Merge branch 'master' into vfdev-5/issue-2651-contributing

Author: vfdev-5

.circleci/config.yml

@@ -132,6 +132,19 @@ jobs:
             sudo mv clang-format /opt/clang-format
             ./travis-scripts/run-clang-format/run-clang-format.py -r torchvision/csrc --clang-format-executable /opt/clang-format
 
+  torchhub_test:
+    docker:
+      - image: circleci/python:3.7
+    steps:
+      - checkout
+      - run:
+          command: |
+            pip install --user --progress-bar off numpy
+            pip install --user --progress-bar off --pre torch -f https://download.pytorch.org/whl/nightly/cpu/torch_nightly.html
+            # need to install torchvision dependencies due to transitive imports
+            pip install --user --progress-bar off --editable .
+            python test/test_hub.py
+
   binary_linux_wheel:
     <<: *binary_common
     docker:
@@ -1060,6 +1073,7 @@ workflows:
       - python_lint
       - python_type_check
       - clang_format
+      - torchhub_test
 
   unittest:
     jobs:
@@ -1174,6 +1188,7 @@ workflows:
       - python_lint
       - python_type_check
       - clang_format
+      - torchhub_test
       - binary_linux_wheel:
           cu_version: cpu
           filters:

.circleci/config.yml.in

@@ -132,6 +132,19 @@ jobs:
             sudo mv clang-format /opt/clang-format
             ./travis-scripts/run-clang-format/run-clang-format.py -r torchvision/csrc --clang-format-executable /opt/clang-format
 
+  torchhub_test:
+    docker:
+      - image: circleci/python:3.7
+    steps:
+      - checkout
+      - run:
+          command: |
+            pip install --user --progress-bar off numpy
+            pip install --user --progress-bar off --pre torch -f https://download.pytorch.org/whl/nightly/cpu/torch_nightly.html
+            # need to install torchvision dependencies due to transitive imports
+            pip install --user --progress-bar off --editable .
+            python test/test_hub.py
+
   binary_linux_wheel:
     <<: *binary_common
     docker:
@@ -686,6 +699,7 @@ workflows:
       - python_lint
       - python_type_check
       - clang_format
+      - torchhub_test
 
   unittest:
     jobs:
@@ -702,6 +716,7 @@ workflows:
       - python_lint
       - python_type_check
       - clang_format
+      - torchhub_test
       {{ build_workflows(prefix="nightly_", filter_branch="nightly", upload=True) }}
   docker_build:
     triggers:

.github/ISSUE_TEMPLATE/---bug-report.md renamed to .github/ISSUE_TEMPLATE/bug-report.md

@@ -21,7 +21,7 @@ def _encode(v):
         return torch.tensor(v, dtype=torch.int32).numpy().tobytes()[::-1]
 
     def _make_image_file(filename, num_images):
-        img = torch.randint(0, 255, size=(28 * 28 * num_images,), dtype=torch.uint8)
+        img = torch.randint(0, 256, size=(28 * 28 * num_images,), dtype=torch.uint8)
         with open(filename, "wb") as f:
             f.write(_encode(2051))  # magic header
             f.write(_encode(num_images))

.github/ISSUE_TEMPLATE/---documentation.md renamed to .github/ISSUE_TEMPLATE/documentation.md

@@ -29,7 +29,7 @@ def get_list_of_videos(num_videos=5, sizes=None, fps=None):
                 f = 5
             else:
                 f = fps[i]
-            data = torch.randint(0, 255, (size, 300, 400, 3), dtype=torch.uint8)
+            data = torch.randint(0, 256, (size, 300, 400, 3), dtype=torch.uint8)
             name = os.path.join(tmp_dir, "{}.mp4".format(i))
             names.append(name)
             io.write_video(name, data, fps=f)

.github/ISSUE_TEMPLATE/--feature-request.md renamed to .github/ISSUE_TEMPLATE/feature-request.md

@@ -22,7 +22,7 @@ def get_list_of_videos(num_videos=5, sizes=None, fps=None):
                 f = 5
             else:
                 f = fps[i]
-            data = torch.randint(0, 255, (size, 300, 400, 3), dtype=torch.uint8)
+            data = torch.randint(0, 256, (size, 300, 400, 3), dtype=torch.uint8)
             name = os.path.join(tmp_dir, "{}.mp4".format(i))
             names.append(name)
             io.write_video(name, data, fps=f)

.github/ISSUE_TEMPLATE/-questions-help-support.md renamed to .github/ISSUE_TEMPLATE/questions-help-support.md

@@ -180,7 +180,7 @@ def test_center_crop(self):
         self._test_op(
             "center_crop", "CenterCrop", fn_kwargs=fn_kwargs, meth_kwargs=meth_kwargs
         )
-        tensor = torch.randint(0, 255, (3, 10, 10), dtype=torch.uint8, device=self.device)
+        tensor = torch.randint(0, 256, (3, 10, 10), dtype=torch.uint8, device=self.device)
         # Test torchscript of transforms.CenterCrop with size as int
         f = T.CenterCrop(size=5)
         scripted_fn = torch.jit.script(f)
@@ -294,7 +294,7 @@ def test_resize(self):
         self.assertEqual(y.shape[2], int(38 * 46 / 32))
 
         tensor, _ = self._create_data(height=34, width=36, device=self.device)
-        batch_tensors = torch.randint(0, 255, size=(4, 3, 44, 56), dtype=torch.uint8, device=self.device)
+        batch_tensors = torch.randint(0, 256, size=(4, 3, 44, 56), dtype=torch.uint8, device=self.device)
         script_fn = torch.jit.script(F.resize)
 
         for dt in [None, torch.float32, torch.float64]:
@@ -323,8 +323,8 @@ def test_resize(self):
             script_fn.save(os.path.join(tmp_dir, "t_resize.pt"))
 
     def test_resized_crop(self):
-        tensor = torch.randint(0, 255, size=(3, 44, 56), dtype=torch.uint8, device=self.device)
-        batch_tensors = torch.randint(0, 255, size=(4, 3, 44, 56), dtype=torch.uint8, device=self.device)
+        tensor = torch.randint(0, 256, size=(3, 44, 56), dtype=torch.uint8, device=self.device)
+        batch_tensors = torch.randint(0, 256, size=(4, 3, 44, 56), dtype=torch.uint8, device=self.device)
 
         for scale in [(0.7, 1.2), [0.7, 1.2]]:
             for ratio in [(0.75, 1.333), [0.75, 1.333]]:
@@ -341,8 +341,8 @@ def test_resized_crop(self):
             s_transform.save(os.path.join(tmp_dir, "t_resized_crop.pt"))
 
     def test_random_affine(self):
-        tensor = torch.randint(0, 255, size=(3, 44, 56), dtype=torch.uint8, device=self.device)
-        batch_tensors = torch.randint(0, 255, size=(4, 3, 44, 56), dtype=torch.uint8, device=self.device)
+        tensor = torch.randint(0, 256, size=(3, 44, 56), dtype=torch.uint8, device=self.device)
+        batch_tensors = torch.randint(0, 256, size=(4, 3, 44, 56), dtype=torch.uint8, device=self.device)
 
         for shear in [15, 10.0, (5.0, 10.0), [-15, 15], [-10.0, 10.0, -11.0, 11.0]]:
             for scale in [(0.7, 1.2), [0.7, 1.2]]:
@@ -363,8 +363,8 @@ def test_random_affine(self):
             s_transform.save(os.path.join(tmp_dir, "t_random_affine.pt"))
 
     def test_random_rotate(self):
-        tensor = torch.randint(0, 255, size=(3, 44, 56), dtype=torch.uint8, device=self.device)
-        batch_tensors = torch.randint(0, 255, size=(4, 3, 44, 56), dtype=torch.uint8, device=self.device)
+        tensor = torch.randint(0, 256, size=(3, 44, 56), dtype=torch.uint8, device=self.device)
+        batch_tensors = torch.randint(0, 256, size=(4, 3, 44, 56), dtype=torch.uint8, device=self.device)
 
         for center in [(0, 0), [10, 10], None, (56, 44)]:
             for expand in [True, False]:
@@ -383,8 +383,8 @@ def test_random_rotate(self):
             s_transform.save(os.path.join(tmp_dir, "t_random_rotate.pt"))
 
     def test_random_perspective(self):
-        tensor = torch.randint(0, 255, size=(3, 44, 56), dtype=torch.uint8, device=self.device)
-        batch_tensors = torch.randint(0, 255, size=(4, 3, 44, 56), dtype=torch.uint8, device=self.device)
+        tensor = torch.randint(0, 256, size=(3, 44, 56), dtype=torch.uint8, device=self.device)
+        batch_tensors = torch.randint(0, 256, size=(4, 3, 44, 56), dtype=torch.uint8, device=self.device)
 
         for distortion_scale in np.linspace(0.1, 1.0, num=20):
             for interpolation in [NEAREST, BILINEAR]:

.travis.yml

@@ -1,13 +1,13 @@
 #include <ATen/ATen.h>
 #include <torch/script.h>
 #include <torch/torch.h>
-#include <torchvision/roi_align.h>
 #include <torchvision/nms.h>
+#include <torchvision/roi_align.h>
 
 #ifdef _WIN32
 // Windows only
 // This is necessary until operators are automatically registered on include
-static auto _nms = &vision::ops::nms_cpu;
+static auto _nms = &vision::ops::nms;
 #endif
 
 int main() {

test/fakedata_generation.py

@@ -66,7 +66,8 @@
 // modified from
 // https://github.com/open-mmlab/mmdetection/blob/master/mmdet/ops/dcn/src/deform_conv_cuda.cpp
 
-#include "deform_conv2d_kernel.h"
+#include <ATen/ATen.h>
+#include <torch/library.h>
 
 namespace vision {
 namespace ops {
@@ -852,9 +853,7 @@ at::Tensor backward_gradient_parameters(
   return grad_weight;
 }
 
-} // namespace
-
-at::Tensor deform_conv2d_forward_cpu(
+at::Tensor deform_conv2d_forward_kernel(
     const at::Tensor& input,
     const at::Tensor& weight,
     const at::Tensor& offset,
@@ -1070,7 +1069,7 @@ at::Tensor deform_conv2d_forward_cpu(
 }
 
 std::tuple<at::Tensor, at::Tensor, at::Tensor, at::Tensor, at::Tensor>
-deform_conv2d_backward_cpu(
+deform_conv2d_backward_kernel(
     const at::Tensor& grad_out,
     const at::Tensor& input,
     const at::Tensor& weight,
@@ -1141,5 +1140,12 @@ deform_conv2d_backward_cpu(
       grad_input, grad_weight, grad_offset, grad_mask, grad_bias);
 }
 
+} // namespace
+
+TORCH_LIBRARY_IMPL(torchvision, CPU, m) {
+  m.impl("deform_conv2d", deform_conv2d_forward_kernel);
+  m.impl("_deform_conv2d_backward", deform_conv2d_backward_kernel);
+}
+
 } // namespace ops
 } // namespace vision

test/test_datasets_samplers.py

@@ -1,4 +1,5 @@
-#include "nms_kernel.h"
+#include <ATen/ATen.h>
+#include <torch/library.h>
 
 namespace vision {
 namespace ops {
@@ -74,9 +75,7 @@ at::Tensor nms_kernel_impl(
   return keep_t.narrow(/*dim=*/0, /*start=*/0, /*length=*/num_to_keep);
 }
 
-} // namespace
-
-at::Tensor nms_cpu(
+at::Tensor nms_kernel(
     const at::Tensor& dets,
     const at::Tensor& scores,
     double iou_threshold) {
@@ -101,11 +100,17 @@ at::Tensor nms_cpu(
 
   auto result = at::empty({0}, dets.options());
 
-  AT_DISPATCH_FLOATING_TYPES(dets.scalar_type(), "nms_cpu", [&] {
+  AT_DISPATCH_FLOATING_TYPES(dets.scalar_type(), "nms_kernel", [&] {
     result = nms_kernel_impl<scalar_t>(dets, scores, iou_threshold);
   });
   return result;
 }
 
+} // namespace
+
+TORCH_LIBRARY_IMPL(torchvision, CPU, m) {
+  m.impl("nms", nms_kernel);
+}
+
 } // namespace ops
 } // namespace vision