Functions — github.com/bytedance/Shot2Story

hub / github.com/bytedance/Shot2Story / functions

Functions1,483 in github.com/bytedance/Shot2Story

⨍ Functions 1,483◇ Types & classes 348↳ Endpoints 10

↓ 4 callersFunctionenhance_level_to_args

(MAX_LEVEL)

code/lavis/processors/randaugment.py:223

↓ 4 callersFunctiongetAttMap

(img, attMap, blur=True, overlap=True)

code/app/utils.py:49

↓ 4 callersMethodget_attention_mask

(self, seq, pad_token=-1)

code/lavis/processors/gpt_processors.py:85

↓ 4 callersMethodget_builder_class

(cls, name)

code/lavis/common/registry.py:232

↓ 4 callersFunctionget_padding

(kernel_size: int, stride: int = 1, dilation: int = 1, **_)

code/lavis/models/timesformer/vit_utils.py:100

↓ 4 callersMethodget_processor_class

(cls, name)

code/lavis/common/registry.py:244

↓ 4 callersMethodget_prompt

(self)

code/lavis/conversation/conversation.py:52

↓ 4 callersFunctionload_blip_itm_model

(device, model_type="base")

code/app/utils.py:77

↓ 4 callersFunctionload_demo_image

()

code/app/__init__.py:16

↓ 4 callersMethodload_from_pretrained

(self, url_or_filename)

code/lavis/models/blip_models/blip.py:34

↓ 4 callersFunctionload_state_dict

(checkpoint_path, use_ema=False)

code/lavis/models/timesformer/helpers.py:24

↓ 4 callersFunctionload_video

(video_path, n_frms=MAX_INT, height=-1, width=-1, sampling="uniform", clips=None)

code/lavis/datasets/data_utils.py:32

↓ 4 callersMethodlog_stats

(self, stats, split_name)

code/lavis/runners/runner_base.py:713

↓ 4 callersMethodmodel

A property to get the DDP-wrapped model on the device.

code/lavis/runners/runner_base.py:88

↓ 4 callersFunctionprepare_sample

(samples, cuda_enabled=True)

code/lavis/datasets/data_utils.py:140

↓ 4 callersMethodprocessPunctuation

(self, inText)

code/lavis/common/vqa_tools/vqa_eval.py:249

↓ 4 callersMethodsetup_task

(cls, cfg)

code/lavis/tasks/vqa.py:48

↓ 4 callersFunctiontile

(x, dim, n_tile)

code/lavis/models/base_model.py:242

↓ 3 callersMethod__init__

vis_root (string): Root directory of images (e.g. coco/images/) ann_root (string): directory to store the annotation file

code/lavis/datasets/datasets/retrieval_datasets.py:30

↓ 3 callersMethod__init__

( self, in_features, hidden_features=None, out_features=None, act_laye

code/lavis/models/vit.py:29

↓ 3 callersMethod__init__

(self, feature_info: List[Dict], out_indices: Tuple[int])

code/lavis/models/timesformer/features.py:22

↓ 3 callersMethod_convert_node_to_json

(self, node)

code/lavis/common/config.py:161

↓ 3 callersMethod_save_checkpoint

Save the checkpoint at the current epoch.

code/lavis/runners/runner_base.py:632

↓ 3 callersFunctionapply_rot_embed

(x: torch.Tensor, sin_emb, cos_emb)

code/lavis/models/clip_models/timm_model.py:434

↓ 3 callersMethodattention

(self, x: torch.Tensor, attn_mask: Optional[torch.Tensor] = None)

code/lavis/models/clip_models/model.py:279

↓ 3 callersMethodbuild_datasets

(self, cfg)

code/lavis/tasks/vqa.py:71

↓ 3 callersMethodchunk

(iterable, chunk_size, drop_last_batch)

code/lavis/datasets/datasets/hdfs_distributed_sampler.py:31

↓ 3 callersMethodeval_epoch

Evaluate the model on a given split. Args: split_name (str): name of the split to evaluate on. cur_epoch (in

code/lavis/runners/runner_base.py:482

↓ 3 callersMethodforward_cap

Args: samples (dict): A dictionary containing the following keys: - image (torch.Tensor): A tensor of shape (batc

code/lavis/models/pnp_vqa_models/pnp_vqa.py:84

↓ 3 callersMethodforward_itm

Args: samples (dict): A dictionary containing the following keys: - image (torch.Tensor): A tensor of shape (batc

code/lavis/models/pnp_vqa_models/pnp_vqa.py:54

↓ 3 callersMethodforward_qa

Args: samples (dict): A dictionary containing the following keys: - image (torch.Tensor): A tensor of shape (batc

code/lavis/models/pnp_vqa_models/pnp_vqa.py:174

↓ 3 callersFunctionfreeze_batch_norm_2d

Converts all `BatchNorm2d` and `SyncBatchNorm` layers of provided module into `FrozenBatchNorm2d`. If `module` is itself an instance of eithe

code/lavis/models/clip_models/utils.py:14

↓ 3 callersMethodget_extended_attention_mask

Makes broadcastable attention and causal masks so that future and masked tokens are ignored. Arguments: attention_mask (

code/lavis/models/blip2_models/Qformer.py:713

↓ 3 callersFunctionget_pretrained_url

(model: str, tag: str)

code/lavis/models/clip_models/pretrained.py:120

↓ 3 callersFunctionget_world_size

()

code/lavis/common/dist_utils.py:41

↓ 3 callersFunctionhexists

hdfs capable to check whether a file_path is exists

code/lavis/datasets/hdfs_utils.py:237

↓ 3 callersFunctionhload_pkl

load pickle from hdfs

code/lavis/datasets/hdfs_utils.py:29

↓ 3 callersFunctioninit_bert_tokenizer

()

code/app/utils.py:44

↓ 3 callersMethodloadRes

Load result file and return a result object. :param resFile (str) : file name of result file :return: res (obj)

code/lavis/common/vqa_tools/vqa.py:173

↓ 3 callersFunctionload_dataset_config

(cfg_path)

code/lavis/datasets/builders/base_dataset_builder.py:310

↓ 3 callersFunctionmakedir

Create the directory if it does not exist.

code/lavis/common/utils.py:64

↓ 3 callersFunctionpixel_freq_bands

( num_bands: int, max_freq: float = 224.0, linear_bands: bool = True, dtype: torch.dtype = tor

code/lavis/models/clip_models/timm_model.py:250

↓ 3 callersMethodprocessDigitArticle

(self, inText)

code/lavis/common/vqa_tools/vqa_eval.py:261

↓ 3 callersMethodregister

r"""Register an item to registry with key 'name' Args: name: Key with which the item will be registered. Usage::

code/lavis/common/registry.py:205

↓ 3 callersFunctionreorg_datasets_by_split

Organizes datasets by split. Args: datasets: dict of torch.utils.data.Dataset objects by name. Returns: Dict of dataset

code/lavis/datasets/data_utils.py:149

↓ 3 callersMethodreset_classifier

(self, num_classes, global_pool="")

code/lavis/models/timesformer/vit.py:401

↓ 3 callersFunctionrot

(x)

code/lavis/models/clip_models/timm_model.py:430

↓ 3 callersFunctionshow_samples

(dataset, offset=0, is_next=False)

code/app/dataset_browser.py:94

↓ 3 callersMethodstep

(self, cur_epoch, cur_step)

code/lavis/common/optims.py:37

↓ 3 callersMethodto_dict

(self)

code/lavis/common/config.py:165

↓ 3 callersFunctiontrunc_normal_

r"""Fills the input Tensor with values drawn from a truncated normal distribution. The values are effectively drawn from the normal distributi

code/lavis/models/clip_models/timm_model.py:544

↓ 2 callersMethod__init__

(self)

code/lavis/processors/alpro_processors.py:52

↓ 2 callersMethod_build_class_labels

(self)

code/lavis/datasets/datasets/nlvr_datasets.py:40

↓ 2 callersMethod_collect

(self, x)

code/lavis/models/timesformer/features.py:207

↓ 2 callersFunction_file_name

(row)

code/lavis/datasets/download_scripts/DownloadConceptualCaptions/download_data_cc3m.py:82

↓ 2 callersFunction_file_name

(row)

code/lavis/datasets/download_scripts/DownloadConceptualCaptions/download_data_cc12m.py:82

↓ 2 callersFunction_get_feature_info

(net, out_indices)

code/lavis/models/timesformer/features.py:137

↓ 2 callersMethod_load_checkpoint

Resume from a checkpoint.

code/lavis/runners/runner_base.py:680

↓ 2 callersFunction_module_list

(module, flatten_sequential=False)

code/lavis/models/timesformer/features.py:123

↓ 2 callersMethod_predict_class

Args: samples (dict): A dictionary containing the following keys: - image (torch.Tensor): A tensor of shape (batc

code/lavis/models/blip2_models/blip2_t5_instruct.py:503

↓ 2 callersMethod_predict_class

( self, samples, candidates, n_segments=1, )

code/lavis/models/blip2_models/blip2_vicuna_instruct.py:469

↓ 2 callersFunction_rescan_model_configs

()

code/lavis/models/clip_models/model.py:896

↓ 2 callersMethod_save_checkpoint

(self, cur_iters, is_best=False)

code/lavis/runners/runner_iter.py:156

↓ 2 callersMethod_save_result_leaderboard

Saving the results in the format required for leaderboard evaluation. [TODO] add support for multi-choice.

code/lavis/tasks/vqa.py:296

↓ 2 callersMethod_shift_right

(self, input_ids)

code/lavis/models/blip2_models/modeling_t5.py:919

↓ 2 callersMethod_train_inner_loop

An inner training loop compatible with both epoch-based and iter-based training. When using epoch-based, training stops after one ep

code/lavis/tasks/base_task.py:156

↓ 2 callersFunction_urlretrieve

(url: str, filename: str, chunk_size: int = 1024)

code/lavis/common/utils.py:154

↓ 2 callersMethodadd_meter

(self, name, meter)

code/lavis/common/logger.py:119

↓ 2 callersMethodask

(self, text, conv)

code/lavis/conversation/conversation.py:502

↓ 2 callersMethodattention

(self, x: torch.Tensor)

code/lavis/models/clip_vit.py:132

↓ 2 callersFunctionbuild_fourier_pos_embed

( feat_shape: List[int], bands: Optional[torch.Tensor] = None, num_bands: int = 64, max_res: i

code/lavis/models/clip_models/timm_model.py:329

↓ 2 callersMethodbuild_model

(self, cfg)

code/lavis/tasks/base_task.py:29

↓ 2 callersFunctionbuild_model_from_openai_state_dict

(state_dict: dict)

code/lavis/models/clip_models/model.py:792

↓ 2 callersFunctionbytes_to_unicode

Returns list of utf-8 byte and a corresponding list of unicode strings. The reversible bpe codes work on unicode strings. This means you

code/lavis/models/clip_models/tokenizer.py:32

↓ 2 callersFunctioncombine_splits

(s1,s2)

code/lavis/conversation/conversation.py:239

↓ 2 callersMethodcompute_itm

(self, image_inputs, text_ids, text_atts)

code/lavis/models/blip2_models/blip2_qformer.py:347

↓ 2 callersFunctioncompute_sim_matrix

(model, data_loader, **kwargs)

code/lavis/models/albef_models/__init__.py:76

↓ 2 callersFunctionconcat_datasets

Concatenates multiple datasets into a single dataset. It supports may-style datasets and DataPipeline from WebDataset. Currently, does not s

code/lavis/datasets/data_utils.py:175

↓ 2 callersMethodconcat_text_input_output

(self, input_ids, input_atts, output_ids, output_atts)

code/lavis/models/blip2_models/blip2_vicuna_instruct.py:115

↓ 2 callersFunctionconvert_weights_to_fp16

Convert applicable model parameters to fp16

code/lavis/models/eva_vit.py:397

↓ 2 callersFunctionconvert_weights_to_fp16

Convert applicable model parameters to fp16

code/lavis/models/clip_models/model.py:763

↓ 2 callersMethodcreateIndex

(self)

code/lavis/common/vqa_tools/vqa.py:53

↓ 2 callersMethodcreate_loaders

Create dataloaders for training and validation.

code/lavis/runners/runner_base.py:523

↓ 2 callersFunctioncreate_model

( model_name: str, pretrained: str = "", precision: str = "fp32", device: torch.device = torch

code/lavis/models/clip_models/model.py:934

↓ 2 callersFunctioncrop

Args: clip (torch.tensor): Video clip to be cropped. Size is (C, T, H, W)

code/lavis/processors/functional_video.py:23

↓ 2 callersFunctiondownload_pretrained

(url: str, root: str = os.path.expanduser("~/.cache/clip"))

code/lavis/models/clip_models/pretrained.py:130

↓ 2 callersFunctiondrop_path

Drop paths (Stochastic Depth) per sample (when applied in main path of residual blocks). This is the same as the DropConnect impl I created for Ef

code/lavis/models/timesformer/vit_utils.py:161

↓ 2 callersMethodencode_img

(self, image)

code/lavis/models/blip2_models/video_minigpt4.py:189

↓ 2 callersMethodformat_arguments

(self)

code/lavis/common/config.py:248

↓ 2 callersMethodformat_help

(self)

code/lavis/common/config.py:251

↓ 2 callersMethodforward

( self, hidden_states, attention_mask=None, head_mask=None, encoder_hi

code/lavis/models/med.py:422

↓ 2 callersMethodforward_encoder

(self, samples)

code/lavis/models/blip_models/blip_caption_video.py:53

↓ 2 callersMethodforward_encoder

(self, samples)

code/lavis/models/blip_models/blip_caption.py:53

↓ 2 callersMethodforward_encoder

(self, samples)

code/lavis/models/albef_models/albef_vqa.py:137

↓ 2 callersMethodfrom_config

(cls, cfg=None)

code/lavis/processors/base_processor.py:20

↓ 2 callersMethodgenerate

(self, input_ids, attention_mask, num_beams=1, min_length=0, max_length=20)

code/lavis/models/pnp_vqa_models/pnp_unifiedqav2_fid.py:43

↓ 2 callersMethodgenerate_from_encoder

( self, tokenized_prompt, visual_embeds, sep_token_id, pad_token_id,

code/lavis/models/med.py:1316

↓ 2 callersMethodgetQuesIds

Get question ids that satisfy given filter conditions. default skips that filter :param imgIds (int array) : get question ids f

code/lavis/common/vqa_tools/vqa.py:79

↓ 2 callersMethodget_dicts

return info dicts for specified keys (or all if None) at specified indices (or out_indices if None)

code/lavis/models/timesformer/features.py:49

↓ 2 callersMethodget_embed

(self, shape: List[int])

code/lavis/models/clip_models/timm_model.py:497

← previous next →101–200 of 1,483, ranked by callers