Merge branch 'master' into tests/doctest-examples

.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

@@ -5,7 +5,7 @@ All notable changes to this project will be documented in this file.
 The format is based on [Keep a Changelog](http://keepachangelog.com/en/1.0.0/).
 
 
-## [unreleased.Features] - YYYY-MM-DD
+## [unreleased.BugFix] - YYYY-MM-DD
 
 ### Added
 
@@ -22,28 +22,39 @@ The format is based on [Keep a Changelog](http://keepachangelog.com/en/1.0.0/).
 ### Fixed
 
 
-
-## [unreleased.BugFix] - YYYY-MM-DD
+## [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
 
-
-### Deprecated
+- 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)
+- Fixed the saved filename in `ModelCheckpoint` when it already exists ([#4861](https://github.com/PyTorchLightning/pytorch-lightning/pull/4861))
 
 
+=======
 ### 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` exposes optimizer attributes ([#5095](https://github.com/PyTorchLightning/pytorch-lightning/pull/5095))
-
+- 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)
+
+- Fixed `DDPHPCAccelerator` hangs in DDP construction by calling `init_device` ([#5157](https://github.com/PyTorchLightning/pytorch-lightning/pull/5157))
 
 
 ## [1.1.0] - 2020-12-09

benchmarks/test_parity.py

@@ -4,8 +4,8 @@
 import pytest
 import torch
 
+from pytorch_lightning import seed_everything, Trainer
 import tests.base.develop_utils as tutils
-from pytorch_lightning import Trainer, seed_everything
 from tests.base.models import ParityModuleMNIST, ParityModuleRNN
 
 

benchmarks/test_sharded_parity.py

@@ -6,7 +6,7 @@
 import pytest
 import torch
 
-from pytorch_lightning import Trainer, seed_everything
+from pytorch_lightning import seed_everything, Trainer
 from pytorch_lightning.plugins.ddp_plugin import DDPPlugin
 from pytorch_lightning.plugins.sharded_plugin import DDPShardedPlugin
 from pytorch_lightning.utilities import FAIRSCALE_AVAILABLE, NATIVE_AMP_AVAILABLE

dockers/base-xla/Dockerfile

@@ -97,6 +97,8 @@ RUN \
     python -c "fname = 'requirements.txt' ; lines = [line for line in open(fname).readlines() if not line.startswith('torch')] ; open(fname, 'w').writelines(lines)" && \
     # drop Horovod as it is not needed
     python -c "fname = 'requirements/extra.txt' ; lines = [line for line in open(fname).readlines() if not line.startswith('horovod')] ; open(fname, 'w').writelines(lines)" && \
+    # drop fairscale as it is not needed
+    python -c "fname = 'requirements/extra.txt' ; lines = [line for line in open(fname).readlines() if 'fairscale' not in line] ; open(fname, 'w').writelines(lines)" && \
     # drop TorchVision as it was installed with XLA
     python -c "fname = 'requirements/examples.txt' ; lines = [line for line in open(fname).readlines() if not line.startswith('torchvision')] ; open(fname, 'w').writelines(lines)" && \
     pip install --requirement ./requirements/devel.txt --upgrade-strategy only-if-needed && \

dockers/tpu-tests/Dockerfile

@@ -27,8 +27,10 @@ COPY ./ ./pytorch-lightning/
 RUN \
     # Install pytorch-lightning at the current PR, plus dependencies.
     #pip install -r pytorch-lightning/requirements.txt --no-cache-dir && \
-    # drop Horovod
+    # drop Horovod as it is not needed
     python -c "fname = 'pytorch-lightning/requirements/extra.txt' ; lines = [line for line in open(fname).readlines() if not line.startswith('horovod')] ; open(fname, 'w').writelines(lines)" && \
+    # drop fairscale as it is not needed
+    python -c "fname = 'pytorch-lightning/requirements/extra.txt' ; lines = [line for line in open(fname).readlines() if 'fairscale' not in line] ; open(fname, 'w').writelines(lines)" && \
     pip install -r pytorch-lightning/requirements/devel.txt --no-cache-dir --upgrade-strategy only-if-needed
 
 #RUN python -c "import pytorch_lightning as pl; print(pl.__version__)"

docs/source/introduction_guide.rst

@@ -601,8 +601,8 @@ In this method we do all the preparation we need to do once (instead of on every
         def setup(self, stage):
             # transform
             transform=transforms.Compose([transforms.ToTensor()])
-            MNIST(os.getcwd(), train=True, download=False, transform=transform)
-            MNIST(os.getcwd(), train=False, download=False, transform=transform)
+            mnist_train = MNIST(os.getcwd(), train=True, download=False, transform=transform)
+            mnist_test = MNIST(os.getcwd(), train=False, download=False, transform=transform)
 
             # train/val split
             mnist_train, mnist_val = random_split(mnist_train, [55000, 5000])

docs/source/multi_gpu.rst

@@ -663,7 +663,7 @@ It is highly recommended to use Sharded Training in multi-GPU environments where
 A technical note: as batch size scales, storing activations for the backwards pass becomes the bottleneck in training. As a result, sharding optimizer state and gradients becomes less impactful.
 Work within the future will bring optional sharding to activations and model parameters to reduce memory further, but come with a speed cost.
 
-To use Sharded Training, you need to first install FairScale using the command below or install all extras using ``pip install pytorch-lightning["extra"]``.
+To use Sharded Training, you need to first install FairScale using the command below.
 
 .. code-block:: bash
 

notebooks/04-transformers-text-classification.ipynb

@@ -1,5 +1,12 @@
 {
  "cells": [
+  {
+   "cell_type": "markdown",
+   "metadata": {},
+   "source": [
+    "<a href=\"https://colab.research.google.com/github/PytorchLightning/pytorch-lightning/blob/master/notebooks/04-transformers-text-classification.ipynb\" target=\"_parent\"><img src=\"https://colab.research.google.com/assets/colab-badge.svg\" alt=\"Open In Colab\"/></a>"
+   ]
+  },
   {
    "cell_type": "markdown",
    "metadata": {

notebooks/05-trainer-flags-overview.ipynb

@@ -1,5 +1,12 @@
 {
  "cells": [
+  {
+   "cell_type": "markdown",
+   "metadata": {},
+   "source": [
+    "<a href=\"https://colab.research.google.com/github/PytorchLightning/pytorch-lightning/blob/master/notebooks/05-trainer-flags-overview.ipynb\" target=\"_parent\"><img src=\"https://colab.research.google.com/assets/colab-badge.svg\" alt=\"Open In Colab\"/></a>"
+   ]
+  },
   {
    "cell_type": "markdown",
    "metadata": {

pyproject.toml

@@ -16,7 +16,7 @@ exclude = "(.eggs|.git|.hg|.mypy_cache|.nox|.tox|.venv|.svn|_build|buck-out|buil
 
 [tool.isort]
 known_first_party = [
-    "bencharmks",
+    "benchmarks",
     "docs",
     "pl_examples",
     "pytorch_lightning",
@@ -52,3 +52,5 @@ skip_glob = [
 ]
 profile = "black"
 line_length = 120
+force_sort_within_sections = "True"
+order_by_type = "False"

pytorch_lightning/__init__.py

@@ -1,6 +1,6 @@
 """Root package info."""
 
-__version__ = '1.1.1rc0'
+__version__ = '1.1.1'
 __author__ = 'William Falcon et al.'
 __author_email__ = 'waf2107@columbia.edu'
 __license__ = 'Apache-2.0'

pytorch_lightning/accelerators/ddp_cpu_hpc_accelerator.py

@@ -48,3 +48,6 @@ def model_to_device(self, model, process_idx):
     def get_device_ids(self):
         device_ids = None
         return device_ids
+
+    def init_device(self, process_idx):
+        pass

pytorch_lightning/accelerators/ddp_hpc_accelerator.py

@@ -126,6 +126,7 @@ def ddp_train(self, process_idx, model):
         """
         # determine which process we are and world size
         self.set_world_ranks(process_idx)
+        self.init_device(process_idx)
 
         # toggle prog bar
         if (self.trainer.node_rank != 0 or process_idx != 0) and self.trainer.progress_bar_callback is not None:

pytorch_lightning/callbacks/early_stopping.py

@@ -19,14 +19,16 @@
 Monitor a metric and stop training when it stops improving.
 
 """
+import numbers
 import os
 
 import numpy as np
 import torch
 
 from pytorch_lightning import _logger as log
 from pytorch_lightning.callbacks.base import Callback
-from pytorch_lightning.utilities import rank_zero_info, rank_zero_warn, TPU_AVAILABLE
+from pytorch_lightning.metrics.metric import Metric
+from pytorch_lightning.utilities import TPU_AVAILABLE, rank_zero_info, rank_zero_warn
 
 
 class EarlyStopping(Callback):
@@ -201,8 +203,11 @@ def _run_early_stopping_check(self, trainer, pl_module):
         # when in dev debugging
         trainer.dev_debugger.track_early_stopping_history(self, current)
 
-        if not isinstance(current, torch.Tensor):
-            current = torch.tensor(current, device=pl_module.device)
+        if current is not None:
+            if isinstance(current, Metric):
+                current = current.compute()
+            elif isinstance(current, numbers.Number):
+                current = torch.tensor(current, device=pl_module.device, dtype=torch.float)
 
         if trainer.use_tpu and TPU_AVAILABLE:
             current = current.cpu()

pytorch_lightning/callbacks/lr_monitor.py

@@ -157,7 +157,7 @@ def _find_names(self, lr_schedulers) -> List[str]:
         names = []
         for scheduler in lr_schedulers:
             sch = scheduler['scheduler']
-            if 'name' in scheduler:
+            if scheduler['name'] is not None:
                 name = scheduler['name']
             else:
                 opt_name = 'lr-' + sch.optimizer.__class__.__name__