Fix setup and pre-commit.ci (#209)

* Fix getting long description in setup.py

* [pre-commit.ci] auto fixes from pre-commit.com hooks

* Small fixes to the setup and pyproject

* [pre-commit.ci] auto fixes from pre-commit.com hooks

* Fixing format

* [pre-commit.ci] auto fixes from pre-commit.com hooks

for more information, see https://pre-commit.ci

* Fix pre-commit.ci

Co-authored-by: pre-commit-ci[bot] <66853113+pre-commit-ci[bot]@users.noreply.github.com>

.pre-commit-config.yaml

@@ -1,5 +1,7 @@
-repos:
+default_language_version:
+  python: python3.8
 
+repos:
   - repo: https://github.com/pre-commit/pre-commit-hooks
     rev: v4.0.1
     hooks:

deployment/ncnn/export_onnx.py

@@ -7,9 +7,7 @@
 
 def get_parser():
     parser = argparse.ArgumentParser()
-    parser.add_argument(
-        "--weights", type=str, default="./yolov5s.pt", help="weights path"
-    )
+    parser.add_argument("--weights", type=str, default="./yolov5s.pt", help="weights path")
     parser.add_argument(
         "--output_path",
         type=str,
@@ -25,12 +23,8 @@ def get_parser():
     )
     parser.add_argument("--num_classes", type=int, default=80, help="number of classes")
     parser.add_argument("--batch_size", type=int, default=1, help="batch size")
-    parser.add_argument(
-        "--device", default="cpu", help="cuda device, i.e. 0 or 0,1,2,3 or cpu"
-    )
-    parser.add_argument(
-        "--half", action="store_true", help="FP16 half-precision export"
-    )
+    parser.add_argument("--device", default="cpu", help="cuda device, i.e. 0 or 0,1,2,3 or cpu")
+    parser.add_argument("--half", action="store_true", help="FP16 half-precision export")
     parser.add_argument("--dynamic", action="store_true", help="ONNX: dynamic axes")
     parser.add_argument("--simplify", action="store_true", help="ONNX: simplify model")
     parser.add_argument("--opset", type=int, default=11, help="ONNX: opset version")

deployment/ncnn/tools/yolort_deploy_friendly.py

@@ -61,17 +61,13 @@ def yolov5s_r40_deploy_ncnn(
     width_multiple = 0.5
     version = "r4.0"
 
-    backbone = darknet_pan_backbone(
-        backbone_name, depth_multiple, width_multiple, version=version
-    )
+    backbone = darknet_pan_backbone(backbone_name, depth_multiple, width_multiple, version=version)
 
     model = YOLODeployFriendly(backbone, num_classes, **kwargs)
     if pretrained:
         if model_urls.get(weights_name, None) is None:
             raise ValueError(f"No checkpoint is available for model {weights_name}")
-        state_dict = load_state_dict_from_url(
-            model_urls[weights_name], progress=progress
-        )
+        state_dict = load_state_dict_from_url(model_urls[weights_name], progress=progress)
         model.load_state_dict(state_dict)
 
     del model.anchor_generator

docs/source/conf.py

@@ -68,7 +68,7 @@
 # List of patterns, relative to source directory, that match files and
 # directories to ignore when looking for source files.
 # This pattern also affects html_static_path and html_extra_path.
-exclude_patterns = [u"_build", "**.ipynb_checkpoints", "Thumbs.db", ".DS_Store"]
+exclude_patterns = ["_build", "**.ipynb_checkpoints", "Thumbs.db", ".DS_Store"]
 
 pygments_style = "sphinx"
 

pyproject.toml

@@ -0,0 +1,8 @@
+[tool.usort]
+
+first_party_detection = false
+
+[tool.black]
+
+line-length = 110
+target-version = ["py36"]

setup.cfg

@@ -3,6 +3,7 @@ universal=1
 
 [metadata]
 license_file = LICENSE
+description-file = README.md
 
 [pep8]
 max-line-length = 110

setup.py

@@ -7,6 +7,7 @@
 Adopted from TorchVision, see:
 https://github.com/pytorch/vision/blob/master/setup.py
 """
+import io
 import subprocess
 from pathlib import Path
 
@@ -19,11 +20,7 @@
 sha = "Unknown"
 
 try:
-    sha = (
-        subprocess.check_output(["git", "rev-parse", "HEAD"], cwd=PATH_ROOT)
-        .decode("ascii")
-        .strip()
-    )
+    sha = subprocess.check_output(["git", "rev-parse", "HEAD"], cwd=PATH_ROOT).decode("ascii").strip()
 except Exception:
     pass
 
@@ -38,9 +35,13 @@ def write_version_file():
         f.write("    cuda = _check_cuda_version()\n")
 
 
-def load_requirements(
-    path_dir=PATH_ROOT, file_name="requirements.txt", comment_char="#"
-):
+def get_long_description():
+    # Get the long description from the README file
+    with io.open(str(PATH_ROOT / "README.md"), encoding="utf-8") as f:
+        return f.read()
+
+
+def load_requirements(path_dir=PATH_ROOT, file_name="requirements.txt", comment_char="#"):
     with open(path_dir / file_name, "r", encoding="utf-8", errors="ignore") as file:
         lines = [ln.rstrip() for ln in file.readlines() if not ln.startswith("#")]
     reqs = []
@@ -59,16 +60,13 @@ def load_requirements(
 
     write_version_file()
 
-    # Get the long description from the README file
-    long_description = (PATH_ROOT / "README.md").read_text(encoding="utf-8")
-
     setup(
         name=PACKAGE_NAME,
         version=VERSION,
-        description="Yet Another YOLOv5 and its Additional Runtime Stack",
+        description="Yet another yolov5 and its runtime stack for specialized accelerators.",
         author="Zhiqiang Wang",
         author_email="[email protected]",
-        long_description=long_description,
+        long_description=get_long_description(),
         long_description_content_type="text/markdown",
         url="https://github.com/zhiqwang/yolov5-rt-stack",
         license="GPL-3.0",

test/test_data_pipeline.py

@@ -27,9 +27,7 @@ def test_get_dataset():
 
 def test_get_dataloader():
     batch_size = 8
-    data_loader = data_helper.get_dataloader(
-        data_root="data-bin", mode="train", batch_size=batch_size
-    )
+    data_loader = data_helper.get_dataloader(data_root="data-bin", mode="train", batch_size=batch_size)
     # Test the dataloader
     images, targets = next(iter(data_loader))
 
@@ -67,7 +65,5 @@ def test_prepare_coco128():
     data_path = Path("data-bin")
     coco128_dirname = "coco128"
     data_helper.prepare_coco128(data_path, dirname=coco128_dirname)
-    annotation_file = (
-        data_path / coco128_dirname / "annotations" / "instances_train2017.json"
-    )
+    annotation_file = data_path / coco128_dirname / "annotations" / "instances_train2017.json"
     assert annotation_file.is_file()

test/test_engine.py

@@ -110,9 +110,7 @@ def test_test_epoch_end():
     data_path = Path("data-bin")
     coco128_dirname = "coco128"
     data_helper.prepare_coco128(data_path, dirname=coco128_dirname)
-    annotation_file = (
-        data_path / coco128_dirname / "annotations" / "instances_train2017.json"
-    )
+    annotation_file = data_path / coco128_dirname / "annotations" / "instances_train2017.json"
 
     # Get dataloader to test
     val_dataloader = data_helper.get_dataloader(data_root=data_path, mode="val")

test/test_models.py

@@ -54,9 +54,7 @@ def get_export_import_copy(m):
             results_from_imported = m_import(*args)
         tol = 3e-4
         try:
-            torch.testing.assert_close(
-                results, results_from_imported, atol=tol, rtol=tol
-            )
+            torch.testing.assert_close(results, results_from_imported, atol=tol, rtol=tol)
         except ValueError:
             # custom check for the models that return named tuples:
             # we compare field by field while ignoring None as assert_close can't handle None
@@ -140,9 +138,7 @@ def _get_feature_shapes(self, height, width, width_multiple=0.5, use_p6=False):
 
         return [(c, height // s, width // s) for (c, s) in zip(in_channels, strides)]
 
-    def _get_feature_maps(
-        self, batch_size, height, width, width_multiple=0.5, use_p6=False
-    ):
+    def _get_feature_maps(self, batch_size, height, width, width_multiple=0.5, use_p6=False):
         feature_shapes = self._get_feature_shapes(
             height,
             width,
@@ -152,9 +148,7 @@ def _get_feature_maps(
         feature_maps = [torch.rand(batch_size, *f_shape) for f_shape in feature_shapes]
         return feature_maps
 
-    def _get_head_outputs(
-        self, batch_size, height, width, width_multiple=0.5, use_p6=False
-    ):
+    def _get_head_outputs(self, batch_size, height, width, width_multiple=0.5, use_p6=False):
         feature_shapes = self._get_feature_shapes(
             height,
             width,
@@ -163,9 +157,7 @@ def _get_head_outputs(
         )
 
         num_outputs = self.num_outputs
-        head_shapes = [
-            (batch_size, 3, *f_shape[1:], num_outputs) for f_shape in feature_shapes
-        ]
+        head_shapes = [(batch_size, 3, *f_shape[1:], num_outputs) for f_shape in feature_shapes]
         head_outputs = [torch.rand(*h_shape) for h_shape in head_shapes]
 
         return head_outputs
@@ -202,9 +194,7 @@ def _init_test_backbone_with_pan(
             (0.67, 0.75, "r6.0", False, False),
         ],
     )
-    @pytest.mark.parametrize(
-        "batch_size, height, width", [(4, 448, 320), (2, 384, 640)]
-    )
+    @pytest.mark.parametrize("batch_size, height, width", [(4, 448, 320), (2, 384, 640)])
     def test_backbone_with_pan(
         self,
         depth_multiple,
@@ -216,9 +206,7 @@ def test_backbone_with_pan(
         height,
         width,
     ):
-        out_shape = self._get_feature_shapes(
-            height, width, width_multiple=width_multiple, use_p6=use_p6
-        )
+        out_shape = self._get_feature_shapes(height, width, width_multiple=width_multiple, use_p6=use_p6)
 
         x = torch.rand(batch_size, 3, height, width)
         model = self._init_test_backbone_with_pan(
@@ -243,9 +231,7 @@ def _init_test_anchor_generator(self, use_p6=False):
         "width_multiple, use_p6",
         [(0.5, False), (0.5, True)],
     )
-    @pytest.mark.parametrize(
-        "batch_size, height, width", [(4, 448, 320), (2, 384, 640)]
-    )
+    @pytest.mark.parametrize("batch_size, height, width", [(4, 448, 320), (2, 384, 640)])
     def test_anchor_generator(self, width_multiple, use_p6, batch_size, height, width):
         feature_maps = self._get_feature_maps(
             batch_size, height, width, width_multiple=width_multiple, use_p6=use_p6
@@ -344,15 +330,9 @@ def test_torchscript(arch):
     out = model(x)
     out_script = scripted_model(x)
 
-    torch.testing.assert_close(
-        out[0]["scores"], out_script[1][0]["scores"], rtol=0, atol=0
-    )
-    torch.testing.assert_close(
-        out[0]["labels"], out_script[1][0]["labels"], rtol=0, atol=0
-    )
-    torch.testing.assert_close(
-        out[0]["boxes"], out_script[1][0]["boxes"], rtol=0, atol=0
-    )
+    torch.testing.assert_close(out[0]["scores"], out_script[1][0]["scores"], rtol=0, atol=0)
+    torch.testing.assert_close(out[0]["labels"], out_script[1][0]["labels"], rtol=0, atol=0)
+    torch.testing.assert_close(out[0]["boxes"], out_script[1][0]["boxes"], rtol=0, atol=0)
 
 
 @pytest.mark.parametrize(
@@ -403,12 +383,6 @@ def test_load_from_yolov5(
     model.eval()
     out = model.predict(img_path)
 
-    torch.testing.assert_close(
-        out_from_yolov5[0]["scores"], out[0]["scores"], rtol=0, atol=0
-    )
-    torch.testing.assert_close(
-        out_from_yolov5[0]["labels"], out[0]["labels"], rtol=0, atol=0
-    )
-    torch.testing.assert_close(
-        out_from_yolov5[0]["boxes"], out[0]["boxes"], rtol=0, atol=0
-    )
+    torch.testing.assert_close(out_from_yolov5[0]["scores"], out[0]["scores"], rtol=0, atol=0)
+    torch.testing.assert_close(out_from_yolov5[0]["labels"], out[0]["labels"], rtol=0, atol=0)
+    torch.testing.assert_close(out_from_yolov5[0]["boxes"], out[0]["boxes"], rtol=0, atol=0)

test/test_models_anchor_utils.py

@@ -18,13 +18,9 @@ def test_anchor_generator(self):
         model.eval()
         anchors = model(features)
 
-        expected_anchor_output = torch.tensor(
-            [[-0.5, -0.5], [0.5, -0.5], [-0.5, 0.5], [0.5, 0.5]]
-        )
+        expected_anchor_output = torch.tensor([[-0.5, -0.5], [0.5, -0.5], [-0.5, 0.5], [0.5, 0.5]])
         expected_wh_output = torch.tensor([[4.0], [4.0], [4.0], [4.0]])
-        expected_xy_output = torch.tensor(
-            [[6.0, 14.0], [6.0, 14.0], [6.0, 14.0], [6.0, 14.0]]
-        )
+        expected_xy_output = torch.tensor([[6.0, 14.0], [6.0, 14.0], [6.0, 14.0], [6.0, 14.0]])
 
         assert len(anchors) == 3
         assert tuple(anchors[0].shape) == (4, 2)

test/test_models_common.py

@@ -3,9 +3,7 @@
 from yolort.v5 import focus_transform, space_to_depth
 
 
-@pytest.mark.parametrize(
-    "n, b, h, w", [(1, 3, 480, 640), (4, 3, 416, 320), (4, 3, 320, 416)]
-)
+@pytest.mark.parametrize("n, b, h, w", [(1, 3, 480, 640), (4, 3, 416, 320), (4, 3, 320, 416)])
 def test_space_to_depth(n, b, h, w):
     tensor_input = torch.rand((n, b, h, w))
     out1 = focus_transform(tensor_input)

test/test_models_transform.py

@@ -17,9 +17,5 @@ def test_yolo_transform():
     assert targets.shape[1] == 6
 
     # Test annotations after transformation
-    torch.testing.assert_close(
-        annotations[0]["boxes"], annotations_copy[0]["boxes"], rtol=0, atol=0
-    )
-    torch.testing.assert_close(
-        annotations[1]["boxes"], annotations_copy[1]["boxes"], rtol=0, atol=0
-    )
+    torch.testing.assert_close(annotations[0]["boxes"], annotations_copy[0]["boxes"], rtol=0, atol=0)
+    torch.testing.assert_close(annotations[1]["boxes"], annotations_copy[1]["boxes"], rtol=0, atol=0)

test/test_onnx.py

@@ -62,9 +62,7 @@ def run_model(
                 test_ouputs = model(*test_inputs)
                 if isinstance(test_ouputs, Tensor):
                     test_ouputs = (test_ouputs,)
-            self.ort_validate(
-                onnx_io, test_inputs, test_ouputs, tolerate_small_mismatch
-            )
+            self.ort_validate(onnx_io, test_inputs, test_ouputs, tolerate_small_mismatch)
 
     def ort_validate(self, onnx_io, inputs, outputs, tolerate_small_mismatch=False):
 
@@ -82,16 +80,12 @@ def to_numpy(tensor):
 
         ort_session = onnxruntime.InferenceSession(onnx_io.getvalue())
         # compute onnxruntime output prediction
-        ort_inputs = dict(
-            (ort_session.get_inputs()[i].name, inpt) for i, inpt in enumerate(inputs)
-        )
+        ort_inputs = dict((ort_session.get_inputs()[i].name, inpt) for i, inpt in enumerate(inputs))
         ort_outs = ort_session.run(None, ort_inputs)
 
         for i in range(0, len(outputs)):
             try:
-                torch.testing.assert_close(
-                    outputs[i], ort_outs[i], rtol=1e-03, atol=1e-05
-                )
+                torch.testing.assert_close(outputs[i], ort_outs[i], rtol=1e-03, atol=1e-05)
             except AssertionError as error:
                 if tolerate_small_mismatch:
                     self.assertIn("(0.00%)", str(error), str(error))

test/test_utils.py

@@ -105,9 +105,7 @@ def test_load_from_ultralytics_voc(
 
     torch.testing.assert_allclose(out_from_yolort[0]["boxes"], out_from_yolov5[:, :4])
     torch.testing.assert_allclose(out_from_yolort[0]["scores"], out_from_yolov5[:, 4])
-    torch.testing.assert_allclose(
-        out_from_yolort[0]["labels"], out_from_yolov5[:, 5].to(dtype=torch.int64)
-    )
+    torch.testing.assert_allclose(out_from_yolort[0]["labels"], out_from_yolov5[:, 5].to(dtype=torch.int64))
 
 
 def test_read_image_to_tensor():
@@ -194,13 +192,8 @@ def test_feature_extractor(batch_size, height, width, arch, width_multiple):
     in_channels = [int(gw * width_multiple) for gw in grow_widths]
     strides = [8, 16, 32]
     num_outputs = 85
-    expected_features = [
-        (batch_size, inc, height // s, width // s)
-        for inc, s in zip(in_channels, strides)
-    ]
-    expected_head_outputs = [
-        (batch_size, c, height // s, width // s, num_outputs) for s in strides
-    ]
+    expected_features = [(batch_size, inc, height // s, width // s) for inc, s in zip(in_channels, strides)]
+    expected_head_outputs = [(batch_size, c, height // s, width // s, num_outputs) for s in strides]
 
     model = models.__dict__[arch]()
     model = model.train()

tools/eval_metric.py

@@ -38,9 +38,7 @@ def get_parser():
         default="./data-bin/coco128/images/train2017",
         help="Root path of the dataset containing images",
     )
-    parser.add_argument(
-        "--annotation_path", default=None, help="Path of the annotation file"
-    )
+    parser.add_argument("--annotation_path", default=None, help="Path of the annotation file")
     parser.add_argument(
         "--eval_type",
         default="yolov5",

tools/export_model.py

@@ -45,9 +45,7 @@ def get_parser():
         help="Image size for evaluation (default: 640).",
     )
     parser.add_argument("--batch_size", default=1, type=int, help="Batch size.")
-    parser.add_argument(
-        "--opset", default=DEFAULT_OPSET, type=int, help="opset_version"
-    )
+    parser.add_argument("--opset", default=DEFAULT_OPSET, type=int, help="opset_version")
     parser.add_argument("--simplify", action="store_true", help="ONNX: simplify model.")
 
     return parser