Returns the smallest integer greater than or equal to 'number' that is divisible by 'factor'.

Check the whether the configuration is valid, if not, we will edit it/them. Args: config (Configuration): The configuration must inc

Converts a sequence of [tokens|ids] using the vocab.

Args: video_sample(np.ndArray): b,c,h,w, segment-based sampled video concatenated along channel-dim Returns:

Get current attributes/fields registered under the SampleList. Returns: List[str]: list of attributes of the SampleList.

Construct configuration from any mapping.

Get image ids that satisfy given filter conditions. default skips that filter :param quesIds (int array) : get image ids for gi

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

:param sim_matrix: #text(bsz) x #video(bsz*n_pair) :return:

get node from hier labels Note: Using this func to get tree node, especially when label exists more than once, the hierarchic

Rename the given file.

Updates a dictionary with other dictionary recursively. Parameters ---------- dictionary : dict Dictionary to be updated. upd

Perform the computation Parameters: predictions: raw outputs of the model target_sizes: tensor of dimension [batch_siz

r""" | .. image:: https://raw.githubusercontent.com/rusty1s/pytorch_scatter/ master/docs/source/_figures/add.svg?sanitize=true

Python 3.8 changes the default mode of multiprocessing on MacOS to spawn instead of fork. This requires all parameters passed to Sanic worker

Rescales the image so that the following conditions are met: 1. Both dimensions (height and width) are divisible by 'factor'. 2. The to

Args: image(torch.Tensor or np.ndArray): image with shape [h,w,c] coords(torch.Tensor or np.ndArray): Nx4, [xc,yc,w,h] or [x1,y1,

Call requires item to have either "tokens" attribute or either "text" attribute. If "text" is present, it will be tokenized using the

Calculate accuracy and return it back. Args: sample_list (SampleList): SampleList provided by DataLoader for

Applies Rotary Position Embedding with Multimodal Sections to the query and key tensors (https://qwenlm.github.io/blog/qwen2-vl/). Explanation:

TODO: add document here.

Args: text_sample_info(Dict): {'text': ''} image_sample_info(Dict): {'images': [image_tensor] }, with image_tensor's

Calculates the output image size when using Conv2dSamePadding with a stride. Necessary for static padding. Thanks to mannatsingh for pointing t

1x1 convolution

3x3 convolution with padding

Converts `text` to Unicode (if it's not already), assuming utf-8 input.

Cosine similarity between all the image and sentence pairs

This method forwards all its arguments to LlamaTokenizerFast's [`~PreTrainedTokenizer.decode`]. Please refer to the docstring of this

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

Returns the largest integer less than or equal to 'number' that is divisible by 'factor'.