Refactor integration test (step 10) - miscellaneous docs and help messages (#4901)

docs/source/experiment/training_service/remote.rst

@@ -106,7 +106,7 @@ For example, with anaconda you can specify:
 Configure shared storage
 ^^^^^^^^^^^^^^^^^^^^^^^^^^^
 
-Remote training service support shared storage, which can help use your own storage during using NNI. Follow the guide `here <./shared_storage.rst>`__ to learn how to use shared storage.
+Remote training service support shared storage, which can help use your own storage during using NNI. Follow the guide :doc:`here <./shared_storage>` to learn how to use shared storage.
 
 Monitor via TensorBoard
 ^^^^^^^^^^^^^^^^^^^^^^^

nni/experiment/launcher.py

@@ -222,6 +222,9 @@ def get_stopped_experiment_config_json(exp_id, exp_dir=None):
             _logger.error('Id %s not exist!', exp_id)
             return None
         if experiment_metadata['status'] != 'STOPPED':
-            _logger.error('Only stopped experiments can be resumed or viewed!')
+            _logger.error(
+                'Only stopped experiments can be resumed or viewed! But retrieved metadata for %s is:\n%s',
+                exp_id, experiment_metadata
+            )
             return None
         return Config(exp_id, experiment_metadata['logDir']).get_config()

pipelines/full-test-compression.yml

@@ -8,7 +8,7 @@ pr: none
 jobs:
 - job: linux
   pool: nni-it
-  timeoutInMinutes: 120
+  timeoutInMinutes: 60
 
   steps:
   - template: templates/install-dependencies.yml

pipelines/full-test-hpo.yml

@@ -8,7 +8,7 @@ pr: none
 jobs:
 - job: linux
   pool: nni-it
-  timeoutInMinutes: 120
+  timeoutInMinutes: 60
 
   steps:
   - template: templates/install-dependencies.yml
@@ -18,12 +18,7 @@ jobs:
 
   - template: templates/install-nni.yml
 
-  - script: |
-      set -e
-      cd examples/tuners/customized_tuner
-      python setup.py develop
-      nnictl algo register --meta meta_file.yml
-    displayName: Install customized tuner
+  - template: templates/install-customized-tuner.yml
 
   - script: |
       set -e

pipelines/full-test-nas.yml

@@ -8,7 +8,7 @@ schedules:
 jobs:
 - job: linux
   pool: nni-it
-  timeoutInMinutes: 120
+  timeoutInMinutes: 60
 
   steps:
   - template: templates/install-dependencies.yml

pipelines/integration-test-hybrid.yml

@@ -8,7 +8,7 @@ schedules:
 jobs:
 - job: hybrid
   pool: nni-it
-  timeoutInMinutes: 120
+  timeoutInMinutes: 90
 
   steps:
   # FIXME: should use GPU here

pipelines/integration-test-local-linux.yml

@@ -8,7 +8,7 @@ schedules:
 jobs:
 - job: linux
   pool: nni-it
-  timeoutInMinutes: 120
+  timeoutInMinutes: 60
 
   steps:
   - template: templates/install-dependencies.yml

pipelines/integration-test-local-windows.yml

@@ -15,7 +15,7 @@ jobs:
   - template: templates/install-dependencies.yml
     parameters:
       platform: windows
-      python_venv: noop
+      python_env: noop
 
   - template: templates/install-nni.yml
     parameters:

pipelines/integration-test-remote-l2l.yml

@@ -13,7 +13,7 @@ pr: none
 jobs:
 - job: remote_linux2linux
   pool: nni-it
-  timeoutInMinutes: 140
+  timeoutInMinutes: 120
 
   steps:
   # FIXME: GPU is not supported yet.

pipelines/integration-test-remote-w2w.yml

@@ -12,14 +12,14 @@ variables:
 jobs:
 - job: remote_windows2windows
   pool: nni-it-windows
-  timeoutInMinutes: 60
+  timeoutInMinutes: 120
 
   steps:
   # FIXME: GPU is not supported yet.
   - template: templates/install-dependencies.yml
     parameters:
       platform: windows
-      python_venv: noop
+      python_env: noop
 
   - template: templates/install-nni.yml
     parameters:

test/ut/retiarii/test_oneshot_utils.py

@@ -110,7 +110,7 @@ def test_concat_loader_with_ddp(
     model = BoringModel()
     trainer = Trainer(
         strategy='ddp',
-        accelerator='auto',
+        accelerator='cpu',
         devices=num_devices,
         replace_sampler_ddp=replace_sampler_ddp,
     )