MCPcopy Create free account

hub / github.com/Tencent-Hunyuan/HunyuanVideo-I2V / functions

Functions333 in github.com/Tencent-Hunyuan/HunyuanVideo-I2V

⨍ Functions 333◇ Types & classes 54
↓ 1 callersMethodcompute_d_alpha_alpha_ratio_t
Compute the ratio between d_alpha and alpha
hyvideo/diffusion/flow/path.py:36
↓ 1 callersMethodcompute_diffusion
Compute the diffusion term of the SDE Args: x: [batch_dim, ...], data point t: [batch_dim,], time vector form: s
hyvideo/diffusion/flow/path.py:50
↓ 1 callersMethodcompute_mu_t
Compute the mean of time-dependent density p_t
hyvideo/diffusion/flow/path.py:119
↓ 1 callersMethodcompute_ut
Compute the vector field corresponding to p_t
hyvideo/diffusion/flow/path.py:134
↓ 1 callersMethodcompute_xt
Sample xt from time-dependent density p_t; rng is required
hyvideo/diffusion/flow/path.py:129
↓ 1 callersFunctionconvert_fp8_linear
(module, dit_weight_path, original_dtype, params_to_keep={})
hyvideo/modules/fp8_optimization.py:82
↓ 1 callersFunctionconvert_kohya_to_peft_keys
( kohya_dict: dict, kohya_prefix="", peft_prefix: str = "base_model.model", device="cpu", )
hyvideo/utils/train_utils.py:24
↓ 1 callersFunctioncreate_transport
( *, path_type, prediction, loss_weight=None, train_eps=None,
hyvideo/diffusion/__init__.py:5
↓ 1 callersMethoddisable_spatial_tiling
(self)
hyvideo/vae/autoencoder_kl_causal_3d.py:147
↓ 1 callersMethoddisable_temporal_tiling
(self)
hyvideo/vae/autoencoder_kl_causal_3d.py:141
↓ 1 callersFunctiondump_args
(args, save_path, extra_args=None)
hyvideo/utils/file_utils.py:104
↓ 1 callersFunctiondump_codes
Dump codes to the experiment directory. Args: save_path (str): Path to the experiment directory. root (Path): Path to the ro
hyvideo/utils/file_utils.py:66
↓ 1 callersFunctionempty_logger
()
hyvideo/utils/file_utils.py:116
↓ 1 callersMethodencode
Encode a batch of images/videos into latents. Args: x (`torch.FloatTensor`): Input batch of images/videos.
hyvideo/vae/autoencoder_kl_causal_3d.py:260
↓ 1 callersFunctionextract
( vae: torch.nn.Module, meta_files: List[str], output_base_dir: str, sample_n_frames: int,
hyvideo/hyvae_extract/run.py:31
↓ 1 callersFunctionfp8_activation_dequant
(qdq_out, scale, dtype)
hyvideo/modules/fp8_optimization.py:50
↓ 1 callersFunctionfp8_linear_forward
(cls, original_dtype, input)
hyvideo/modules/fp8_optimization.py:55
↓ 1 callersFunctionfp8_tensor_quant
(x, scale, bits=8, mantissa_bit=3, sign_bits=1)
hyvideo/modules/fp8_optimization.py:43
↓ 1 callersFunctiongenerate_crop_size_list
generate crop size list Args: base_size (int, optional): the base size for generate bucket. Defaults to 256. patch_size (int, opt
hyvideo/utils/data_utils.py:64
↓ 1 callersFunctionget_1d_rotary_pos_embed
Precompute the frequency tensor for complex exponential (cis) with given dimensions. (Note: `cis` means `cos + i * sin`, where i is the imagi
hyvideo/modules/posemb_layers.py:261
↓ 1 callersFunctionget_build_config
Obtain the build information of PyTorch or Parrots.
utils/collect_env.py:27
↓ 1 callersMethodget_closest_ratio
(width: float, height: float, ratios: list, buckets: list)
hyvideo/hyvae_extract/dataset.py:216
↓ 1 callersFunctionget_cond_images
get conditioned images by decode the first frame latents
hyvideo/utils/train_utils.py:115
↓ 1 callersFunctionget_cond_latents
get conditioned latent by decode and encode the first frame latents
hyvideo/utils/train_utils.py:83
↓ 1 callersFunctionget_cu_seqlens
Calculate cu_seqlens_q, cu_seqlens_kv using text_mask and img_len Args: text_mask (torch.Tensor): the mask of text img_len (int):
hyvideo/modules/attenion.py:34
↓ 1 callersMethodget_data_info
(self, index)
hyvideo/dataset/video_loader.py:96
↓ 1 callersFunctionget_deepspeed_config
(args: argparse.Namespace, micro_batch_size: int, global_bat
hyvideo/ds_config.py:14
↓ 1 callersFunctionget_down_block3d
( down_block_type: str, num_layers: int, in_channels: int, out_channels: int, temb_ch
hyvideo/vae/unet_causal_3d_blocks.py:418
↓ 1 callersMethodget_drift
member function for obtaining the drift of the probability flow ODE
hyvideo/diffusion/flow/transport.py:236
↓ 1 callersFunctionget_experiment_max_number
(experiments)
hyvideo/utils/file_utils.py:130
↓ 1 callersFunctionget_fp_maxval
(bits=8, mantissa_bit=3, sign_bits=1)
hyvideo/modules/fp8_optimization.py:7
↓ 1 callersMethodget_item
(self, idx)
hyvideo/hyvae_extract/dataset.py:143
↓ 1 callersFunctionget_meshgrid_nd
Get n-D meshgrid with start, stop and num. Args: start (int or tuple): If len(args) == 0, start is num; If len(args) == 1, start is
hyvideo/modules/posemb_layers.py:14
↓ 1 callersMethodget_model_t
(self, t)
hyvideo/diffusion/flow/transport.py:157
↓ 1 callersFunctionget_module_kohya_state_dict
(module, prefix, dtype, adapter_name="default")
hyvideo/utils/train_utils.py:425
↓ 1 callersFunctionget_rope_freq_from_size
( args, model, latents_size, ndim, target_ndim, rope_theta_rescale_factor=1.0, rop
hyvideo/utils/train_utils.py:374
↓ 1 callersMethodget_rotary_pos_embed
(self, video_length, height, width)
hyvideo/inference.py:609
↓ 1 callersMethodget_score
member function for obtaining score of x_t = alpha_t * x + sigma_t * eps
hyvideo/diffusion/flow/transport.py:269
↓ 1 callersMethodget_score_from_velocity
Wrapper function: transfrom velocity prediction model to score Args: velocity: [batch_dim, ...] shaped tensor; velocity model outp
hyvideo/diffusion/flow/path.py:75
↓ 1 callersFunctionget_tensorboard_config
(output_dir: str, job_name: str)
hyvideo/ds_config.py:5
↓ 1 callersFunctionget_up_block3d
( up_block_type: str, num_layers: int, in_channels: int, out_channels: int, prev_outp
hyvideo/vae/unet_causal_3d_blocks.py:471
↓ 1 callersMethodindex_for_timestep
(self, timestep, schedule_timesteps=None)
hyvideo/diffusion/schedulers/scheduling_flow_match_discrete.py:158
↓ 1 callersFunctionis_cuda_available
Returns True if cuda devices exist.
utils/collect_env.py:44
↓ 1 callersFunctionis_musa_available
()
utils/collect_env.py:41
↓ 1 callersFunctionis_rocm_pytorch
Check whether the PyTorch is compiled on ROCm.
utils/collect_env.py:13
↓ 1 callersFunctionis_valid_experiment
(path)
hyvideo/utils/file_utils.py:123
↓ 1 callersFunctionload_denoiser
(args)
hyvideo/diffusion/__init__.py:74
↓ 1 callersMethodload_diffusion_pipeline
( self, args, vae, text_encoder, text_encoder_2, model,
hyvideo/inference.py:571
↓ 1 callersFunctionload_lora
(model, lora_path, device)
hyvideo/utils/train_utils.py:56
↓ 1 callersFunctionload_lora_for_pipeline
( pipeline, lora_path, LORA_PREFIX_TRANSFORMER="", LORA_PREFIX_TEXT_ENCODER="", alpha=1.0,
hyvideo/utils/lora_utils.py:6
↓ 1 callersFunctionload_state_dict
(args, model, logger)
hyvideo/utils/train_utils.py:136
↓ 1 callersFunctionload_text_encoder
( text_encoder_type, text_encoder_precision=None, text_encoder_path=None, logger=None, dev
hyvideo/text_encoder/__init__.py:25
↓ 1 callersFunctionload_tokenizer
( tokenizer_type, tokenizer_path=None, padding_side="right", logger=None )
hyvideo/text_encoder/__init__.py:69
↓ 1 callersFunctionmain
()
sample_image2video.py:12
↓ 1 callersFunctionmain
(args)
train_image2video_lora.py:278
↓ 1 callersFunctionmain
( local_rank: int, vae_path: str, meta_files: str, output_base_dir: str, sample_n_frames:
hyvideo/hyvae_extract/run.py:102
↓ 1 callersFunctionparallelize_transformer
(pipe)
hyvideo/inference.py:110
↓ 1 callersMethodparams_count
(self)
hyvideo/modules/models.py:816
↓ 1 callersMethodplan
(self, t, x0, x1)
hyvideo/diffusion/flow/path.py:144
↓ 1 callersMethodpredict
( self, prompt, height=192, width=336, video_length=129, seed=
hyvideo/inference.py:703
↓ 1 callersFunctionprepare_causal_attention_mask
(n_frame: int, n_hw: int, dtype, device, batch_size: int = None)
hyvideo/vae/unet_causal_3d_blocks.py:37
↓ 1 callersMethodprepare_latents
( self, batch_size, num_channels_latents, height, width, video
hyvideo/diffusion/pipelines/pipeline_hunyuan_video.py:566
↓ 1 callersFunctionprepare_model_inputs
( args, batch: tuple, device: Union[int, str], model, vae, text_encoder, text_enco
hyvideo/utils/train_utils.py:243
↓ 1 callersFunctionpreprocess_text_encoder_tokenizer
(args)
hyvideo/utils/preprocess_text_encoder_tokenizer_utils.py:9
↓ 1 callersMethodpreprocess_train
(self, frames)
hyvideo/hyvae_extract/dataset.py:192
↓ 1 callersMethodpreprocess_url
(self, data_json_path)
hyvideo/hyvae_extract/dataset.py:132
↓ 1 callersMethodprior_logp
Standard multivariate normal prior Assume z is batched
hyvideo/diffusion/flow/transport.py:89
↓ 1 callersFunctionprofiler_context
(enable, exp_dir, worker_name)
hyvideo/utils/helpers.py:37
↓ 1 callersFunctionquantize_to_fp8
Default is E4M3.
hyvideo/modules/fp8_optimization.py:20
↓ 1 callersMethodrequest_ceph_data
(self, path)
hyvideo/hyvae_extract/dataset.py:124
↓ 1 callersFunctionrescale_noise_cfg
Rescale `noise_cfg` according to `guidance_rescale`. Based on findings of [Common Diffusion Noise Schedules and Sample Steps are Flawed](http
hyvideo/diffusion/pipelines/pipeline_hunyuan_video.py:57
↓ 1 callersMethodreset
(self)
train_image2video_lora.py:218
↓ 1 callersFunctionretrieve_timesteps
Calls the scheduler's `set_timesteps` method and retrieves timesteps from the scheduler after the call. Handles custom timesteps. Any kwargs
hyvideo/diffusion/pipelines/pipeline_hunyuan_video.py:75
↓ 1 callersMethodsample
(self, generator: Optional[torch.Generator] = None)
hyvideo/vae/vae.py:313
↓ 1 callersFunctionsanity_check_args
(args)
hyvideo/config.py:582
↓ 1 callersFunctionsave_videos_grid
save videos by video tensor copy from https://github.com/guoyww/AnimateDiff/blob/e92bd5671ba62c0d774a32951453e328018b7c5b/animatediff/utils/uti
hyvideo/utils/file_utils.py:170
↓ 1 callersMethodscale_model_input
( self, sample: torch.Tensor, timestep: Optional[int] = None )
hyvideo/diffusion/schedulers/scheduling_flow_match_discrete.py:180
↓ 1 callersMethodsd3_time_shift
(self, t: torch.Tensor)
hyvideo/diffusion/schedulers/scheduling_flow_match_discrete.py:185
↓ 1 callersFunctionseed_everything
(seed)
hyvideo/hyvae_extract/run.py:24
↓ 1 callersFunctionset_reproducibility
(enable, global_seed=None)
hyvideo/utils/helpers.py:61
↓ 1 callersFunctionsetup_distributed_training
(args)
train_image2video_lora.py:62
↓ 1 callersFunctionsetup_experiment_directory
(args, rank)
train_image2video_lora.py:112
↓ 1 callersFunctionsplit_video_urls
(meta_files: str, global_rank: int, world_size: int)
hyvideo/hyvae_extract/dataset.py:17
↓ 1 callersMethodtemporal_tiled_decode
(self, z: torch.FloatTensor, return_dict: bool = True)
hyvideo/vae/autoencoder_kl_causal_3d.py:500
↓ 1 callersMethodtemporal_tiled_encode
(self, x: torch.FloatTensor, return_dict: bool = True)
hyvideo/vae/autoencoder_kl_causal_3d.py:465
↓ 1 callersFunctiontimestep_embedding
Create sinusoidal timestep embeddings. Args: t (torch.Tensor): a 1-D Tensor of N indices, one per batch element. These may be fracti
hyvideo/modules/embed_layers.py:93
↓ 1 callersMethodtraining_losses
(self, model, x1, model_kwargs=None, timestep=None, n_tokens=None, i2v_mode=False, con
hyvideo/diffusion/flow/transport.py:163
↓ 1 callersFunctiontry_save
(_save_name)
train_image2video_lora.py:257
↓ 1 callersMethodunpatchify
x: (N, T, patch_size**2 * C) imgs: (N, H, W, C)
hyvideo/modules/models.py:801
↓ 1 callersMethodwrite_events
(self, events)
hyvideo/utils/helpers.py:30
Method__Euler_Maruyama_step
(self, x, mean_x, t, model, **model_kwargs)
hyvideo/diffusion/flow/integrators.py:26
Method__Heun_step
(self, x, _, t, model, **model_kwargs)
hyvideo/diffusion/flow/integrators.py:36
Method__call__
r""" The call function to the pipeline for generation. Args: prompt (`str` or `List[str]`): The prompt or
hyvideo/diffusion/pipelines/pipeline_hunyuan_video.py:688
Method__call__
(self, worker_id)
hyvideo/utils/train_utils.py:224
Method__getitem__
(self, key)
hyvideo/diffusion/flow/utils.py:9
Method__getitem__
(self, idx)
hyvideo/hyvae_extract/dataset.py:246
Method__init__
( self, args, vae, vae_kwargs, text_encoder, model, te
hyvideo/inference.py:520
Method__init__
( self, in_channels: int = 3, out_channels: int = 3, up_block_types: Tuple
hyvideo/vae/vae.py:141
Method__init__
(self, parameters: torch.Tensor, deterministic: bool = False)
hyvideo/vae/vae.py:295
Method__init__
( self, channels: int, use_conv: bool = False, use_conv_transpose: bool =
hyvideo/vae/unet_causal_3d_blocks.py:82
← previousnext →101–200 of 333, ranked by callers