Fix windows pipeline timeout (#2050)

microsoft · Feb 12, 2020 · 1958adb · 1958adb
1 parent 73a6990
commit 1958adb
Show file tree

Hide file tree

Showing 6 changed files with 15 additions and 5 deletions.
diff --git a/examples/trials/cifar10_pytorch/main.py b/examples/trials/cifar10_pytorch/main.py
@@ -99,7 +99,7 @@ def prepare(args):
 
 
 # Training
-def train(epoch):
+def train(epoch, batches=-1):
     global trainloader
     global testloader
     global net
@@ -129,6 +129,9 @@ def train(epoch):
         progress_bar(batch_idx, len(trainloader), 'Loss: %.3f | Acc: %.3f%% (%d/%d)'
             % (train_loss/(batch_idx+1), 100.*correct/total, correct, total))
 
+        if batches > 0 and (batch_idx+1) >= batches:
+            return
+
 def test(epoch):
     global best_acc
     global trainloader
@@ -176,6 +179,10 @@ def test(epoch):
 if __name__ == '__main__':
     parser = argparse.ArgumentParser()
     parser.add_argument("--epochs", type=int, default=200)
+
+    # Maximum mini-batches per epoch, for code testing purpose
+    parser.add_argument("--batches", type=int, default=-1)
+
     args, _ = parser.parse_known_args()
 
     try:
@@ -187,7 +194,7 @@ def test(epoch):
         acc = 0.0
         best_acc = 0.0
         for epoch in range(start_epoch, start_epoch+args.epochs):
-            train(epoch)
+            train(epoch, args.batches)
             acc, best_acc = test(epoch)
             nni.report_intermediate_result(acc)
 

diff --git a/test/config_test/examples/cifar10-pytorch.test.yml b/test/config_test/examples/cifar10-pytorch.test.yml
@@ -13,7 +13,7 @@ assessor:
     optimize_mode: maximize
 trial:
   codeDir: ../../../examples/trials/cifar10_pytorch
-  command: python3 main.py --epochs 2
+  command: python3 main.py --epochs 1 --batches 1
   gpuNum: 1
 
 useAnnotation: false

diff --git a/test/pipelines-it-local-windows.yml b/test/pipelines-it-local-windows.yml
@@ -1,5 +1,6 @@
 jobs:
-- job: 'Test'
+- job: 'integration_test_local_windows'
+  timeoutInMinutes: 0
 
   steps:
   - script: |
@@ -9,7 +10,6 @@ jobs:
       python -m pip install scikit-learn==0.20.0 --user
       python -m pip install keras==2.1.6 --user
       python -m pip install torch===1.2.0 torchvision===0.4.1 -f https://download.pytorch.org/whl/torch_stable.html --user
-      python -m pip install torchvision --user
       python -m pip install tensorflow-gpu==1.11.0 --user
     displayName: 'Install dependencies for integration tests'
   - script: |

diff --git a/test/pipelines-it-local.yml b/test/pipelines-it-local.yml
@@ -1,5 +1,6 @@
 jobs:
 - job: 'integration_test_local_ubuntu'
+  timeoutInMinutes: 0
 
   steps:
   - script: python3 -m pip install --upgrade pip setuptools --user

diff --git a/test/pipelines-it-remote-windows.yml b/test/pipelines-it-remote-windows.yml
@@ -1,5 +1,6 @@
 jobs:
 - job: 'integration_test_remote_windows'
+  timeoutInMinutes: 0
 
   steps:
   - script: python -m pip install --upgrade pip setuptools

diff --git a/test/pipelines-it-remote.yml b/test/pipelines-it-remote.yml
@@ -1,5 +1,6 @@
 jobs:
 - job: 'integration_test_remote'
+  timeoutInMinutes: 0
 
   steps:
   - script: python3 -m pip install --upgrade pip setuptools --user