torch.Tensor)
helios/utils/utils_base.py:711
↓ 2 callersFunctionupdate_videos
Update videos based on selected idx and two steps
tools/gradio/comparison/gradio_compare_diff-video.py:170
↓ 2 callersFunctionupdate_videos_from_selectors
Update videos based on selected step and idx
tools/gradio/comparison/gradio_compare_diff-ablation.py:200
↓ 2 callersFunctionupdate_videos_from_selectors
Update videos based on selected step and idx
tools/gradio/comparison/gradio_compare_diff-ckpt.py:205
↓ 2 callersFunctionwrite
(file, data)
eval/utils/third_party/amt/utils/utils.py:120
↓ 2 callersFunctionwritePFM
(file, image, scale=1)
eval/utils/third_party/amt/utils/utils.py:177
↓ 1 callersMethod__init__
( self, dim: int, dim_out: Optional[int] = None, mult: int = 4, dropou
helios/modules/helios_kernels/tiled_linear.py:225
↓ 1 callersMethod_apply_global_ratio_sampling
(self)
helios/dataset/dataloader_history_latents_dist.py:350
↓ 1 callersMethod_apply_global_ratio_sampling
(self)
helios/dataset/dataloader_mp4_dist.py:505
↓ 1 callersMethod_build_file_metadata
(self, json_file, video_folder)
helios/dataset/dataloader_mp4_dist.py:220
↓ 1 callersMethod_build_folder_metadata
(self, folder)
tools/others/get_mean_std_size9.py:84
↓ 1 callersMethod_build_folder_metadata
(self, folder)
helios/dataset/dataloader_history_latents_dist.py:78
↓ 1 callersMethod_build_folder_metadata
(self, folder, data_type)
helios/dataset/dataloader_dmd.py:119
↓ 1 callersFunction_critic_loss
( args, critic_accelerator, fake_score_model, transformer, scheduler, noise, promp
helios/utils/utils_helios_post.py:2788
↓ 1 callersFunction_downscale_maps
Resize flow maps for score calculation
eval/1_get_motion_amplitude.py:16
↓ 1 callersFunction_flash_attn_varlen_wrapper
(q, k, v, cu_seqlens_q, cu_seqlens_kv, max_seqlen_q, max_seqlen_kv)
helios/modules/helios_kernels/attention_dispatch.py:128
↓ 1 callersFunction_flash_attn_wrapper
(q, k, v)
helios/modules/helios_kernels/attention_dispatch.py:123
↓ 1 callersFunction_flow_loss
( args, accelerator, lr_scheduler, transformer, prompt_embeds, prompt_attention_masks,
helios/utils/utils_helios_base.py:20
↓ 1 callersMethod_forward
(self, **kwargs)
eval/utils/third_party/amt/losses/loss.py:28
↓ 1 callersFunction_generator_loss
( args, accelerator, real_fake_score_model, transformer, scheduler, noise, prompt_
helios/utils/utils_helios_post.py:2141
↓ 1 callersFunction_get_qkv_projections
(attn: "HeliosAttention", hidden_states: torch.Tensor, encoder_hidden_states: torch.Tensor)
helios/modules/transformer_helios.py:75
↓ 1 callersFunction_get_qkv_projections
(attn: "HeliosAttention", hidden_states: torch.Tensor, encoder_hidden_states: torch.Tensor)
helios/diffusers_version/transformer_helios_diffusers.py:65
↓ 1 callersMethod_get_spatial_meshgrid
(self, height, width, device_str)
helios/modules/transformer_helios.py:689
↓ 1 callersMethod_get_spatial_meshgrid
(self, height, width, device_str)
helios/diffusers_version/transformer_helios_diffusers.py:347
↓ 1 callersFunction_get_t5_prompt_embeds
( tokenizer, text_encoder, prompt: Union[str, List[str]] = None, num_videos_per_prompt: int =
helios/utils/utils_base.py:593
↓ 1 callersMethod_init_dataset
(self)
eval/utils/third_party/amt/trainers/base_trainer.py:97
↓ 1 callersMethod_init_logger
(self)
eval/utils/third_party/amt/trainers/base_trainer.py:62
↓ 1 callersMethod_init_loss
(self)
eval/utils/third_party/amt/trainers/base_trainer.py:113
↓ 1 callersFunction_insert_adapter_name_into_state_dict
Utility function to remap the state_dict keys to fit the PEFT model by inserting the adapter name.
helios/videoalign/utils.py:146
↓ 1 callersFunction_motion_score
Calculate mean score from maps or masks
eval/1_get_motion_amplitude.py:26
↓ 1 callersMethod_norm
(self, reward)
helios/videoalign/inference.py:72
↓ 1 callersMethod_prepare_dataset_buckets
(self)
helios/dataset/dataloader_history_latents_dist.py:304
↓ 1 callersMethod_prepare_dataset_buckets
(self)
helios/dataset/dataloader_mp4_dist.py:459
↓ 1 callersMethod_prepare_input
Prepare `inputs` before feeding them to the model, converting them to tensors if they are not already and handling potential state.
helios/videoalign/inference.py:99
↓ 1 callersMethod_prepare_inputs
Prepare `inputs` before feeding them to the model, converting them to tensors if they are not already and handling potential state.
helios/videoalign/inference.py:119
↓ 1 callersMethod_process_folder
(self, folder, cache_file)
tools/others/get_mean_std_size9.py:56
↓ 1 callersMethod_process_folder
(self, folder, cache_file)
helios/dataset/dataloader_history_latents_dist.py:52
↓ 1 callersMethod_process_folder
(self, folder, cache_file, data_type)
helios/dataset/dataloader_dmd.py:98
↓ 1 callersMethod_process_json_file
(self, json_file, video_folder, cache_file)
helios/dataset/dataloader_mp4_dist.py:191
↓ 1 callersMethod_save_pretrained_with_prefix
(self, state_dict, save_dir, weight_file_prefix)
helios/utils/create_ema_zero3.py:195
↓ 1 callersMethod_shard_indices_for_sp_group
Shard indices across SP groups, similar to DP_SP_BatchSampler. Each SP group gets a disjoint subset of the data.
helios/dataset/dataloader_history_latents_dist.py:319
↓ 1 callersMethod_shard_indices_for_sp_group
Shard indices across SP groups, similar to DP_SP_BatchSampler. Each SP group gets a disjoint subset of the data.
helios/dataset/dataloader_mp4_dist.py:474
↓ 1 callersMethod_shard_indices_for_sp_group
Shard indices across SP groups. Each SP group gets a disjoint subset of the data.
helios/dataset/dataloader_dmd.py:354
↓ 1 callersMethod_time_shift_exponential
(self, mu, sigma, t)
helios/diffusers_version/scheduling_helios_diffusers.py:281
↓ 1 callersMethod_time_shift_linear
(self, mu, sigma, t)
helios/diffusers_version/scheduling_helios_diffusers.py:285
↓ 1 callersMethodadd_noise
(self, original_samples, noise, timestep, sigmas, timesteps)
helios/diffusers_version/scheduling_helios_diffusers.py:830
↓ 1 callersFunctionapply_error_injection
( args, recycle_vars, model_input, noise, timesteps, latents_history_long, latents
helios/utils/utils_recycle_batch.py:8
↓ 1 callersMethodattention
(self, x: torch.Tensor)
eval/utils/third_party/ViCLIP/viclip_text.py:56
↓ 1 callersMethodattention
(self, x)
eval/utils/third_party/ViCLIP/viclip_vision.py:76
↓ 1 callersMethodbackward
(self, gradOutput)
eval/utils/third_party/amt/flow_generation/liteflownet/correlation/correlation.py:332
↓ 1 callersFunctionbase_build_fn
(module, cls, params)
eval/utils/third_party/amt/utils/build_utils.py:10
↓ 1 callersFunctionbasic_clean
(text)
eval/utils/third_party/ViCLIP/simple_tokenizer.py:62
↓ 1 callersFunctionbasic_clean
(text)
helios/utils/utils_base.py:576
↓ 1 callersFunctionbasic_clean
(text)
helios/diffusers_version/pipeline_helios_diffusers.py:93
↓ 1 callersFunctionbasic_clean
(text)
helios/pipelines/pipeline_helios_ode.py:99
↓ 1 callersFunctionbasic_clean
(text)
helios/pipelines/pipeline_helios.py:100
↓ 1 callersMethodbpe
(self, token)
eval/utils/third_party/ViCLIP/simple_tokenizer.py:95
↓ 1 callersMethodbuild_attention_mask
(self)
eval/utils/third_party/ViCLIP/viclip_text.py:118
↓ 1 callersMethodbuild_text_encoder
build text_encoder and possiblly video-to-text multimodal fusion encoder. Returns: nn.Module. The text encoder
eval/utils/third_party/ViCLIP/viclip.py:173
↓ 1 callersMethodbuild_vision_encoder
build vision encoder Returns: (vision_encoder, vision_layernorm). Each is a `nn.Module`.
eval/utils/third_party/ViCLIP/viclip.py:154
↓ 1 callersFunctioncalculate_shift
( image_seq_len, base_seq_len: int = 256, max_seq_len: int = 4096, base_shift: float = 0.5,
helios/utils/utils_base.py:518
↓ 1 callersFunctioncalculate_weighted_rating
(summary_data, is_long)
eval/9_merge_all_scores.py:104
↓ 1 callersFunctioncall_gpt
Call GPT API to evaluate video naturalness
eval/8_get_drifting_naturalness.py:105
↓ 1 callersFunctioncall_gpt
Call GPT API to evaluate video naturalness
eval/4_get_naturalness.py:82
↓ 1 callersMethodcheck_inputs
( self, prompt, negative_prompt, height, width, prompt_embeds=
helios/diffusers_version/pipeline_helios_diffusers.py:297
↓ 1 callersMethodcheck_inputs
( self, prompt, negative_prompt, height, width, prompt_embeds=
helios/pipelines/pipeline_helios_ode.py:291
↓ 1 callersMethodcheck_inputs
( self, prompt, negative_prompt, height, width, prompt_embeds=
helios/pipelines/pipeline_helios.py:292
↓ 1 callersMethodclear_cache
(self)
helios/modules/transformer_helios.py:193
← previousnext →201–300 of 847, ranked by callers