Flambe

Breaking Changes
------------------
- fastBPE was removed, until a better alternative is settled on

Bug Fixes
----------
- Better default behavior when no devices is specified inside an Experiment
- Fixes a vulnerability with PyYAML < 5.3.1

New features
- When creating new AWS instances, they now receive an initial temporary name until the final name is assigned.

Bug Fixes
- Fixed an issue with running trials in parallel on a single node with multiple GPUs
- Set the default Metric aggregate behavior to better reflect common use cases

New Features
--------------

- ``TextField``: now accepts a vocabulary as input, supports additional special tokens and can operate over nested lists and dictionaries of strings
- ``MultiClassAUC``: new metric to compute non binary AUC
- ``Recall``: new metric to compute non binary recall
- ``BaseSampler``: has a new argument ``downsample_max_samples`` to downsample to a specific number of samples as opposed to a percentage of the data.
- We now use Github Actions for testing and publishing

Bug fixes
---------

- Fixed an issue with serialization on loading Tensors
- Fixed learning rate logging in the ``Trainer``
- ``Perplexity`` now aggregates correctly

New Features
* Now TextField can store all the embeddings provided, not just the ones that are on the datasets (236 )
* Changed most of the wrapper logic around metric computation for more robustness and generalization ability (226 )

Bug fixes
* We fixed the examples tests that were not able to run before on the CI because of hardware limitations.

**We now run fully on Github Actions!**