.item() added to the 'target' variable in fakedataset.py

[avijit9]

solves #3517

[avijit9]

cc @pmeier do I need to add any unit test case for this?

[pmeier]

@avijit9

do I need to add any unit test case for this?

We recently added tests for most of our datasets. In fact, the one for FakeData is failing now, because it checks the return types.

You need to change

vision/test/test_datasets.py

Line 1327 in 85fcea6

FEATURE_TYPES = (PIL.Image.Image, torch.Tensor)

to (PIL.Image.Image, int). This indicates that the test should now expect an int opposed to a torch.Tensor as label type.

[pmeier]

Thanks for the PR @avijit9! LGTM when CI is happy.

[codecov]

Codecov Report

Merging #3587 (71d9b3e) into master (c00c4d3) will not change coverage.
The diff coverage is 100.00%.

❗ Current head 71d9b3e differs from pull request most recent head c3df7f4. Consider uploading reports for the commit c3df7f4 to get more accurate results

@@           Coverage Diff           @@
##           master    #3587   +/-   ##
=======================================
  Coverage   79.07%   79.07%           
=======================================
  Files         105      105           
  Lines        9786     9786           
  Branches     1572     1572           
=======================================
  Hits         7738     7738           
  Misses       1567     1567           
  Partials      481      481           

Impacted Files	Coverage Δ
torchvision/transforms/transforms.py	84.30% <ø> (ø)
torchvision/datasets/fakedata.py	92.59% <100.00%> (ø)

---

Continue to review full report at Codecov.

Legend - Click here to learn more
Δ = absolute <relative> (impact), ø = not affected, ? = missing data
Powered by Codecov. Last update 5f0532d...c3df7f4. Read the comment docs.