ci: Fix dataset downloading errors

.github/workflows/ci_test-full.yml

@@ -80,12 +80,33 @@ jobs:
  path: ./datasets
  key: pl-datasets-${{ hashFiles('tests/conftest.py') }}
 
+ - name: Check cached datasets
+ shell: bash
+ run: |
+ ls -lh
+ echo =======
+ ls -lh datasets
+ echo =======
+ ls -lh datasets/*
+
  - name: Tests
  run: |
  # tox --sitepackages
- coverage run --source pl_bolts -m py.test pl_bolts tests -v --junitxml=junit/test-results-${{ runner.os }}-${{ matrix.python-version }}-${{ matrix.requires }}.xml
+ coverage run --source pl_bolts -m pytest pl_bolts tests -v --junitxml=junit/test-results-${{ runner.os }}-${{ matrix.python-version }}-${{ matrix.requires }}.xml
  coverage xml
 
+ - name: Check cached datasets
+ shell: bash
+ run: |
+ ls -lh
+ echo =======
+ ls -lh datasets
+ echo =======
+ find . -name "MNIST"
+ find . -name "mnist"
+ find . -name "CIFAR"
+ find . -name "cifar"
+
  - name: Upload pytest test results
  uses: actions/upload-artifact@v2
  with:

pl_bolts/datasets/cifar10_dataset.py

@@ -37,11 +37,10 @@ class CIFAR10(LightDataset):
  downloaded again.
 
  Examples:
-
  >>> from torchvision import transforms
  >>> from pl_bolts.transforms.dataset_normalizations import cifar10_normalization
  >>> cf10_transforms = transforms.Compose([transforms.ToTensor(), cifar10_normalization()])
- >>> dataset = CIFAR10(download=True, transform=cf10_transforms)
+ >>> dataset = CIFAR10(download=True, transform=cf10_transforms, data_dir="datasets")
  >>> len(dataset)
  50000
  >>> torch.bincount(dataset.targets)
@@ -167,7 +166,7 @@ class TrialCIFAR10(CIFAR10):
  without the torchvision dependency.
  Examples:
 
- >>> dataset = TrialCIFAR10(download=True, num_samples=150, labels=(1, 5, 8))
+ >>> dataset = TrialCIFAR10(download=True, num_samples=150, labels=(1, 5, 8), data_dir="datasets")
  >>> len(dataset)
  450
  >>> sorted(set([d.item() for d in dataset.targets]))

tests/callbacks/test_data_monitor.py

@@ -16,11 +16,11 @@
 )
 @mock.patch("pl_bolts.callbacks.data_monitor.DataMonitorBase.log_histogram")
 def test_base_log_interval_override(
- log_histogram, tmpdir, log_every_n_steps, max_steps, expected_calls
+ log_histogram, tmpdir, log_every_n_steps, max_steps, expected_calls, datadir
 ):
  """ Test logging interval set by log_every_n_steps argument. """
  monitor = TrainingDataMonitor(log_every_n_steps=log_every_n_steps)
- model = LitMNIST(num_workers=0)
+ model = LitMNIST(data_dir=datadir, num_workers=0)
  trainer = Trainer(
  default_root_dir=tmpdir,
  log_every_n_steps=1,
@@ -43,11 +43,11 @@ def test_base_log_interval_override(
 )
 @mock.patch("pl_bolts.callbacks.data_monitor.DataMonitorBase.log_histogram")
 def test_base_log_interval_fallback(
- log_histogram, tmpdir, log_every_n_steps, max_steps, expected_calls
+ log_histogram, tmpdir, log_every_n_steps, max_steps, expected_calls, datadir
 ):
  """ Test that if log_every_n_steps not set in the callback, fallback to what is defined in the Trainer. """
  monitor = TrainingDataMonitor()
- model = LitMNIST(num_workers=0)
+ model = LitMNIST(data_dir=datadir, num_workers=0)
  trainer = Trainer(
  default_root_dir=tmpdir,
  log_every_n_steps=log_every_n_steps,
@@ -81,10 +81,10 @@ def test_base_unsupported_logger_warning(tmpdir):
 
 
 @mock.patch("pl_bolts.callbacks.data_monitor.TrainingDataMonitor.log_histogram")
-def test_training_data_monitor(log_histogram, tmpdir):
+def test_training_data_monitor(log_histogram, tmpdir, datadir):
  """ Test that the TrainingDataMonitor logs histograms of data points going into training_step. """
  monitor = TrainingDataMonitor()
- model = LitMNIST()
+ model = LitMNIST(data_dir=datadir, num_workers=0)
  trainer = Trainer(
  default_root_dir=tmpdir, log_every_n_steps=1, callbacks=[monitor],
  )