[cleanup] consistent __init__.py for import * (#550)

- fixes #471
- one less thing to worry about during development.

CONTRIBUTING.md

 # Contributing to fairscale
+
 We want to make contributing to this project as easy and transparent as
 possible.
 
 ## Our Development Process
+
 Minor changes and improvements will be released on an ongoing basis. Larger
 changes (e.g., changesets implementing a new paper) will be released on a
 more periodic basis.
 
 ## Pull Requests
+
 We actively welcome your pull requests.
 
 1. Fork the repo and create your branch from `master`.
@@ -18,12 +21,14 @@ We actively welcome your pull requests.
 6. If you haven't already, complete the Contributor License Agreement ("CLA").
 
 ## Contributor License Agreement ("CLA")
+
 In order to accept your pull request, we need you to submit a CLA. You only need
 to do this once to work on any of Facebook's open source projects.
 
 Complete your CLA here: <https://code.facebook.com/cla>
 
 ## Issues
+
 We use GitHub issues to track public bugs. Please ensure your description is
 clear and has sufficient instructions to be able to reproduce the issue.
 
@@ -41,17 +46,21 @@ outlined on that page and do not file a public issue.
 ```
 
 ## Coding Style
+
 * In your editor, install the [editorconfig](https://editorconfig.org/) extension
   which should ensure that you are following the same standards as us.
-* Ideally, run black and isort before opening up your PR.
+* Please run black and isort before opening up your PR.
 
 ```
 black .
-isort
+isort .
 flake8
 ```
-* Read the [editorconfig](.editorconfig) file to understand the exact coding style preferences.
-* Place Python code related to models in fairscale/nn. Place Python code related to optimizers in fairscale/optim. Place C++ extensions in fairscale/clib.
+* Please read the [editorconfig](.editorconfig) file to understand the exact coding style preferences.
+* Please place Python code related to models in fairscale/nn. Place Python code related to optimizers
+  in fairscale/optim. Place C++ extensions in fairscale/clib.
+* Please put `__all__:List[str] = []` in new `__init__.py` files for consistent importing behavior
+  and less development overhead in maintaining an importing list.
 
 ## Testing
 
@@ -72,7 +81,8 @@ python -m pytest tests/nn/data_parallel/test_oss_ddp.py::test_on_cpu
 ### Check test coverage
 
 ```
-python -m pytest --cov-report term --cov=fairscale/nn/data_parallel  tests/nn/data_parallel/test_oss_ddp.py::test_on_cpu
+python -m pytest --cov-report term --cov=fairscale/nn/data_parallel \
+   tests/nn/data_parallel/test_oss_ddp.py::test_on_cpu
 ```
 
 ### CircleCI status
@@ -108,6 +118,7 @@ Any line of the commit message cannot be longer 100 characters! This allows the
 to read on github as well as in various git tools.
 
 ### Type
+
 Must be one of the following:
 
 * **feat**: A new feature
@@ -122,5 +133,6 @@ generation
 * **docs**: Documentation only changes
 
 ## License
+
 By contributing to fairscale, you agree that your contributions will be licensed
 under the LICENSE file in the root directory of this source tree.

fairscale/__init__.py

@@ -10,4 +10,8 @@ __version__ = "0.3.2"
 # Import most common subpackages
 ################################################################################
 
+from typing import List
+
 from . import nn
+
+__all__: List[str] = []

fairscale/experimental/__init__.py

@@ -7,4 +7,8 @@
 # Import most common subpackages
 ################################################################################
 
+from typing import List
+
 from . import nn
+
+__all__: List[str] = []

fairscale/experimental/nn/__init__.py

@@ -2,3 +2,7 @@
 #
 # This source code is licensed under the BSD license found in the
 # LICENSE file in the root directory of this source tree.
+
+from typing import List
+
+__all__: List[str] = []

fairscale/experimental/nn/ampnet_pipe/__init__.py

+# Copyright (c) Facebook, Inc. and its affiliates. All rights reserved.
+#
+# This source code is licensed under the BSD license found in the
+# LICENSE file in the root directory of this source tree.
+
+from typing import List
+
+__all__: List[str] = []

fairscale/nn/__init__.py

@@ -3,23 +3,12 @@
 # This source code is licensed under the BSD license found in the
 # LICENSE file in the root directory of this source tree.
 
+from typing import List
+
 from .data_parallel import FullyShardedDataParallel, ShardedDataParallel
 from .misc import FlattenParamsWrapper, checkpoint_wrapper
 from .moe import MOELayer, Top2Gate
 from .pipe import Pipe, PipeRPCWrapper
 from .wrap import auto_wrap, default_auto_wrap_policy, enable_wrap, wrap
 
-__all__ = [
-    "FlattenParamsWrapper",
-    "checkpoint_wrapper",
-    "FullyShardedDataParallel",
-    "ShardedDataParallel",
-    "Pipe",
-    "PipeRPCWrapper",
-    "MOELayer",
-    "Top2Gate",
-    "auto_wrap",
-    "default_auto_wrap_policy",
-    "enable_wrap",
-    "wrap",
-]
+__all__: List[str] = []

fairscale/nn/data_parallel/__init__.py

@@ -3,5 +3,9 @@
 # This source code is licensed under the BSD license found in the
 # LICENSE file in the root directory of this source tree.
 
+from typing import List
+
 from .fully_sharded_data_parallel import FullyShardedDataParallel, TrainingState, auto_wrap_bn
 from .sharded_ddp import ShardedDataParallel
+
+__all__: List[str] = []

fairscale/nn/misc/__init__.py

@@ -3,6 +3,10 @@
 # This source code is licensed under the BSD license found in the
 # LICENSE file in the root directory of this source tree.
 
+from typing import List
+
 from .checkpoint_activations import checkpoint_wrapper
 from .flatten_params_wrapper import FlattenParamsWrapper
 from .param_bucket import GradBucket, ParamBucket
+
+__all__: List[str] = []

fairscale/nn/model_parallel/__init__.py

@@ -3,6 +3,8 @@
 # This source code is licensed under the BSD license found in the
 # LICENSE file in the root directory of this source tree.
 
+from typing import List
+
 from .cross_entropy import vocab_parallel_cross_entropy
 from .initialize import (
     destroy_model_parallel,
@@ -20,3 +22,5 @@ from .initialize import (
 from .layers import ColumnParallelLinear, RowParallelLinear, VocabParallelEmbedding
 from .mappings import copy_to_model_parallel_region, gather_from_model_parallel_region
 from .random import get_cuda_rng_tracker, model_parallel_cuda_manual_seed
+
+__all__: List[str] = []

fairscale/nn/moe/__init__.py

@@ -3,5 +3,9 @@
 # This source code is licensed under the BSD license found in the
 # LICENSE file in the root directory of this source tree.
 
+from typing import List
+
 from .moe_layer import MOELayer
 from .top2gate import Top2Gate
+
+__all__: List[str] = []

fairscale/nn/wrap/__init__.py

+# Copyright (c) Facebook, Inc. and its affiliates. All rights reserved.
+#
+# This source code is licensed under the BSD license found in the
+# LICENSE file in the root directory of this source tree.
+
+from typing import List
+
 from .auto_wrap import auto_wrap, default_auto_wrap_policy, enable_wrap, wrap
+
+__all__: List[str] = []

fairscale/optim/__init__.py

@@ -7,6 +7,7 @@
 :mod:`fairscale.optim` is a package implementing various torch optimization algorithms.
 """
 import logging
+from typing import List
 
 from .adascale import AdaScale, AdaScaleWrapper
 from .oss import OSS
@@ -19,3 +20,5 @@ try:
     from .grad_scaler import GradScaler
 except ImportError:
     logging.warning("Torch AMP is not available on this platform")
+
+__all__: List[str] = []

fairscale/utils/__init__.py

@@ -2,3 +2,7 @@
 #
 # This source code is licensed under the BSD license found in the
 # LICENSE file in the root directory of this source tree.
+
+from typing import List
+
+__all__: List[str] = []