Merge main into v0.4.0

.circleci/config.yml

@@ -47,7 +47,7 @@ run_pytorch_container: &run_pytorch_container
       environment:
         wd: << pipeline.parameters.workingdir >>
       command: |
-        docker run --gpus=all --rm -itd -v ${wd}:/ignite -w /ignite --name pthd << pipeline.parameters.pytorch_stable_image >>
+        docker run --gpus=all --rm -itd --shm-size 16G -v ${wd}:/ignite -w /ignite --name pthd << pipeline.parameters.pytorch_stable_image >>
         docker exec -it pthd nvidia-smi
         docker exec -it pthd ls
 
@@ -80,7 +80,7 @@ jobs:
 
             # pytest on cuda
             export test_cmd='sh tests/run_gpu_tests.sh'
-            docker exec -it pthd /bin/bash -c "$test_cmd"
+            docker exec -it pthd /bin/bash -c "${test_cmd}"
 
             # MNIST tests
 
@@ -118,7 +118,7 @@ jobs:
       - run:
           name: Codecov upload
           command: |
-            codecov -F gpu || echo 'Codecov upload failed'
+            bash <(curl -s https://codecov.io/bash) -Z -F gpu
 
 
   two_gpus_tests:
@@ -135,7 +135,64 @@ jobs:
           name: Run 1 Node 2 GPUs Unit Tests
           command: |
             export test_cmd='sh tests/run_gpu_tests.sh 2'
-            docker exec -it pthd /bin/bash -c "$test_cmd"
+            docker exec -it pthd /bin/bash -c "${test_cmd}"
+
+      - run:
+          name: Codecov upload
+          command: |
+            bash <(curl -s https://codecov.io/bash) -Z -F gpu-2
+
+
+  two_gpus_check_dist_cifar10_example:
+    <<: *two_gpus
+
+    working_directory: << pipeline.parameters.workingdir >>
+
+    steps:
+      - checkout
+      - <<: *pull_pytorch_stable_image
+      - <<: *run_pytorch_container
+      - <<: *install_dependencies
+      - run:
+          name: "Install additional example dependencies"
+          command: |
+            docker exec -it pthd pip install fire
+      - run:
+          name: "Run without backend"
+          command: |
+            export example_path="examples/contrib/cifar10"
+            # initial run
+            export stop_cmd="--stop_iteration=500"
+            export test_cmd="CI=1 python ${example_path}/main.py run"
+            docker exec -it pthd /bin/bash -c "${test_cmd} ${stop_cmd}"
+            # resume
+            export resume_opt="--resume-from=/tmp/output-cifar10/resnet18_backend-None-1_stop-on-500/training_checkpoint_400.pt"
+            docker exec -it pthd /bin/bash -c "${test_cmd} --num_epochs=7 ${resume_opt}"
+
+      - run:
+          name: "Run with NCCL backend using torch dist launch"
+          command: |
+            export example_path="examples/contrib/cifar10"
+            # initial run
+            export stop_cmd="--stop_iteration=500"
+            export test_cmd="CI=1 python -u -m torch.distributed.launch --nproc_per_node=2 --use_env ${example_path}/main.py run --backend=nccl"
+            docker exec -it pthd /bin/bash -c "${test_cmd} ${stop_cmd}"
+            # resume
+            export resume_opt="--resume-from=/tmp/output-cifar10/resnet18_backend-nccl-2_stop-on-500/training_checkpoint_400.pt"
+            docker exec -it pthd /bin/bash -c "${test_cmd} --num_epochs=7 ${resume_opt}"
+
+      - run:
+          name: "Run with NCCL backend using spawn"
+          command: |
+            export example_path="examples/contrib/cifar10"
+            # initial run
+            export stop_cmd="--stop_iteration=500"
+            export test_cmd="CI=1 python -u ${example_path}/main.py run --backend=nccl --nproc_per_node=2"
+            docker exec -it pthd /bin/bash -c "${test_cmd} ${stop_cmd}"
+            # resume
+            export resume_opt="--resume-from=/tmp/output-cifar10/resnet18_backend-nccl-2_stop-on-500/training_checkpoint_400.pt"
+            docker exec -it pthd /bin/bash -c "${test_cmd} --num_epochs=7 ${resume_opt}"
+
 
 # -------------------------------------------------------------------------------------
 # Workflows
@@ -146,3 +203,4 @@ workflows:
     jobs:
       - one_gpu_tests
       - two_gpus_tests
+      - two_gpus_check_dist_cifar10_example

.github/workflows/pytorch-version-tests.yml

@@ -1,7 +1,7 @@
 on: 
   schedule:
-    # Run at 00:00 UTC on Sunday
-    - cron:  '0 0 * * 0'
+    # Run at 00:00 UTC Every Day
+    - cron:  '0 0 * * *'
 
 jobs:
   build:
@@ -11,16 +11,10 @@ jobs:
       fail-fast: false
       matrix:
         python-version: [3.5, 3.6, 3.7, 3.8]
-        pytorch-version: [1.4.0, 1.3.1, 1.2.0, 1.1.0, 1.0.1]
+        pytorch-version: [1.4.0, 1.3.1]
         exclude:
           - pytorch-version: 1.3.1
             python-version: 3.8
-          - pytorch-version: 1.2.0
-            python-version: 3.8
-          - pytorch-version: 1.1.0
-            python-version: 3.8
-          - pytorch-version: 1.0.1
-            python-version: 3.8
 
     steps:
     - uses: actions/checkout@v2

.github/workflows/unittests.yml

@@ -2,54 +2,63 @@ on: [push, pull_request]
 
 jobs:
   build:
-    runs-on: ubuntu-latest
+    runs-on: ${{ matrix.os }}
     strategy:
       max-parallel: 10
       fail-fast: false
       matrix:
+        os: [ubuntu-latest, ]
         python-version: [3.5, 3.6, 3.7, 3.8]
         pytorch-channel: [pytorch, pytorch-nightly]
         exclude:
           # excludes pytorch-nightly python 3.5 as it was dropped
           - pytorch-channel: pytorch-nightly
             python-version: 3.5
+        include:
+          # includes a single build on windows
+          - os: windows-latest
+            pytorch-channel: pytorch
+            python-version: 3.7
+            skip-distrib-tests: 1
+          # includes a single build on macosx
+          - os: macos-latest
+            pytorch-channel: pytorch
+            python-version: 3.7
+            skip-distrib-tests: 1
 
     steps:
     - uses: actions/checkout@v2
-    - name: Setup Conda
-      run: |
-        wget https://repo.continuum.io/miniconda/Miniconda3-latest-Linux-x86_64.sh -O miniconda.sh
-        bash miniconda.sh -b -p $HOME/miniconda
-        export PATH="$HOME/miniconda/bin:$PATH"
-        conda config --set always_yes yes --set changeps1 no
-        conda update -q conda
-        # Useful for debugging any issues with conda
-        conda info -a
-        conda create -q -n test-environment pytorch cpuonly python=${{ matrix.python-version }} -c ${{ matrix.pytorch-channel }}
-
+    - name: Setup Miniconda
+      uses: goanpeca/setup-miniconda@v1
+      with:
+        miniconda-version: "latest"
+        python-version: ${{ matrix.python-version }}
+
     - name: Install dependencies
+      shell: bash -l {0}
       run: |
-        export PATH="$HOME/miniconda/bin:$PATH"
-        source activate test-environment
-        # Keep fix in case of problem with torchvision nightly releases
-        # if [[ "${{ matrix.pytorch-channel }}" == "pytorch-nightly" ]]; then pip install --upgrade git+https://github.com/pytorch/vision.git; else conda install torchvision cpuonly python=${{ matrix.python-version }} -c ${{ matrix.pytorch-channel }}; fi
-        conda install torchvision cpuonly python=${{ matrix.python-version }} -c ${{ matrix.pytorch-channel }}
+        conda install pytorch torchvision cpuonly -c ${{ matrix.pytorch-channel }}
         pip install -r requirements-dev.txt
+        # Fixes #1153
+        pip install --upgrade scipy==1.4.1
         python setup.py install
 
     - name: Run Tests
+      shell: bash -l {0}
       run: |
-        export PATH="$HOME/miniconda/bin:$PATH"
-        source activate test-environment
-        CI_PYTHON_VERSION="${{ matrix.python-version }}" sh tests/run_cpu_tests.sh
+        SKIP_DISTRIB_TESTS=${{ matrix.skip-distrib-tests }} CI_PYTHON_VERSION="${{ matrix.python-version }}" sh tests/run_cpu_tests.sh
 
     - name: Run MNIST Examples
+      shell: bash -l {0}
       run: |
-        export PATH="$HOME/miniconda/bin:$PATH"
-        source activate test-environment
         # MNIST
         # 1) mnist.py
         python examples/mnist/mnist.py --epochs=1
+
+    - name: Run MNIST with loggers Examples
+      shell: bash -l {0}
+      if: ${{ matrix.os == 'ubuntu-latest' }}
+      run: |
         # 2) mnist_with_visdom.py
         python -c "from visdom.server import download_scripts; download_scripts()" # download scripts : https://github.com/facebookresearch/visdom/blob/master/py/server.py#L929
         python -m visdom.server &
@@ -63,40 +72,40 @@ jobs:
         python examples/mnist/mnist_with_tensorboard.py --epochs=1
 
     - name: Run MNIST Example With Crash
+      shell: bash -l {0}
+      if: ${{ matrix.os == 'ubuntu-latest' }}
       continue-on-error: true
       run: |
-        export PATH="$HOME/miniconda/bin:$PATH"
-        source activate test-environment
         # 4) mnist_save_resume_engine.py
         python examples/mnist/mnist_save_resume_engine.py --epochs=2 --crash_iteration 1100
 
     - name: Resume MNIST from previous crash
+      shell: bash -l {0}
+      if: ${{ matrix.os == 'ubuntu-latest' }}
       run: |
-        export PATH="$HOME/miniconda/bin:$PATH"
-        source activate test-environment
         python examples/mnist/mnist_save_resume_engine.py --epochs=2 --resume_from=/tmp/mnist_save_resume/checkpoint_1.pt
 
     - name: Run GAN example
+      shell: bash -l {0}
+      if: ${{ matrix.os == 'ubuntu-latest' }}
       run: |
-        export PATH="$HOME/miniconda/bin:$PATH"
-        source activate test-environment
         # DCGAN
         python examples/gan/dcgan.py --dataset fake --dataroot /tmp/fakedata --output-dir /tmp/outputs-dcgan --batch-size 2 --epochs 2  --workers 0
 
     - name: Run RL Examples
+      shell: bash -l {0}
+      if: ${{ matrix.os == 'ubuntu-latest' }}
       run: |
-        export PATH="$HOME/miniconda/bin:$PATH"
-        source activate test-environment
         # RL
         # 1) Actor-Critic
         python examples/reinforcement_learning/actor_critic.py --max-episodes=2
         # 2) Reinforce
         python examples/reinforcement_learning/reinforce.py --max-episodes=2
 
     - name: Run Neural Style Example
+      shell: bash -l {0}
+      if: ${{ matrix.os == 'ubuntu-latest' }}
       run: |
-        export PATH="$HOME/miniconda/bin:$PATH"
-        source activate test-environment
         #fast-neural-style
         #train
         mkdir -p ~/.cache/torch/checkpoints/ && wget "https://download.pytorch.org/models/vgg16-397923af.pth" -O ~/.cache/torch/checkpoints/vgg16-397923af.pth