update version to v0.2.2 (facebookresearch#882)

update version to v0.2.2 (facebookresearch#882)

Fix a minor bug about `generate_video` (facebookresearch#879)

* Fix a minor bug about `generate_video`

`fps` is not passed to `images_to_video`

* Update common.py

Follow black format

[ObsTransforms] Add support for semantic sensor observation transforms (

facebookresearch#847)

* [ObsTransforms] Add support for semantic sensor observation transforms

* Incoporate review

* Handle tensor

* Use semantic key to set interpolation mode

* Update black

* [BugFix] Return pytorch tensor in GPU2GPU mode

* Fix signature

* Fix signature

Version bump v0.2.1 (facebookresearch#699)

Version bump v0.2.1

Added HM3D dataset to Habitat Lab readme and new stickers (facebookre…

…search#669)

Added HM3D dataset to Habitat Lab readme and new stickers

Version change from 0.1.6 to 0.1.7 (facebookresearch#596)

* Version change from 0.1.6 to 0.1.7

* Made flake8 happy

Version change from 0.1.6 to 0.1.7 (facebookresearch#596)

* Version change from 0.1.6 to 0.1.7

* Made flake8 happy

Collapse PPO and DD-PPO trainers, faster RNN code, and double buffere…

…d sampling (facebookresearch#538)

- Collapse PPO and DD-PPO trainers
- Faster RNN code -- it is definitely faster and can make a noticeable impact during early training (~20% faster in some cases), but good luck reading it :-)
- Rename NUM_PROCESSES to NUM_ENVIRONMENTS. The fact that the simulators are in different processes is an implementation detail. A backwards compatibility check has been added tho.
- Support specifying training length both in terms of number of updates and number of frames
- Specify the number of checkpoints as the number of checkpoints instead of a checkpoint interval
- Introduce a TensorDict class for more cleanly interacting with dictionaries of tensors (potentially recursive). This also makes RolloutStorage about 100x cleaner.
- Store RGB observations as their proper dtype in the rollout storage (this can save a lot of memory)
- Some refactoring of PPOTrainer.train to be less of a script wrapped in a function
- Double buffered sampling. This can improve performance when simulation time is equal or larger than policy inference time

Move baseline utils from common to a dedicated utils folder (facebook…

…research#482)

This is to move habitat_baselines utils from habitat_baselines/common to a dedicated habitat_baselines/util