Functions — github.com/WillDreamer/Aurora

CodeHub MCPcopy Create free account

hub / github.com/WillDreamer/Aurora / functions

Functions226 in github.com/WillDreamer/Aurora

⨍ Functions 226◇ Types & classes 41

↓ 55 callersFunctionprint

(*args, **kwargs)

↓ 26 callersFunction_n2p

↓ 13 callersMethod__init__

(self, config, layer_num, adapter_config=None, adapter=None,R=None)

↓ 9 callersFunctionconcat_all_gather

Performs all_gather operation on the provided tensors. *** Warning ***: torch.distributed.all_gather has no gradient.

CP/blip_retrieval.py:352

↓ 9 callersMethodtranspose_for_scores

↓ 8 callersMethodadd_meter

(self, name, meter)

↓ 8 callersMethodlog_every

(self, iterable, print_freq, header=None)

↓ 8 callersFunctionpre_caption

(caption,max_words=50)

data/utils.py:21

↓ 8 callersMethodupdate

(self, **kwargs)

↓ 7 callersMethodmax

↓ 5 callersFunctioncreate_vit

(vit, image_size, use_grad_checkpointing=False, ckpt_layer=0, drop_path_rate=0, config=None, adapters=None, de

↓ 4 callersFunctionenhance_level_to_args

transform/randaugment.py:202

↓ 4 callersFunctioninit_tokenizer

↓ 4 callersFunctionload_checkpoint

(model, url_or_filename)

↓ 3 callersMethod__init__

(self, in_features, hidden_features=None, out_features=None, act_layer=nn.GELU, drop=0., R = 64)

↓ 3 callersMethodbackward

CP/blip_retrieval.py:378

↓ 3 callersFunctioncosine_lr_schedule

Decay the learning rate

↓ 3 callersFunctioncreate_dataset

(dataset, config, min_scale=0.5)

data/__init__.py:15

↓ 3 callersFunctioncreate_loader

(datasets, samplers, batch_size, num_workers, is_trains, collate_fns)

data/__init__.py:77

↓ 3 callersFunctioncreate_sampler

(datasets, shuffles, num_tasks, global_rank)

data/__init__.py:69

↓ 3 callersMethodglobal_avg

↓ 3 callersFunctionis_dist_avail_and_initialized

↓ 3 callersMethodsynchronize_between_processes

↓ 3 callersMethodvalue

↓ 2 callersFunctionblip_retrieval

(pretrained='',**kwargs)

CP/blip_retrieval.py:342

↓ 2 callersFunctioncalculate_score

(json_path='', target_path='')

train_vqa.py:30

↓ 2 callersFunctionevaluation

(model, data_loader, device, config)

train_vqa.py:88

↓ 2 callersFunctionextract_frames_from_video_binary

Args: in_mem_bytes_io: binary from read file object >>> with open(video_path, "rb") as f: >>> input_bytes = f

data/video_dataset_didemo.py:103

↓ 2 callersFunctionflat_list_of_lists

flatten a list of lists [[1,2], [3,4]] to [1,2,3,4]

data/video_dataset_didemo.py:167

↓ 2 callersFunctionget_start_end_idx

Sample a clip of size clip_size from a video of size video_size and return the indices of the first and last frame of the clip. If clip_idx i

data/utils.py:146

↓ 2 callersMethodimgfea2videofea

CP/blip_retrieval.py:85

↓ 2 callersFunctioninterpolate_pos_embed

(pos_embed_checkpoint, visual_encoder)

↓ 2 callersFunctionload_jsonl

data/video_dataset_didemo.py:47

↓ 2 callersFunctionload_jsonl

data/video_dataset_msrvtt.py:28

↓ 2 callersFunctionmk_input_group

Re-organize examples into groups. Each input group will have a single image paired with X (X=max_n_example_per_img) examples. Images with total #

data/video_dataset_didemo.py:198

↓ 2 callersFunctionmk_video_ret_datalist

Args: raw_datalist: list(dict) cfg: Returns:

data/video_dataset_didemo.py:144

↓ 2 callersFunctionpre_question

(question,max_ques_words=50)

data/utils.py:42

↓ 2 callersFunctionpyav_decode_stream

Decode the video with PyAV decoder. Args: container (container): PyAV container. start_pts (int): the starting Presentation T

data/utils.py:178

↓ 2 callersFunctionsave_result

(result, result_dir, filename, remove_duplicate='')

data/utils.py:58

↓ 2 callersFunctionshear_level_to_args

(MAX_LEVEL, replace_value)

transform/randaugment.py:208

↓ 2 callersFunctiontile

(x, dim, n_tile)

CP/blip_vqa.py:196

↓ 2 callersFunctiontranslate_level_to_args

(translate_const, MAX_LEVEL, replace_value)

transform/randaugment.py:217

↓ 2 callersFunctiontune_channel

transform/randaugment.py:16

↓ 2 callersMethodupdate

(self, value, n=1)

↓ 1 callersMethod__init__

Args: med_config (str): path for the mixture of encoder-decoder model's configuration file image_size (int): input im

↓ 1 callersMethod_dequeue_and_enqueue

(self, image_feat, text_feat, idxs)

CP/blip_retrieval.py:323

↓ 1 callersMethod_load_video_from_path_decord

(self, video_path, height=None, width=None, start_time=None, end_time=None, fps=-1)

data/videovqa_dataset.py:69

↓ 1 callersMethod_load_video_from_path_decord

(self, video_path, height=None, width=None, start_time=None, end_time=None, fps=-1)

data/video_dataset_msrvtt.py:77

↓ 1 callersMethod_load_video_from_path_decord

(self, video_path, height=None, width=None, start_time=None, end_time=None, fps=-1)

data/video_dataset_msrvtt.py:157

↓ 1 callersFunction_load_weights

(model: VisionTransformer, checkpoint_path: str, prefix: str = '')

↓ 1 callersMethod_momentum_update

CP/blip_retrieval.py:316

↓ 1 callersFunctionall_gather_with_grad

Performs all_gather operation on the provided tensors. Graph remains connected for backward grad computation.

CP/blip_retrieval.py:384

↓ 1 callersFunctionblip_vqa

(pretrained='',**kwargs)

CP/blip_vqa.py:188

↓ 1 callersFunctionchunk_list

Args: examples: iterable, examples grouped by image/video chunk_size: int, number of examples in each chunk. pad_to_divis

data/video_dataset_didemo.py:171

↓ 1 callersMethodcopy_params

CP/blip_retrieval.py:308

↓ 1 callersFunctiondecode

Decode the video and perform temporal sampling. Args: container (container): pyav container. sampling_rate (int): frame sampl

data/utils.py:317

↓ 1 callersFunctionevaluation

(model, data_loader, device, config)

train_video_retrieval.py:66

↓ 1 callersFunctionevaluation

(model, data_loader, device, config)

train_retrieval.py:63

↓ 1 callersMethodgenerate

(self, image, sample=False, num_beams=3, max_length=30, min_length=10, top_p=0.9, repetition_penalty=1.0)

↓ 1 callersMethodget_extended_attention_mask

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

↓ 1 callersMethodget_random_ops

transform/randaugment.py:315

↓ 1 callersFunctionget_rank

↓ 1 callersFunctionget_video_decoding_kwargs

(container, num_frames, target_fps, num_clips=None, clip_idx=None,

data/video_dataset_didemo.py:52

↓ 1 callersFunctionis_main_process

↓ 1 callersFunctionis_url

(url_or_filename)

↓ 1 callersFunctionitm_eval

(scores_v2t, scores_t2v, txt2vmg, vid2txt)

train_video_retrieval.py:178

↓ 1 callersFunctionitm_eval

(scores_i2t, scores_t2i, txt2img, img2txt)

train_retrieval.py:169

↓ 1 callersFunctionmain

train_vqa.py:123

↓ 1 callersFunctionmain

train_video_retrieval.py:222

↓ 1 callersFunctionmain

train_retrieval.py:215

↓ 1 callersMethodmedian

↓ 1 callersFunctionposterize_level_to_args

transform/randaugment.py:245

↓ 1 callersMethodprune_heads

↓ 1 callersFunctionpyav_decode

Convert the video from its original fps to the target_fps. If the video support selective decoding (contain decoding information in the video

data/utils.py:220

↓ 1 callersMethodrank_answer

(self, question_states, question_atts, answer_ids, answer_atts, k, CP_U, CP_V, CP_C)

CP/blip_vqa.py:137

↓ 1 callersFunctionrotate_level_to_args

(MAX_LEVEL, replace_value)

transform/randaugment.py:252

↓ 1 callersMethodsave_attention_map

(self, attention_map)

↓ 1 callersFunctionsetup_for_distributed

This function disables printing when not in master process

↓ 1 callersFunctionsolarize_level_to_args

transform/randaugment.py:234

↓ 1 callersFunctiontemporal_sampling

Given the start and end frame index, sample num_samples frames between the start and end with equal interval. Args: frames (list(

data/utils.py:126

↓ 1 callersFunctiontrain

(model, data_loader, optimizer, epoch, device)

train_vqa.py:57

↓ 1 callersFunctiontrain

(model, data_loader, optimizer, epoch, device, config)

train_video_retrieval.py:28

↓ 1 callersFunctiontrain

(model, data_loader, optimizer, epoch, device, config)

train_retrieval.py:29

data/videovqa_dataset.py:22

data/video_dataset_didemo.py:39

data/video_dataset_msrvtt.py:22

transform/randaugment.py:319

Method__getattr__

Method__getitem__

data/flickr30k_dataset.py:39

Method__getitem__

data/flickr30k_dataset.py:86

Method__getitem__

data/videovqa_dataset.py:50

Method__getitem__

data/coco_karpathy_dataset.py:36

Method__getitem__

data/coco_karpathy_dataset.py:85

Method__getitem__

data/video_dataset_didemo.py:287

Method__getitem__

data/video_dataset_didemo.py:350

Method__getitem__

data/vqa_dataset.py:43

Method__getitem__

data/video_dataset_msrvtt.py:68

Method__getitem__

data/video_dataset_msrvtt.py:143

(self, window_size=20, fmt=None)

(self, delimiter="\t")

next →1–100 of 226, ranked by callers