Functions — github.com/bytedance/StyleSSP

hub / github.com/bytedance/StyleSSP / functions

Functions440 in github.com/bytedance/StyleSSP

⨍ Functions 440◇ Types & classes 84

↓ 10 callersMethodencode_prompt

r""" Encodes the prompt into text encoder hidden states. Args: prompt (`str` or `List[str]`, *optional*):

ip_adapter/pipeline_stable_diffusion_extra_cfg.py:305

↓ 10 callersMethodget_decouple_embeds

( self, pil_image=None, clip_image_embeds=None, prompt=None, query=Non

ip_adapter/ip_adapter_instruct.py:126

↓ 10 callersMethodscale_model_input

Ensures interchangeability with schedulers that need to scale the denoising model input depending on the current timestep. Scales the

src/schedulers/euler_scheduler.py:37

↓ 8 callersFunctionget_image_paths

Returns the paths of the images in the specified directory, filtered by allowed file extensions. Args: path (str): Path to image director

evaluation/eval_artfid.py:158

↓ 7 callersMethod__init__

( self, num_classes1: int = 1000, num_classes2: int = 1000, aux_logits: bool =

src/evaluate/inception.py:22

↓ 7 callersMethod__init__

( self, num_classes1: int = 1000, num_classes2: int = 1000, aux_logits: bool =

evaluation/inception.py:22

↓ 6 callersMethodnormalize

(self, input)

evaluation/image_metrics.py:160

↓ 6 callersMethodstep

Predict the sample from the previous timestep by reversing the SDE. This function propagates the diffusion process from the learned m

src/schedulers/lcm_scheduler.py:32

↓ 5 callersFunctionFeedForward

(dim, mult=4)

ip_adapter/resampler_Instruct.py:81

↓ 5 callersMethod__init__

(self)

src/metrics/lpips.py:79

↓ 4 callersFunctionFeedForward

(dim, mult=4)

ip_adapter/resampler_SD3.py:15

↓ 4 callersMethod__init__

(self, input_dim, hidden_dims, output_dim, dropout_rate=0.1)

ip_adapter/resampler_Instruct.py:17

↓ 4 callersMethod__init__

(self)

src/evaluate/image_metrics.py:68

↓ 4 callersMethod__init__

(self)

evaluation/image_metrics.py:68

↓ 4 callersMethod_get_clip_prompt_embeds

( self, prompt: Union[str, List[str]], num_images_per_prompt: int = 1, device:

ip_adapter/pipeline_stable_diffusion_sd3_extra_cfg.py:267

↓ 4 callersMethodgram_matrix

(self, input)

src/evaluate/image_metrics.py:163

↓ 4 callersMethodgram_matrix

(self, input)

evaluation/image_metrics.py:163

↓ 4 callersFunctionnormalize

(x)

src/evaluate/image_metrics.py:27

↓ 4 callersMethodnormalize

(self, input)

src/evaluate/image_metrics.py:160

↓ 4 callersMethodset_scale

(self, scale)

ip_adapter/ip_adapter_instruct.py:613

↓ 4 callersFunctionspherical_dist_loss

(x, y)

src/renoise_inversion.py:252

↓ 3 callersMethod__init__

( self, hidden_size=None, cross_attention_dim=None, )

ip_adapter/attention_processor.py:16

↓ 3 callersMethodcheck_inputs

( self, prompt, prompt_2, strength, num_inference_steps, callb

pipeline_controlnet_sd_xl_img2img.py:552

↓ 3 callersMethodencode_image

(self, image, device, num_images_per_prompt, output_hidden_states=None)

pipeline_controlnet_sd_xl_img2img.py:765

↓ 3 callersMethodencode_prompt

r""" Encodes the prompt into text encoder hidden states. Args: prompt (`str` or `List[str]`, *optional*):

pipeline_controlnet_sd_xl_img2img.py:300

↓ 3 callersMethodgenerate

( self, pil_image=None, clip_image_embeds=None, prompt=None, negative_

ip_adapter/ip_adapter.py:164

↓ 3 callersFunctionget_activations

Computes the activations of for all images. Args: files (list): List of image file paths. model (torch.nn.Module): Model for comp

evaluation/eval_artfid.py:42

↓ 3 callersFunctionget_canny_map

(input_image_cv2)

infer_style.py:132

↓ 3 callersMethodget_feats

(self, img)

src/evaluate/image_metrics.py:152

↓ 3 callersMethodget_feats

(self, img)

evaluation/image_metrics.py:152

↓ 3 callersMethodget_timesteps

(self, num_inference_steps, strength, device, denoising_start=None)

pipeline_controlnet_sd_xl_img2img.py:642

↓ 3 callersMethodinv_step

Predict the sample from the previous timestep by reversing the SDE. This function propagates the diffusion process from the learned m

src/schedulers/lcm_scheduler.py:126

↓ 3 callersMethodprepare_extra_step_kwargs

(self, generator, eta)

pipeline_controlnet_sd_xl_img2img.py:535

↓ 3 callersMethodprepare_latents

( self, image, timestep, batch_size, num_images_per_prompt, dtype, device, generator=None, add_noise=T

pipeline_controlnet_sd_xl_img2img.py:678

↓ 3 callersFunctionreshape_tensor

(x, heads)

ip_adapter/resampler_Instruct.py:94

↓ 3 callersFunctionreshape_tensor

(x, heads)

ip_adapter/resampler_SD3.py:25

↓ 3 callersFunctionreshape_tensor

(x, heads)

ip_adapter/resampler.py:25

↓ 3 callersFunctionreshape_tensor_multihead

(x, heads)

ip_adapter/resampler_Instruct.py:91

↓ 3 callersMethodset_requires_grad

(self, state: bool)

src/metrics/lpips.py:36

↓ 3 callersMethodset_scale

(self, scale)

ip_adapter/ip_adapter.py:159

↓ 3 callersMethodupcast_vae

(self)

pipeline_controlnet_sd_xl_img2img.py:887

↓ 2 callersMethod__init__

( self, dim=1024, depth=8, dim_head=64, heads=16, num_queries=

ip_adapter/resampler_SD3.py:84

↓ 2 callersMethod__init__

(self, sd_pipe, image_encoder_path, ip_ckpt, device, num_tokens=4)

ip_adapter/ip_adapter.py:72

↓ 2 callersMethod_encode_vae_image

(self, image: torch.Tensor, generator: torch.Generator)

pipeline_controlnet_inpaint_sd_xl.py:1138

↓ 2 callersMethod_forward

(self, x: Tensor, return_features: bool=False)

src/evaluate/inception.py:92

↓ 2 callersMethod_forward

(self, x: Tensor, return_features: bool=False)

evaluation/inception.py:92

↓ 2 callersMethod_get_add_time_ids

( self, original_size, crops_coords_top_left, target_size, aesthetic_s

pipeline_controlnet_sd_xl_img2img.py:835

↓ 2 callersMethod_get_add_time_ids

( self, original_size, crops_coords_top_left, target_size, dtype, text_encoder_projection_dim=None

ip_adapter/pipeline_stable_diffusion_sdxl_extra_cfg.py:730

↓ 2 callersMethod_get_t5_prompt_embeds

( self, prompt: Union[str, List[str]] = None, num_images_per_prompt: int = 1,

ip_adapter/pipeline_stable_diffusion_sd3_extra_cfg.py:211

↓ 2 callersMethodcalc_mean_std

(self, feat, eps=1e-5)

src/evaluate/net.py:88

↓ 2 callersMethodcalc_mean_std

(self, feat, eps=1e-5)

evaluation/net.py:88

↓ 2 callersMethodcheck_image

(self, image, prompt, prompt_embeds)

pipeline_controlnet_inpaint_sd_xl.py:769

↓ 2 callersMethodcheck_image

(self, image, prompt, prompt_embeds)

pipeline_controlnet_sd_xl_img2img_plus.py:851

↓ 2 callersFunctioncompute_activation_statistics

Computes the activation statistics used by the FID. Args: files (list): List of image file paths. model (torch.nn.Module): Mo

evaluation/eval_artfid.py:138

↓ 2 callersFunctioncompute_content_distance

Computes the distance for the given paths. Args: path_to_stylized (str): Path to the stylized images. path_to_style (str): Path t

evaluation/eval_artfid.py:264

↓ 2 callersFunctioncompute_frechet_distance

Numpy implementation of the Frechet Distance. Args: mu1 (np.ndarray): Sample mean of activations of stylized images. mu2 (np.

evaluation/eval_artfid.py:90

↓ 2 callersMethodencode_prompt

r""" Encodes the prompt into text encoder hidden states. Args: prompt (`str` or `List[str]`, *optional*):

ip_adapter/pipeline_stable_diffusion_sdxl_extra_cfg.py:271

↓ 2 callersMethodencode_with_intermediate

(self, input)

src/evaluate/net.py:81

↓ 2 callersMethodencode_with_intermediate

(self, input)

evaluation/net.py:81

↓ 2 callersFunctionfreq_2d

Noise reinitialization. Args: x: diffused latent noise: randomly sampled noise LPF: low pass filter

src/frequency_utils.py:16

↓ 2 callersFunctiongenerate_caption

( image: Image.Image, text: str = None, decoding_method: str = "Nucleus sampling", temperature

infer_style.py:54

↓ 2 callersFunctionget_depth_map

(image)

infer_style.py:110

↓ 2 callersMethodget_feats

(self, img)

src/evaluate/image_metrics.py:100

↓ 2 callersMethodget_feats

(self, img)

evaluation/image_metrics.py:100

↓ 2 callersFunctionget_freq_filter

Form the frequency filter for noise reinitialization. Args: shape: shape of latent (B, C, T, H, W) filter_type: type of the

src/frequency_utils.py:190

↓ 2 callersFunctionget_generator

(seed, device)

ip_adapter/utils.py:86

↓ 2 callersMethodget_guidance_scale_embedding

See https://github.com/google-research/vdm/blob/dc27b98a554f65cdc654b800da5aa1846545d41b/model_vdm.py#L298 Args: w (`tor

pipeline_controlnet_sd_xl_img2img.py:905

↓ 2 callersMethodget_image_embeds

(self, pil_image=None, clip_image_embeds=None)

ip_adapter/ip_adapter_instruct.py:592

↓ 2 callersMethodget_image_embeds

(self, pil_image=None, clip_image_embeds=None,dtype=torch.float32)

ip_adapter/ip_adapter.py:147

↓ 2 callersMethodget_single_image_embeds

(self, pil_image=None,clip_image=None,clip_image_embeds=None,instruct_embeds=None,prompt_embeds=None)

ip_adapter/ip_adapter_instruct.py:114

↓ 2 callersFunctioninversion_step

( pipe, z_t: torch.tensor, t: torch.tensor, prompt_embeds, added_cond_kwargs, num_reno

src/renoise_inversion.py:122

↓ 2 callersMethodload_ip_adapter

(self)

ip_adapter/ip_adapter.py:130

↓ 2 callersFunctionnoise_regularization

( e_t, noise_pred_optimal, lambda_kl, lambda_ac, num_reg_steps, num_ac_rolls, generator=None )

src/renoise_inversion.py:13

↓ 2 callersFunctionnormalize

(x)

evaluation/image_metrics.py:27

↓ 2 callersMethodnormalize

(self, input)

src/evaluate/image_metrics.py:108

↓ 2 callersMethodnormalize

(self, input)

evaluation/image_metrics.py:108

↓ 2 callersFunctionpatchify_tensor

(input_tensor)

src/renoise_inversion.py:64

↓ 2 callersMethodprepare_control_image

( self, image, width, height, batch_size, num_images_per_promp

pipeline_controlnet_inpaint_sd_xl.py:1023

↓ 2 callersMethodprepare_control_image

( self, image, width, height, batch_size, num_images_per_promp

pipeline_controlnet_sd_xl_img2img_plus.py:889

↓ 2 callersFunctionreplace_transformer_blocks

(original_model)

ip_adapter/ip_adapter_instruct.py:475

↓ 2 callersFunctionretrieve_latents

( encoder_output: torch.Tensor, generator: Optional[torch.Generator] = None, sample_mode: str = "sample" )

pipeline_controlnet_inpaint_sd_xl.py:80

↓ 2 callersFunctionretrieve_latents

( encoder_output: torch.Tensor, generator: Optional[torch.Generator] = None, sample_mode: str = "sample" )

pipeline_controlnet_sd_xl_img2img_plus.py:159

↓ 2 callersFunctionretrieve_latents

( encoder_output: torch.Tensor, generator: Optional[torch.Generator] = None, sample_mode: str = "sample" )

pipeline_controlnet_sd_xl_img2img.py:111

↓ 2 callersFunctionretrieve_timesteps

Calls the scheduler's `set_timesteps` method and retrieves timesteps from the scheduler after the call. Handles custom timesteps. Any kwargs

pipeline_controlnet_sd_xl_img2img.py:125

↓ 2 callersFunctionretrieve_timesteps

Calls the scheduler's `set_timesteps` method and retrieves timesteps from the scheduler after the call. Handles custom timesteps. Any kwargs

ip_adapter/pipeline_stable_diffusion_extra_cfg.py:76

↓ 2 callersMethodset_ip_adapter

(self, n=2)

ip_adapter/ip_adapter_instruct.py:530

↓ 2 callersMethodset_noise_list

(self, noise_list)

src/schedulers/lcm_scheduler.py:29

↓ 2 callersFunctionspherical_dist_loss

(x, y)

pipeline_controlnet_sd_xl_img2img_plus.py:77

↓ 2 callersFunctionunet_pass

(pipe, z_t, t, prompt_embeds, added_cond_kwargs)

src/renoise_inversion.py:376

↓ 1 callersFunctionFeedForward

(dim, mult=4)

ip_adapter/resampler.py:15

↓ 1 callersMethod__init__

(self,num_tokens)

ip_adapter/ip_joint_attention.py:155

↓ 1 callersMethod__init__

( self, dim=1024, depth=8, dim_head=64, heads=16, num_queries=

ip_adapter/resampler.py:84

↓ 1 callersMethod__init__

(self, gpu_ids=[])

src/evaluate/net.py:118

↓ 1 callersMethod__init__

(self, gpu_ids=[])

evaluation/net.py:118

↓ 1 callersMethod_forward

(self, x: Tensor)

src/evaluate/inception.py:190

↓ 1 callersMethod_forward

(self, x: Tensor)

src/evaluate/inception.py:222

↓ 1 callersMethod_forward

(self, x: Tensor)

src/evaluate/inception.py:261

↓ 1 callersMethod_forward

(self, x: Tensor)

src/evaluate/inception.py:298

↓ 1 callersMethod_forward

(self, x: Tensor)

src/evaluate/inception.py:334

↓ 1 callersMethod_forward

(self, x: Tensor)

evaluation/inception.py:190

next →1–100 of 440, ranked by callers