Vision-datasets

What's Changed
* put torch back to extra_requires by pj-ms in https://github.com/microsoft/vision-datasets/pull/69


**Full Changelog**: https://github.com/microsoft/vision-datasets/compare/v1.0.6...v1.0.7

What's Changed
* vision_datasets/visual_object_grounding/manifest.py by msxingpeng in https://github.com/microsoft/vision-datasets/pull/68


**Full Changelog**: https://github.com/microsoft/vision-datasets/compare/v1.0.5...v1.0.6

What's Changed
* Pinjin/jsonl convert by pj-ms in https://github.com/microsoft/vision-datasets/pull/67
Add additional info support for all kinds of manifest
"grounding" => "groundings" in coco json format
image text matching accepts match value in float between 0 to 1
proper support manifest => JSONL conversion

**Full Changelog**: https://github.com/microsoft/vision-datasets/compare/v1.0.4...v1.0.5

What's Changed
* support VQA; support grounding by pj-ms in https://github.com/microsoft/vision-datasets/pull/66


**Full Changelog**: https://github.com/microsoft/vision-datasets/compare/v1.0.3...v1.0.4

What's Changed
* remove print; unify return tuple types;remove broken few-shot; by pj-ms in https://github.com/microsoft/vision-datasets/pull/64


**Full Changelog**: https://github.com/microsoft/vision-datasets/compare/v1.0.2...v1.0.3

What's Changed
* Add ltrb to iris manifest by pj-ms in https://github.com/microsoft/vision-datasets/pull/63


**Full Changelog**: https://github.com/microsoft/vision-datasets/compare/v1.0.1...v1.0.2