Functions — github.com/QwenLM/Qwen3-VL-Embedding

hub / github.com/QwenLM/Qwen3-VL-Embedding / functions

Functions211 in github.com/QwenLM/Qwen3-VL-Embedding

⨍ Functions 211◇ Types & classes 22

↓ 1 callersFunctionload_score

Load score results for a single task

src/evaluation/mmeb_v2/gather_results.py:92

↓ 1 callersFunctionload_topk_from_pred

Load embedding retrieval TopK from *_pred.jsonl written by eval_embedding.

src/evaluation/mmeb_v2/eval_reranker.py:56

↓ 1 callersFunctionload_video_class_dataset

(model_args, data_args, **kwargs)

src/evaluation/mmeb_v2/data/datasets/video_classification_datasets.py:63

↓ 1 callersFunctionload_vidore_dataset

(model_args, data_args, **kwargs)

src/evaluation/mmeb_v2/data/datasets/vidore_dataset.py:92

↓ 1 callersFunctionmain

()

main.py:1

↓ 1 callersFunctionmain

()

src/evaluation/mmeb_v2/constant.py:160

↓ 1 callersFunctionmain

()

src/evaluation/mmeb_v2/eval_reranker.py:165

↓ 1 callersFunctionmain

()

src/evaluation/mmeb_v2/eval_embedding.py:107

↓ 1 callersFunctionmain

()

src/evaluation/mmeb_v2/gather_results.py:255

↓ 1 callersFunctionnormalize

Args: clip (torch.tensor): Video clip to be normalized. Size is (T, C, H, W) mean (tuple): pixel RGB mean. Size is (3) st

src/evaluation/mmeb_v2/utils/vision_utils/video_transforms.py:368

↓ 1 callersFunctionpad_dataset_to_divisible

(dataset, world_size)

src/evaluation/mmeb_v2/eval_reranker.py:30

↓ 1 callersFunctionrandom_shift_crop

Slide along the long edge, with the short edge as crop size

src/evaluation/mmeb_v2/utils/vision_utils/video_transforms.py:332

↓ 1 callersFunctionread_image_from_path

(path, transform=None, transform_name="center", num_frames=1, image_size=(256, 256))

src/evaluation/mmeb_v2/utils/vision_utils/video_transforms.py:113

↓ 1 callersFunctionread_image_from_path

(path, transform=None, transform_name="center", num_frames=1, image_size=(256, 256))

src/evaluation/mmeb_v2/utils/vision_utils/vision_utils.py:137

↓ 1 callersFunctionread_video_from_path

(path, transform=None, transform_name="center", image_size=(256, 256))

src/evaluation/mmeb_v2/utils/vision_utils/video_transforms.py:123

↓ 1 callersFunctionread_video_from_path

(path, transform=None, transform_name="center", image_size=(256, 256))

src/evaluation/mmeb_v2/utils/vision_utils/vision_utils.py:147

↓ 1 callersFunctionrerank_topk_for_queries

For each query, rerank its retrieved TopK candidates by Qwen3VLReranker. Return pred_dicts compatible with RankingMetrics.

src/evaluation/mmeb_v2/eval_reranker.py:69

↓ 1 callersFunctionresize_crop_to_fill

(pil_image, image_size)

src/evaluation/mmeb_v2/utils/vision_utils/vision_utils.py:207

↓ 1 callersFunctionresize_scale

(clip, target_size, interpolation_mode)

src/evaluation/mmeb_v2/utils/vision_utils/video_transforms.py:256

↓ 1 callersFunctionsample_frames

(frames: List[Union[str, Image.Image]], max_segments: int)

src/models/qwen3_vl_embedding.py:119

↓ 1 callersFunctionsample_frames

( frames: List[Union[str, Image.Image]], max_segments: int )

src/models/qwen3_vl_reranker.py:60

↓ 1 callersFunctionsample_frames

(frames, num_segments)

src/evaluation/mmeb_v2/utils/vision_utils/vision_utils.py:254

↓ 1 callersFunctionto_tensor

Convert tensor data type from uint8 to float, divide value by 255.0 and permute the dimensions of clip tensor Args: clip (torch.t

src/evaluation/mmeb_v2/utils/vision_utils/video_transforms.py:352

↓ 1 callersMethodtokenize

(self, pairs: List[Dict], **kwargs)

src/models/qwen3_vl_reranker.py:166

↓ 1 callersMethodtruncate_tokens_optimized

( self, tokens: List[str], max_length: int, special_tokens: List[str] )

src/models/qwen3_vl_reranker.py:139

Method__call__

(self, clip)

src/evaluation/mmeb_v2/utils/vision_utils/video_transforms.py:407

Method__call__

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

src/evaluation/mmeb_v2/utils/vision_utils/video_transforms.py:422

Method__call__

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

src/evaluation/mmeb_v2/utils/vision_utils/video_transforms.py:472

Method__call__

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

src/evaluation/mmeb_v2/utils/vision_utils/video_transforms.py:510

Method__call__

(self, clip)

src/evaluation/mmeb_v2/utils/vision_utils/video_transforms.py:545

Method__call__

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

src/evaluation/mmeb_v2/utils/vision_utils/video_transforms.py:566

Method__call__

Args: clip (torch.tensor): video clip must be normalized. Size is (C, T, H, W)

src/evaluation/mmeb_v2/utils/vision_utils/video_transforms.py:595

Method__call__

Args: clip (torch.tensor, dtype=torch.uint8): Size is (T, C, H, W) Return: clip (torch.tensor, dtype=torch.fl

src/evaluation/mmeb_v2/utils/vision_utils/video_transforms.py:615

Method__call__

Args: clip (torch.tensor): Size is (T, C, H, W) Return: clip (torch.tensor): Size is (T, C, H, W)

src/evaluation/mmeb_v2/utils/vision_utils/video_transforms.py:638

Method__call__

(self, total_frames)

src/evaluation/mmeb_v2/utils/vision_utils/video_transforms.py:666

Method__call__

(self, examples)

src/evaluation/mmeb_v2/data/collator.py:10

Method__init__

(self, config)

src/models/qwen3_vl_embedding.py:47

Method__init__

( self, model_name_or_path: str, max_length: int = MAX_LENGTH, min_pixels: int

src/models/qwen3_vl_reranker.py:76

Method__init__

(self, encoder: Qwen3VLEmbedder, normalize: bool = True, te

src/evaluation/mmeb_v2/models.py:13

Method__init__

(self, size)

src/evaluation/mmeb_v2/utils/vision_utils/video_transforms.py:401

Method__init__

(self, size)

src/evaluation/mmeb_v2/utils/vision_utils/video_transforms.py:416

Method__init__

( self, size, interpolation_mode="bilinear", )

src/evaluation/mmeb_v2/utils/vision_utils/video_transforms.py:458

Method__init__

( self, size, interpolation_mode="bilinear", )

src/evaluation/mmeb_v2/utils/vision_utils/video_transforms.py:496

Method__init__

( self, size, interpolation_mode="bilinear", )

src/evaluation/mmeb_v2/utils/vision_utils/video_transforms.py:531

Method__init__

( self, size, interpolation_mode="bilinear", )

src/evaluation/mmeb_v2/utils/vision_utils/video_transforms.py:552

Method__init__

(self, mean, std, inplace=False)

src/evaluation/mmeb_v2/utils/vision_utils/video_transforms.py:590

Method__init__

(self)

src/evaluation/mmeb_v2/utils/vision_utils/video_transforms.py:612

Method__init__

(self, p=0.5)

src/evaluation/mmeb_v2/utils/vision_utils/video_transforms.py:635

Method__init__

(self, size)

src/evaluation/mmeb_v2/utils/vision_utils/video_transforms.py:663

Method__init__

Initialize retrieval metrics. Args: metric_list (tuple or list): Metrics to compute ("precision", "recall", "ndcg", "map

src/evaluation/mmeb_v2/utils/eval_utils/metrics.py:12

Method__init__

(self, *args, **kwargs)

src/evaluation/mmeb_v2/data/datasets/base_eval_dataset.py:40

Method__init_subclass__

(cls)

src/evaluation/mmeb_v2/data/datasets/base_eval_dataset.py:34

Method__repr__

(self)

src/evaluation/mmeb_v2/utils/vision_utils/video_transforms.py:411

Method__repr__

(self)

src/evaluation/mmeb_v2/utils/vision_utils/video_transforms.py:448

Method__repr__

(self)

src/evaluation/mmeb_v2/utils/vision_utils/video_transforms.py:486

Method__repr__

(self)

src/evaluation/mmeb_v2/utils/vision_utils/video_transforms.py:522

Method__repr__

(self)

src/evaluation/mmeb_v2/utils/vision_utils/video_transforms.py:577

Method__repr__

(self)

src/evaluation/mmeb_v2/utils/vision_utils/video_transforms.py:602

Method__repr__

(self)

src/evaluation/mmeb_v2/utils/vision_utils/video_transforms.py:624

Method__repr__

(self)

src/evaluation/mmeb_v2/utils/vision_utils/video_transforms.py:649

Method_truncate_tokens

(self, token_ids: List[int], max_length: int)

src/models/qwen3_vl_embedding.py:201

Functionadd_metainfo_hook

(f)

src/evaluation/mmeb_v2/data/datasets/base_eval_dataset.py:68

Functionbatch_to_device

(batch, device)

src/evaluation/mmeb_v2/utils/basic_utils.py:23

Methodconfig

(self)

src/evaluation/mmeb_v2/models.py:34

Methoddcg

(rel_list, form)

src/evaluation/mmeb_v2/utils/eval_utils/metrics.py:164

Methodforward

(self, input_ids: torch.LongTensor = None, attention_mask: Optional[torch.Tens

src/models/qwen3_vl_embedding.py:85

Methodget_decoder

(self)

src/models/qwen3_vl_embedding.py:61

Functionget_image

(image_path)

src/evaluation/mmeb_v2/utils/vision_utils/vision_utils.py:227

Methodget_image_features

(self, pixel_values: torch.FloatTensor, image_grid_thw: Optional[torch.LongTensor]

src/models/qwen3_vl_embedding.py:70

Methodget_input_embeddings

(self)

src/models/qwen3_vl_embedding.py:52

Methodget_video_features

(self, pixel_values_videos: torch.FloatTensor, video_grid_thw: Optional[torch.LongT

src/models/qwen3_vl_embedding.py:65

Methodinner_wrapper

(wrapped_class)

src/evaluation/mmeb_v2/data/datasets/base_eval_dataset.py:55

Methodlanguage_model

(self)

src/models/qwen3_vl_embedding.py:76

Functionload_activitynetqa_dataset

(model_args, data_args, *args, **kwargs)

src/evaluation/mmeb_v2/data/datasets/activitynetqa_dataset.py:91

Functionload_didemo_dataset

(model_args, data_args, **kwargs)

src/evaluation/mmeb_v2/data/datasets/didemo_dataset.py:56

Functionload_egoschema_dataset

(model_args, data_args, *args, **kwargs)

src/evaluation/mmeb_v2/data/datasets/egoschema_dataset.py:93

Functionload_image_cls_dataset

(model_args, data_args, *args, **kwargs)

src/evaluation/mmeb_v2/data/datasets/image_cls_dataset.py:49

Functionload_image_i2i_vg_dataset

(model_args, data_args, *args, **kwargs)

src/evaluation/mmeb_v2/data/datasets/image_i2i_vg_dataset.py:66

Functionload_image_i2t_dataset

(model_args, data_args, *args, **kwargs)

src/evaluation/mmeb_v2/data/datasets/image_i2t_eval.py:48

Functionload_image_qa_dataset

(model_args, data_args, *args, **kwargs)

src/evaluation/mmeb_v2/data/datasets/image_qa_dataset.py:46

Functionload_image_t2i_dataset

(model_args, data_args, *args, **kwargs)

src/evaluation/mmeb_v2/data/datasets/image_t2i_eval.py:63

Functionload_local_hf_dataset

Loads a Hugging Face dataset from local Parquet files. Args: dataset_path (str): The base path to the dataset directory subse

src/evaluation/mmeb_v2/utils/dataset_utils.py:58

Functionload_moment_retrieval_dataset

(model_args, data_args, **kwargs)

src/evaluation/mmeb_v2/data/datasets/moment_retrieval_datasets.py:95

Functionload_momentseeker_dataset

(model_args, data_args, *args, **kwargs)

src/evaluation/mmeb_v2/data/datasets/momentseeker_dataset.py:119

Functionload_msvd_dataset

(model_args, data_args, **kwargs)

src/evaluation/mmeb_v2/data/datasets/msvd_dataset.py:62

Functionload_mvbench_dataset

(model_args, data_args, *args, **kwargs)

src/evaluation/mmeb_v2/data/datasets/mvbench_dataset.py:114

Functionload_ssv2_dataset

SSv2-MC setup for zero-shot evaluation.

src/evaluation/mmeb_v2/data/datasets/ssv2_dataset.py:64

Functionload_vatex_dataset

(model_args, data_args, **kwargs)

src/evaluation/mmeb_v2/data/datasets/vatex_dataset.py:59

Functionload_videomme_dataset

(model_args, data_args, *args, **kwargs)

src/evaluation/mmeb_v2/data/datasets/videomme_dataset.py:98

Functionload_videommmu_dataset

(model_args, data_args, training_args, *args, **kwargs)

src/evaluation/mmeb_v2/data/datasets/videommmu_dataset.py:91

Functionload_visrag_dataset

(model_args, data_args, **kwargs)

src/evaluation/mmeb_v2/data/datasets/visrag_dataset.py:98

Functionload_youcook2_dataset

(model_args, data_args, **kwargs)

src/evaluation/mmeb_v2/data/datasets/youcook2_dataset.py:60

Methodmain

(self)

src/evaluation/mmeb_v2/data/datasets/base_eval_dataset.py:64

Methodmean_average_precision_at_k

Helper for MAP@k

src/evaluation/mmeb_v2/utils/eval_utils/metrics.py:120

Methodmean_reciprocal_rank_at_k

Helper for MRR@k

src/evaluation/mmeb_v2/utils/eval_utils/metrics.py:132

Functionnatural_sort_key

Extract numbers from filenames for correct sorting.

src/evaluation/mmeb_v2/utils/vision_utils/vision_utils.py:236

Methodprocess

(self, inputs: List[Dict[str, Any]], normalize: bool = True)

src/models/qwen3_vl_embedding.py:373

Functionread_file

(input_path)

src/evaluation/mmeb_v2/utils/vision_utils/video_transforms.py:29

Functionread_file

(input_path)

src/evaluation/mmeb_v2/utils/vision_utils/vision_utils.py:53

Functionread_from_path

(path, image_size, transform_name="center")

src/evaluation/mmeb_v2/utils/vision_utils/video_transforms.py:132

← previous next →101–200 of 211, ranked by callers