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

Download annotation files if necessary. All the vision-language datasets should have annotations of unified format. storage_

Resume from a checkpoint.

Load the best checkpoint for evaluation.

Create by split datasets inheriting torch.utils.data.Datasets. # build() can be dataset-specific. Overwrite to customize.

Build a dictionary of datasets, keyed by split 'train', 'valid', 'test'. Download dataset and annotations automatically if not exist.

Concatenate the batched input embedding and batched output embedding together. Both the input and the output embedding should be righ

Convert applicable model parameters to fp16

Decay the learning rate

Create dataloaders for training and validation.

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

Given a URL, returns the URL it redirects to or the original URL in case of no indirection

Get runner class from config. Default to epoch-based runner.

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

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

Example >>> dataset = load_dataset("coco_caption", cfg=None) >>> splits = dataset.keys() >>> print([len(dataset[split]) for split in

Load preprocessor configs and construct preprocessors. If no preprocessor is specified, return BaseProcessor, which does not do any preproce

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

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