Functions — github.com/UCSC-VLAA/OpenVision

hub / github.com/UCSC-VLAA/OpenVision / functions

Functions716 in github.com/UCSC-VLAA/OpenVision

⨍ Functions 716◇ Types & classes 90 ↳ Endpoints 1

Method_ikok_pp_fn

(data)

src/helpers/registry.py:58

Function_image_wider_than_requested_aspect_ratio

()

src/transforms/simclr_aug.py:205

Function_inception_crop

(image)

src/transforms/ops_image.py:141

Function_keep

(data)

src/transforms/ops_general.py:145

Function_make_global_arr

(x, shard, shape)

src/helpers/sharding.py:77

Method_merge_heads

(self, hidden_states, embed_dim)

src/models/text_decoder.py:179

Function_on_commit_callback

(array_names)

src/helpers/utils.py:1042

Function_onehot

(data)

src/transforms/ops_general.py:125

Function_pad_to_shape

(x)

src/transforms/ops_general.py:231

Function_pp_ass_keys

(labels)

src/transforms/bert_ops.py:259

Function_pp_autoreg_label

(label)

src/transforms/ops_text.py:45

Function_pp_bert_tokenize

(labels)

src/transforms/bert_ops.py:114

Function_pp_custom_bert_tokenize

(data)

src/transforms/bert_ops.py:474

Function_pp_fn

(example)

src/datasets/input_pipeline.py:180

Function_pp_imagenet_labels

(label)

src/transforms/ops_text.py:35

Function_pp_noun_tokenize

(labels)

src/transforms/bert_ops.py:271

Function_preprocess_fn

The preprocessing function that is returned.

src/datasets/build_transforms.py:60

Function_rand_erase

(image)

src/transforms/ops_image.py:445

Function_randaug

(image)

src/transforms/ops_image.py:420

Function_random_flip_lr_pp

(image)

src/transforms/ops_image.py:314

Method_register

(item)

src/helpers/registry.py:144

Function_requested_aspect_ratio_wider_than_image

()

src/transforms/simclr_aug.py:198

Function_resize

Resizes image to a given size.

src/transforms/ops_image.py:70

Function_resize_small

(image)

src/transforms/ops_image.py:103

Function_rotate_level_to_arg

(level)

src/transforms/autoaugment.py:480

Function_shear_level_to_arg

(level)

src/transforms/autoaugment.py:509

Function_shrink_level_to_arg

Converts level to ratio by which we shrink the image content.

src/transforms/autoaugment.py:486

Function_single_compose_two_images

(image1, image2)

src/transforms/mixup.py:297

Method_split_heads

(self, hidden_states, num_heads, head_dim)

src/models/text_decoder.py:176

Function_squeeze_last_dim

(x)

src/transforms/ops_general.py:198

Function_sync

(x)

src/helpers/utils.py:752

Function_transform

(image)

src/transforms/ops_image.py:270

Function_transform

(image)

src/transforms/simclr_aug.py:364

Method_unlock

(x)

src/convert_upload/open_clip/transformer.py:563

Function_value_range

Scales values in given range.

src/transforms/ops_general.py:51

Function_vgg_value_range

(image)

src/transforms/ops_image.py:340

Method_window_mask

Mask a part of the image.

src/models/vit.py:544

Functionadd_id

(idx, features)

src/evaluators/proj/image_text/retrieval.py:120

Functionadd_model_config

add model config path or file and update registry

src/convert_upload/open_clip/factory.py:72

Functionadd_prompts

(idx, class_name)

src/evaluators/proj/image_text/discriminative_classifier.py:89

Functionautocontrast

Implements Autocontrast function from PIL using TF ops. Args: image: A 3D uint8 tensor. Returns: The image after it has had autocontrast a

src/transforms/autoaugment.py:296

Methodbackward

(ctx, grad_output)

src/convert_upload/open_clip/loss.py:281

Methodbackward

(ctx, *grad_outputs)

src/convert_upload/open_clip/loss.py:298

Functionbce

Binary Cross Entropy Loss Should be numerically stable, built based on: https://github.com/pytorch/pytorch/issues/751 :param

src/losses/common.py:258

Functionbce_logits

(*, logits, labels, weight=None, reduction=True)

src/losses/common.py:256

Functionbrightness

Equivalent of PIL Brightness.

src/transforms/autoaugment.py:227

Functionbrightness_foo

()

src/transforms/simclr_aug.py:139

Methodbuild_cls_mask

(self, text, cast_dtype: torch.dtype)

src/convert_upload/open_clip/transformer.py:772

Functionbuild_zero_shot_classifier

Build zero-shot classifier weights by iterating over class names in batches Args: model: CLIP model instance tokenizer: CLIP toke

src/convert_upload/open_clip/zero_shot_classifier.py:21

Functionbuild_zero_shot_classifier_legacy

Build zero-shot classifier weights by iterating over class names 1 by 1 Args: model: CLIP model instance tokenizer: CLIP tokenize

src/convert_upload/open_clip/zero_shot_classifier.py:71

Functioncenter_crop

Crops to center of image and rescales to desired size. Args: image: Image Tensor to crop. height: Height of image to be cropped.

src/transforms/simclr_aug.py:219

Functioncheck_and_compile

(pattern)

src/helpers/utils.py:781

Functioncheck_zero_image

(test_img, org_img)

src/transforms/autoaugment.py:773

Functioncheckpointing_timeout

(writer, timeout)

src/helpers/utils.py:241

Functioncolor

Equivalent of PIL Color.

src/transforms/autoaugment.py:204

Functioncolor_jitter

Distorts the color of the image. Args: image: The input image tensor. strength: the floating number for the strength of the color augm

src/transforms/simclr_aug.py:55

Functionconcate_two_captions

(data)

src/transforms/bert_ops.py:441

Functioncontrast

Equivalent of PIL Contrast.

src/transforms/autoaugment.py:210

Functioncontrast_foo

()

src/transforms/simclr_aug.py:146

Functionconvert_int8_model_to_inference_mode

(model)

src/convert_upload/open_clip/utils.py:84

Methodcount_correct

Returns count of correct predictions (and optionally embeddings).

src/evaluators/proj/image_text/discriminative_classifier.py:305

Functioncreate_loss

(args)

src/convert_upload/open_clip/factory.py:359

Functioncreate_model_and_transforms

( model_name: str, pretrained: Optional[str] = None, precision: str = 'fp32',

src/convert_upload/open_clip/factory.py:396

Functioncutmix

Composing two images by inserting a patch into another image.

src/transforms/mixup.py:218

Functioncutout

Apply cutout (https://arxiv.org/abs/1708.04552) to image. This operation applies a (2*pad_size x 2*pad_size) mask of zeros to a random location wi

src/transforms/autoaugment.py:139

Functiondecode

(output_ids: torch.Tensor)

src/convert_upload/open_clip/tokenizer.py:272

Methoddistort

Given an image tensor, returns a distorted image with the same shape. Args: image: `Tensor` of shape [height, width, 3] or [num_fram

src/transforms/random_erasing.py:67

Methoddistort

Applies RandomErasing to single `image`. Args: image (tf.Tensor): Of shape [height, width, 3] representing an image. Returns: tf.T

src/transforms/random_erasing.py:139

Functiondistort_image_with_autoaugment

Applies the AutoAugment policy to `image`. AutoAugment is from the paper: https://arxiv.org/abs/1805.09501. Args: image: `Tensor` of shape [he

src/transforms/autoaugment.py:685

Functiondistort_image_with_randaugment

Applies the RandAugment policy to `image`. RandAugment is from the paper https://arxiv.org/abs/1909.13719, Args: image: `Tensor` of shape [hei

src/transforms/autoaugment.py:713

Functiondistort_image_with_randaugment_timm

Applies the RandAugment policy to `image`. RandAugment is from the paper https://arxiv.org/abs/1909.13719, Args: image: `Tensor` of shape [hei

src/transforms/autoaugment.py:778

Functiondistort_image_with_three_aug

Applies the RandAugment policy to `image`. RandAugment is from the paper https://arxiv.org/abs/1909.13719, Args: image: `Tensor` of shape [hei

src/transforms/three_aug.py:187

Methoddistort_with_boxes

Distorts the image and bounding boxes. Args: image: `Tensor` of shape [height, width, 3] or [num_frames, height, width, 3] represent

src/transforms/random_erasing.py:80

Functiondot_product_attention

Computes dot-product attention given query, key, and value. This is the core function for applying attention based on https://arxiv.org/abs/1706.

src/models/common.py:138

Methodembed_images

(train_state, images)

src/evaluators/proj/image_text/retrieval.py:206

Methodembed_texts

(train_state, texts)

src/evaluators/proj/image_text/retrieval.py:210

Methodembed_texts

Returns text embeddings.

src/evaluators/proj/image_text/discriminative_classifier.py:300

Methodencode_image

(self, images, normalize: bool = True)

src/convert_upload/open_clip/coca_model.py:159

Methodencode_text

(self, text, normalize: bool = True)

src/convert_upload/open_clip/coca_model.py:163

Functionequalize

Implements Equalize function from PIL using TF ops.

src/transforms/autoaugment.py:365

Functioneval_logits_fn

(train_state, batch)

src/main_clip.py:620

Functioneval_logits_fn

(state, batch)

src/main_openvision2.py:347

Functionexpand_aliases

(idx, class_name)

src/evaluators/proj/image_text/discriminative_classifier.py:82

Functionf1

()

src/transforms/ops_general.py:185

Functionf2

()

src/transforms/ops_general.py:186

Functionfinal_policy

(image_)

src/transforms/autoaugment.py:672

Functionfix_old_checkpoints

Fix small bwd incompat that can't be resolved with names in model def.

src/models/vit.py:864

Functionflatten

(data)

src/transforms/ops_general.py:246

Functionformat_str

Format string with reference fields from config. This makes it easy to build preprocess strings that contain references to fields tha are edited

src/configs/common.py:143

Methodforward

(self, image_features, text_features, logits, labels, logit_scale, output_dict=False)

src/convert_upload/open_clip/loss.py:160

Methodforward

( self, image_features, text_features, logit_scale,

src/convert_upload/open_clip/loss.py:185

Methodforward

(ctx, from_rank, to_rank, group, tensor)

src/convert_upload/open_clip/loss.py:274

Methodforward

(ctx, left_rank, right_rank, group, tensor_to_left, tensor_to_right)

src/convert_upload/open_clip/loss.py:291

Methodforward

(self, image_features, text_features, logit_scale, logit_bias, output_dict=False)

src/convert_upload/open_clip/loss.py:360

Methodforward

( self, image, text: Optional[torch.Tensor] = None, image_late

src/convert_upload/open_clip/coca_model.py:167

Methodforward

(self, x: torch.Tensor)

src/convert_upload/open_clip/modified_resnet.py:42

Methodforward

(self, x)

src/convert_upload/open_clip/modified_resnet.py:68

Methodforward

(self, x)

src/convert_upload/open_clip/modified_resnet.py:173

Methodforward

Args: img (PIL Image or Tensor): Image to be cropped. Returns: PIL Image or Tensor: Cropped image.

src/convert_upload/open_clip/transform.py:224

Methodforward

(self, x)

src/convert_upload/open_clip/timm_model.py:149

← previous next →401–500 of 716, ranked by callers