Fix PT 1.8 compatibility issue (#580)

* fix torch._six PT 1.8 comp issue

* update torch vision

* update torchvision version

* revert back torch versions

* add user agent

* fix expected od loss tensor shape

* fix unused imports

* move

* format

Co-authored-by: Jirka Borovec <jirka.borovec@seznam.cz>

.gitignore

@@ -2,6 +2,7 @@
 __pycache__/
 *.py[cod]
 *$py.class
+data/
 
 # C extensions
 *.so
@@ -163,3 +164,8 @@ tests/test-reports/
 
 # MLflow
 mlruns/
+
+
+# Downloaded Data through Scripts
+VOCdevkit/
+*.tar

pl_bolts/datamodules/async_dataloader.py

@@ -1,10 +1,11 @@
+import collections.abc as container_abcs
 import re
 from queue import Queue
 from threading import Thread
 from typing import Any, Optional, Union
 
 import torch
-from torch._six import container_abcs, string_classes
+from torch._six import string_classes
 from torch.utils.data import DataLoader, Dataset
 
 

pl_bolts/datasets/__init__.py

@@ -1,3 +1,5 @@
+import urllib
+
 from pl_bolts.datasets.base_dataset import LightDataset
 from pl_bolts.datasets.cifar10_dataset import CIFAR10, TrialCIFAR10
 from pl_bolts.datasets.concat_dataset import ConcatDataset
@@ -31,3 +33,8 @@
     "CIFAR10Mixed",
     "SSLDatasetMixin",
 ]
+
+# TorchVision hotfix https://github.com/pytorch/vision/issues/1938
+opener = urllib.request.build_opener()
+opener.addheaders = [('User-agent', 'Mozilla/5.0')]
+urllib.request.install_opener(opener)

pl_bolts/datasets/imagenet_dataset.py

@@ -3,6 +3,7 @@
 import os
 import sys
 import shutil
+import sys  # noqa F401
 import tarfile
 import tempfile
 import zipfile
@@ -248,6 +249,8 @@ def extract_archive(from_path, to_path=None, remove_finished=False):
     if to_path is None:
         to_path = os.path.dirname(from_path)
 
+    PY3 = sys.version_info[0] == 3
+
     if _is_tar(from_path):
         with tarfile.open(from_path, 'r') as tar:
             tar.extractall(path=to_path)

tests/losses/test_object_detection.py

@@ -10,16 +10,16 @@
 
 @pytest.mark.parametrize(
     "preds, target, expected_loss",
-    [(torch.tensor([[100, 100, 200, 200]]), torch.tensor([[100, 100, 200, 200]]), torch.tensor([0.0]))]
+    [(torch.tensor([[100, 100, 200, 200]]), torch.tensor([[100, 100, 200, 200]]), torch.tensor([[0.0]]))]
 )
 def test_iou_complete_overlap(preds, target, expected_loss):
     torch.testing.assert_allclose(iou_loss(preds, target), expected_loss)
 
 
 @pytest.mark.parametrize(
     "preds, target, expected_loss", [
-        (torch.tensor([[100, 100, 200, 200]]), torch.tensor([[100, 200, 200, 300]]), torch.tensor([1.0])),
-        (torch.tensor([[100, 100, 200, 200]]), torch.tensor([[200, 200, 300, 300]]), torch.tensor([1.0])),
+        (torch.tensor([[100, 100, 200, 200]]), torch.tensor([[100, 200, 200, 300]]), torch.tensor([[1.0]])),
+        (torch.tensor([[100, 100, 200, 200]]), torch.tensor([[200, 200, 300, 300]]), torch.tensor([[1.0]])),
     ]
 )
 def test_iou_no_overlap(preds, target, expected_loss):
@@ -28,16 +28,16 @@ def test_iou_no_overlap(preds, target, expected_loss):
 
 @pytest.mark.parametrize(
     "preds, target, expected_loss",
-    [(torch.tensor([[100, 100, 200, 200]]), torch.tensor([[100, 100, 200, 200]]), torch.tensor([0.0]))]
+    [(torch.tensor([[100, 100, 200, 200]]), torch.tensor([[100, 100, 200, 200]]), torch.tensor([[0.0]]))]
 )
 def test_complete_overlap(preds, target, expected_loss):
     torch.testing.assert_allclose(giou_loss(preds, target), expected_loss)
 
 
 @pytest.mark.parametrize(
     "preds, target, expected_loss", [
-        (torch.tensor([[100, 100, 200, 200]]), torch.tensor([[100, 200, 200, 300]]), torch.tensor([1.0])),
-        (torch.tensor([[100, 100, 200, 200]]), torch.tensor([[200, 200, 300, 300]]), torch.tensor([1.5])),
+        (torch.tensor([[100, 100, 200, 200]]), torch.tensor([[100, 200, 200, 300]]), torch.tensor([[1.0]])),
+        (torch.tensor([[100, 100, 200, 200]]), torch.tensor([[200, 200, 300, 300]]), torch.tensor([[1.5]])),
     ]
 )
 def test_no_overlap(preds, target, expected_loss):

tests/metrics/test_object_detection.py

@@ -10,16 +10,16 @@
 
 @pytest.mark.parametrize(
     "preds, target, expected_iou",
-    [(torch.tensor([[100, 100, 200, 200]]), torch.tensor([[100, 100, 200, 200]]), torch.tensor([1.0]))]
+    [(torch.tensor([[100, 100, 200, 200]]), torch.tensor([[100, 100, 200, 200]]), torch.tensor([[1.0]]))]
 )
 def test_iou_complete_overlap(preds, target, expected_iou):
     torch.testing.assert_allclose(iou(preds, target), expected_iou)
 
 
 @pytest.mark.parametrize(
     "preds, target, expected_iou", [
-        (torch.tensor([[100, 100, 200, 200]]), torch.tensor([[100, 200, 200, 300]]), torch.tensor([0.0])),
-        (torch.tensor([[100, 100, 200, 200]]), torch.tensor([[200, 200, 300, 300]]), torch.tensor([0.0])),
+        (torch.tensor([[100, 100, 200, 200]]), torch.tensor([[100, 200, 200, 300]]), torch.tensor([[0.0]])),
+        (torch.tensor([[100, 100, 200, 200]]), torch.tensor([[200, 200, 300, 300]]), torch.tensor([[0.0]])),
     ]
 )
 def test_iou_no_overlap(preds, target, expected_iou):
@@ -39,16 +39,16 @@ def test_iou_multi(preds, target, expected_iou):
 
 @pytest.mark.parametrize(
     "preds, target, expected_giou",
-    [(torch.tensor([[100, 100, 200, 200]]), torch.tensor([[100, 100, 200, 200]]), torch.tensor([1.0]))]
+    [(torch.tensor([[100, 100, 200, 200]]), torch.tensor([[100, 100, 200, 200]]), torch.tensor([[1.0]]))]
 )
 def test_complete_overlap(preds, target, expected_giou):
     torch.testing.assert_allclose(giou(preds, target), expected_giou)
 
 
 @pytest.mark.parametrize(
     "preds, target, expected_giou", [
-        (torch.tensor([[100, 100, 200, 200]]), torch.tensor([[100, 200, 200, 300]]), torch.tensor([0.0])),
-        (torch.tensor([[100, 100, 200, 200]]), torch.tensor([[200, 200, 300, 300]]), torch.tensor([-0.5])),
+        (torch.tensor([[100, 100, 200, 200]]), torch.tensor([[100, 200, 200, 300]]), torch.tensor([[0.0]])),
+        (torch.tensor([[100, 100, 200, 200]]), torch.tensor([[200, 200, 300, 300]]), torch.tensor([[-0.5]])),
     ]
 )
 def test_no_overlap(preds, target, expected_giou):

tests/models/test_classic_ml.py

@@ -20,8 +20,13 @@ def test_linear_regression_model(tmpdir):
     loader = DataLoader(SklearnDataset(X, y), batch_size=2)
 
     model = LinearRegression(input_dim=2, learning_rate=0.5)
-    trainer = pl.Trainer(max_epochs=300, default_root_dir=tmpdir,
-                         progress_bar_refresh_rate=0, logger=False, checkpoint_callback=False)
+    trainer = pl.Trainer(
+        max_epochs=300,
+        default_root_dir=tmpdir,
+        progress_bar_refresh_rate=0,
+        logger=False,
+        checkpoint_callback=False,
+    )
     trainer.fit(
         model,
         loader,