sync 1.1.2 >> feature 1.2 [full merge, no squash]

.drone.yml

@@ -33,11 +33,10 @@ steps:
     - python --version
     - pip --version
     - nvidia-smi
-    - pip install -r ./requirements/devel.txt --upgrade-strategy only-if-needed -v --no-cache-dir
-    - pip install git+https://${AUTH_TOKEN}@github.com/PyTorchLightning/[email protected] -v --no-cache-dir
+    - pip install -r ./requirements/devel.txt --upgrade-strategy only-if-needed --no-cache-dir
+    - pip install git+https://${AUTH_TOKEN}@github.com/PyTorchLightning/[email protected] --no-cache-dir
     # when Image has defined CUDa version we can switch to this package spec "nvidia-dali-cuda${CUDA_VERSION%%.*}0"
-    # todo: temprarl fix till https://github.com/PyTorchLightning/pytorch-lightning/pull/4922 is resolved
-    - pip install --extra-index-url https://developer.download.nvidia.com/compute/redist "nvidia-dali-cuda100<0.27" --upgrade-strategy only-if-needed
+    - pip install --extra-index-url https://developer.download.nvidia.com/compute/redist nvidia-dali-cuda100 --upgrade-strategy only-if-needed
     - pip list
     - python -m coverage run --source pytorch_lightning -m pytest pytorch_lightning tests -v --durations=25 # --flake8
     # Running special tests

.github/workflows/release-docker.yml

@@ -26,7 +26,7 @@ jobs:
       - name: Get release version
         if: startsWith(github.ref, 'refs/tags/') || github.event_name == 'release'
         id: get_version
-        run: echo ::set-env name=RELEASE_VERSION::$(echo ${GITHUB_REF##*/})
+        run: echo "::set-output name=RELEASE_VERSION::$(echo ${GITHUB_REF##*/})"
 
       - name: Publish Releases to Docker
         # only on releases
@@ -37,6 +37,6 @@ jobs:
           username: ${{ secrets.DOCKER_USERNAME }}
           password: ${{ secrets.DOCKER_PASSWORD }}
           dockerfile: dockers/release/Dockerfile
-          build_args: PYTHON_VERSION=${{ matrix.python_version }},PYTORCH_VERSION=${{ matrix.pytorch_version }},LIGHTNING_VERSION=${{ env.RELEASE_VERSION }}
-          tags: "${{ env.RELEASE_VERSION }}-py${{ matrix.python_version }}-torch${{ matrix.pytorch_version }},latest-py${{ matrix.python_version }}-torch${{ matrix.pytorch_version }}"
+          build_args: PYTHON_VERSION=${{ matrix.python_version }},PYTORCH_VERSION=${{ matrix.pytorch_version }},LIGHTNING_VERSION=${{ steps.get_version.outputs.RELEASE_VERSION }}
+          tags: "${{ steps.get_version.outputs.RELEASE_VERSION }}-py${{ matrix.python_version }}-torch${{ matrix.pytorch_version }},latest-py${{ matrix.python_version }}-torch${{ matrix.pytorch_version }}"
         timeout-minutes: 55

.mergify.yml

@@ -12,59 +12,59 @@
 # See the License for the specific language governing permissions and
 # limitations under the License.
 
-pull_request_rules:
-
-  - name: Automatic merge on approval
-    conditions:
-      - base=master
-      # number of review approvals
-      - "#approved-reviews-by>=3"
-      # no waiting or assigned review
-      - "#review-requested=0"
-      # no requested chnages from any reviewer
-      - "#changes-requested-reviews-by=0"
-      # this serves as ALL check has to pass as we have actually around 40 tests in total
-      - "#status-success>=54"
-      # this is just in case since we rely on GPU tests (note: redundand to the above)
-      - status-success=continuous-integration/drone/pr
-      - "status-success=ci/circleci: TPU-tests"
-      # this is patter-like, unofrunatly serves as `any(...)` (note: redundand to the above)
-      #- "status-success~=^ci/circleci:"
-      # no conflict with master branch
-      - -conflict
-      # was not closed yet
-      - -closed
-      # filter-out GH draft PRs
-      - -draft
-    actions:
-      delete_head_branch: {}
-      merge:
-        # https://doc.mergify.io/merge-action.html#strict-merge
-        # (on head branch) $ git merge --no-ff base
-        # (on head branch) # Wait for CI to go green
-        # (on head branch) # Squash all commits
-        # (on base branch) $ git merge --ff head
-        strict: true
-        method: squash
-      comment:
-        message: Great job! =)
-
-  - name: warn on conflicts
-    conditions:
-      - conflict
-      # filter-out GH draft PRs
-      - -draft
-    actions:
-      comment:
-        message: This pull request is now in conflict... :(
-
-  - name: add core reviewer
-    conditions:
-      # filter-out GH draft PRs
-      - -draft
-      # number of review approvals
-      - "#approved-reviews-by<3"
-    actions:
-      request_reviews:
-        teams:
-          - core-contributors
+#pull_request_rules:
+#
+#  - name: Automatic merge on approval
+#    conditions:
+#      - base=master
+#      # number of review approvals
+#      - "#approved-reviews-by>=3"
+#      # no waiting or assigned review
+#      - "#review-requested=0"
+#      # no requested chnages from any reviewer
+#      - "#changes-requested-reviews-by=0"
+#      # this serves as ALL check has to pass as we have actually around 40 tests in total
+#      - "#status-success>=54"
+#      # this is just in case since we rely on GPU tests (note: redundand to the above)
+#      - status-success=continuous-integration/drone/pr
+#      - "status-success=ci/circleci: TPU-tests"
+#      # this is patter-like, unofrunatly serves as `any(...)` (note: redundand to the above)
+#      #- "status-success~=^ci/circleci:"
+#      # no conflict with master branch
+#      - -conflict
+#      # was not closed yet
+#      - -closed
+#      # filter-out GH draft PRs
+#      - -draft
+#    actions:
+#      delete_head_branch: {}
+#      merge:
+#        # https://doc.mergify.io/merge-action.html#strict-merge
+#        # (on head branch) $ git merge --no-ff base
+#        # (on head branch) # Wait for CI to go green
+#        # (on head branch) # Squash all commits
+#        # (on base branch) $ git merge --ff head
+#        strict: true
+#        method: squash
+#      comment:
+#        message: Great job! =)
+#
+#  - name: warn on conflicts
+#    conditions:
+#      - conflict
+#      # filter-out GH draft PRs
+#      - -draft
+#    actions:
+#      comment:
+#        message: This pull request is now in conflict... :(
+#
+#  - name: add core reviewer
+#    conditions:
+#      # filter-out GH draft PRs
+#      - -draft
+#      # number of review approvals
+#      - "#approved-reviews-by<3"
+#    actions:
+#      request_reviews:
+#        teams:
+#          - core-contributors

CHANGELOG.md

@@ -50,6 +50,64 @@ The format is based on [Keep a Changelog](http://keepachangelog.com/en/1.0.0/).
 - Fixed distributed setting and `ddp_cpu` only with `num_processes>1` ([#5297](https://github.com/PyTorchLightning/pytorch-lightning/pull/5297))
 
 
+- Fixed the saved filename in `ModelCheckpoint` when it already exists ([#4861](https://github.com/PyTorchLightning/pytorch-lightning/pull/4861))
+
+
+- Fixed `DDPHPCAccelerator` hangs in DDP construction by calling `init_device` ([#5157](https://github.com/PyTorchLightning/pytorch-lightning/pull/5157))
+
+
+## [1.1.2] - 2020-12-23
+
+### Added
+
+- Support number for logging with `sync_dist=True` ([#5080](https://github.com/PyTorchLightning/pytorch-lightning/pull/5080)
+- Added offset logging step when resuming for Wandb logger ([#5050](https://github.com/PyTorchLightning/pytorch-lightning/pull/5050)
+
+### Removed
+
+- `enable_pl_optimizer=False` by default to temporarily fix AMP issues ([#5163](https://github.com/PyTorchLightning/pytorch-lightning/pull/5163)
+
+### Fixed
+
+- Metric reduction with Logging ([#5150](https://github.com/PyTorchLightning/pytorch-lightning/pull/5150)
+- Remove nan loss in manual optimization ([#5121](https://github.com/PyTorchLightning/pytorch-lightning/pull/5121)
+- Un-balanced logging properly supported ([#5119](https://github.com/PyTorchLightning/pytorch-lightning/pull/5119)
+- Fix hanging in DDP HPC accelerators ([#5157](https://github.com/PyTorchLightning/pytorch-lightning/pull/5157)
+- Fix saved filename in `ModelCheckpoint` if it already exists ([#4861](https://github.com/PyTorchLightning/pytorch-lightning/pull/4861)
+- Fix reset `TensorRunningAccum` ([#5106](https://github.com/PyTorchLightning/pytorch-lightning/pull/5106)
+- Updated `DALIClassificationLoader` to not use deprecated arguments ([#4925](https://github.com/PyTorchLightning/pytorch-lightning/pull/4925)
+- Corrected call to `torch.no_grad` ([#5124](https://github.com/PyTorchLightning/pytorch-lightning/pull/5124)
+
+
+## [1.1.1] - 2020-12-15
+
+### Added
+
+- Add a notebook example to reach a quick baseline of ~94% accuracy on CIFAR10 using Resnet in Lightning ([#4818](https://github.com/PyTorchLightning/pytorch-lightning/pull/4818))
+
+### Changed
+
+- Simplify accelerator steps ([#5015](https://github.com/PyTorchLightning/pytorch-lightning/pull/5015))
+- Refactor load in checkpoint connector ([#4593](https://github.com/PyTorchLightning/pytorch-lightning/pull/4593))
+
+### Removed
+
+- Drop duplicate metrics ([#5014](https://github.com/PyTorchLightning/pytorch-lightning/pull/5014))
+- Remove beta arg from F1 class and functional ([#5076](https://github.com/PyTorchLightning/pytorch-lightning/pull/5076))
+
+### Fixed
+
+- Fixed trainer by default `None` in `DDPAccelerator` ([#4915](https://github.com/PyTorchLightning/pytorch-lightning/pull/4915))
+- Fixed `LightningOptimizer` to expose optimizer attributes ([#5095](https://github.com/PyTorchLightning/pytorch-lightning/pull/5095))
+- Do not warn when the `name` key is used in the `lr_scheduler` dict ([#5057](https://github.com/PyTorchLightning/pytorch-lightning/pull/5057))
+- Check if optimizer supports closure ([#4981](https://github.com/PyTorchLightning/pytorch-lightning/pull/4981)
+- Extend LightningOptimizer to exposure underlying Optimizer attributes + update doc ([#5095](https://github.com/PyTorchLightning/pytorch-lightning/pull/5095))
+- Add deprecated metric utility functions back to functional (
+    [#5067](https://github.com/PyTorchLightning/pytorch-lightning/pull/5067),
+    [#5068](https://github.com/PyTorchLightning/pytorch-lightning/pull/5068))
+- Allow any input in `to_onnx` and `to_torchscript` ([#4378](https://github.com/PyTorchLightning/pytorch-lightning/pull/4378)
+- Do not warn when the name key is used in the `lr_scheduler` dict ([#5057](https://github.com/PyTorchLightning/pytorch-lightning/pull/5057))
+
 
 ## [1.1.0] - 2020-12-09
 
@@ -65,8 +123,8 @@ The format is based on [Keep a Changelog](http://keepachangelog.com/en/1.0.0/).
 - Added logging using `self.log` in train and evaluation for epoch end hooks (
     [#4552](https://github.com/PyTorchLightning/pytorch-lightning/pull/4552),
     [#4495](https://github.com/PyTorchLightning/pytorch-lightning/pull/4495),
-    [#4439](https://github.com/PyTorchLightning/pytorch-lightning/pull/4439))
-    [#4684](https://github.com/PyTorchLightning/pytorch-lightning/pull/4684))
+    [#4439](https://github.com/PyTorchLightning/pytorch-lightning/pull/4439),
+    [#4684](https://github.com/PyTorchLightning/pytorch-lightning/pull/4684),
     [#4913](https://github.com/PyTorchLightning/pytorch-lightning/pull/4913))
 - Added ability for DDP plugin to modify optimizer state saving ([#4675](https://github.com/PyTorchLightning/pytorch-lightning/pull/4675))
 - Added casting to python types for numpy scalars when logging hparams ([#4647](https://github.com/PyTorchLightning/pytorch-lightning/pull/4647))

Makefile

@@ -1,4 +1,4 @@
-.PHONY: test
+.PHONY: test clean
 
 test:
 	# install APEX, see https://github.com/NVIDIA/apex#linux
@@ -13,3 +13,7 @@ test:
 
 	# specific file
 	# python -m coverage run --source pytorch_lightning -m py.test --flake8 --durations=0 -v -k
+
+clean:
+	# clean all temp runs
+	rm -rf $(shell find . -name "mlruns" )

README.md

@@ -42,6 +42,11 @@ Scale your models, not the boilerplate.**
 
 ---
 
+## NEWS
+[Dec 2020 - Read about how Facebook uses Lightning to standardize deep learning across research and production teams](https://ai.facebook.com/blog/reengineering-facebook-ais-deep-learning-platforms-for-interoperability)
+
+---
+
 ## PyTorch Lightning is just organized PyTorch
 Lightning disentangles PyTorch code to decouple the science from the engineering.
 ![PT to PL](docs/source/_images/general/pl_quick_start_full_compressed.gif)
@@ -73,19 +78,6 @@ Lightning can automatically export to ONNX or TorchScript for those cases.
 
 ---
 
-## Trending contributors
-
-[![](https://sourcerer.io/fame/williamFalcon/pytorchlightning/pytorch-lightning/images/0)](https://sourcerer.io/fame/williamFalcon/pytorchlightning/pytorch-lightning/links/0)
-[![](https://sourcerer.io/fame/williamFalcon/pytorchlightning/pytorch-lightning/images/1)](https://sourcerer.io/fame/williamFalcon/pytorchlightning/pytorch-lightning/links/1)
-[![](https://sourcerer.io/fame/williamFalcon/pytorchlightning/pytorch-lightning/images/2)](https://sourcerer.io/fame/williamFalcon/pytorchlightning/pytorch-lightning/links/2)
-[![](https://sourcerer.io/fame/williamFalcon/pytorchlightning/pytorch-lightning/images/3)](https://sourcerer.io/fame/williamFalcon/pytorchlightning/pytorch-lightning/links/3)
-[![](https://sourcerer.io/fame/williamFalcon/pytorchlightning/pytorch-lightning/images/4)](https://sourcerer.io/fame/williamFalcon/pytorchlightning/pytorch-lightning/links/4)
-[![](https://sourcerer.io/fame/williamFalcon/pytorchlightning/pytorch-lightning/images/5)](https://sourcerer.io/fame/williamFalcon/pytorchlightning/pytorch-lightning/links/5)
-[![](https://sourcerer.io/fame/williamFalcon/pytorchlightning/pytorch-lightning/images/6)](https://sourcerer.io/fame/williamFalcon/pytorchlightning/pytorch-lightning/links/6)
-[![](https://sourcerer.io/fame/williamFalcon/pytorchlightning/pytorch-lightning/images/7)](https://sourcerer.io/fame/williamFalcon/pytorchlightning/pytorch-lightning/links/7)
-
----
-
 ## Continuous Integration
 <center>
 

benchmarks/__init__.py

@@ -0,0 +1,17 @@
+# Copyright The PyTorch Lightning team.
+#
+# Licensed under the Apache License, Version 2.0 (the "License");
+# you may not use this file except in compliance with the License.
+# You may obtain a copy of the License at
+#
+#     http://www.apache.org/licenses/LICENSE-2.0
+#
+# Unless required by applicable law or agreed to in writing, software
+# distributed under the License is distributed on an "AS IS" BASIS,
+# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+# See the License for the specific language governing permissions and
+# limitations under the License.
+import os
+
+BENCHMARK_ROOT = os.path.dirname(__file__)
+PROJECT_ROOT = os.path.dirname(BENCHMARK_ROOT)

benchmarks/generate_comparison.py

@@ -0,0 +1,60 @@
+# Copyright The PyTorch Lightning team.
+#
+# Licensed under the Apache License, Version 2.0 (the "License");
+# you may not use this file except in compliance with the License.
+# You may obtain a copy of the License at
+#
+#     http://www.apache.org/licenses/LICENSE-2.0
+#
+# Unless required by applicable law or agreed to in writing, software
+# distributed under the License is distributed on an "AS IS" BASIS,
+# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+# See the License for the specific language governing permissions and
+# limitations under the License.
+import os
+
+import matplotlib.pylab as plt
+import pandas as pd
+
+from benchmarks.test_basic_parity import lightning_loop, vanilla_loop
+from tests.base.models import ParityModuleMNIST, ParityModuleRNN
+
+NUM_EPOCHS = 20
+NUM_RUNS = 50
+MODEL_CLASSES = (ParityModuleRNN, ParityModuleMNIST)
+PATH_HERE = os.path.dirname(__file__)
+FIGURE_EXTENSION = '.png'
+
+
+def _main():
+    fig, axarr = plt.subplots(nrows=len(MODEL_CLASSES))
+
+    for i, cls_model in enumerate(MODEL_CLASSES):
+        path_csv = os.path.join(PATH_HERE, f'dump-times_{cls_model.__name__}.csv')
+        if os.path.isfile(path_csv):
+            df_time = pd.read_csv(path_csv, index_col=0)
+        else:
+            vanilla = vanilla_loop(cls_model, num_epochs=NUM_EPOCHS, num_runs=NUM_RUNS)
+            lightning = lightning_loop(cls_model, num_epochs=NUM_EPOCHS, num_runs=NUM_RUNS)
+
+            df_time = pd.DataFrame({'vanilla PT': vanilla['durations'][1:], 'PT Lightning': lightning['durations'][1:]})
+            df_time /= NUM_RUNS
+            df_time.to_csv(os.path.join(PATH_HERE, f'dump-times_{cls_model.__name__}.csv'))
+        # todo: add also relative X-axis ticks to see both: relative and absolute time differences
+        df_time.plot.hist(
+            ax=axarr[i],
+            bins=20,
+            alpha=0.5,
+            title=cls_model.__name__,
+            legend=True,
+            grid=True,
+        )
+        axarr[i].set(xlabel='time [seconds]')
+
+    path_fig = os.path.join(PATH_HERE, f'figure-parity-times{FIGURE_EXTENSION}')
+    fig.tight_layout()
+    fig.savefig(path_fig)
+
+
+if __name__ == '__main__':
+    _main()