add FusedBiasLeakyRelu npu adapter

docs/en/understand_mmcv/ops.md

@@ -2,61 +2,61 @@
 
 We implement common ops used in detection, segmentation, etc.
 
-| Device                       | CPU | CUDA | MLU | MPS | Ascend |
-| ---------------------------- | --- | ---- | --- | --- | ------ |
-| ActiveRotatedFilter          | √   | √    |     |     |        |
-| AssignScoreWithK             |     | √    |     |     |        |
-| BallQuery                    |     | √    |     |     |        |
-| BBoxOverlaps                 |     | √    | √   | √   |        |
-| BorderAlign                  |     | √    |     |     |        |
-| BoxIouRotated                | √   | √    |     |     |        |
-| BoxIouQuadri                 | √   | √    |     |     |        |
-| CARAFE                       |     | √    | √   |     |        |
-| ChamferDistance              |     | √    |     |     |        |
-| CrissCrossAttention          |     | √    |     |     |        |
-| ContourExpand                | √   |      |     |     |        |
-| ConvexIoU                    |     | √    |     |     |        |
-| CornerPool                   |     | √    |     |     |        |
-| Correlation                  |     | √    |     |     |        |
-| Deformable Convolution v1/v2 | √   | √    |     |     |        |
-| Deformable RoIPool           |     | √    | √   |     |        |
-| DiffIoURotated               |     | √    |     |     |        |
-| DynamicScatter               |     | √    |     |     |        |
-| FurthestPointSample          |     | √    |     |     |        |
-| FurthestPointSampleWithDist  |     | √    |     |     |        |
-| FusedBiasLeakyrelu           |     | √    |     |     |        |
-| GatherPoints                 |     | √    |     |     |        |
-| GroupPoints                  |     | √    |     |     |        |
-| Iou3d                        |     | √    | √   |     |        |
-| KNN                          |     | √    |     |     |        |
-| MaskedConv                   |     | √    | √   |     | √      |
-| MergeCells                   |     | √    |     |     |        |
-| MinAreaPolygon               |     | √    |     |     |        |
-| ModulatedDeformConv2d        | √   | √    | √   |     | √      |
-| MultiScaleDeformableAttn     |     | √    | √   |     |        |
-| NMS                          | √   | √    | √   |     | √      |
-| NMSRotated                   | √   | √    |     |     |        |
-| NMSQuadri                    | √   | √    |     |     |        |
-| PixelGroup                   | √   |      |     |     |        |
-| PointsInBoxes                | √   | √    |     |     |        |
-| PointsInPolygons             |     | √    |     |     |        |
-| PSAMask                      | √   | √    | √   |     |        |
-| RotatedFeatureAlign          | √   | √    |     |     |        |
-| RoIPointPool3d               |     | √    | √   |     |        |
-| RoIPool                      |     | √    | √   |     |        |
-| RoIAlignRotated              | √   | √    | √   |     |        |
-| RiRoIAlignRotated            |     | √    |     |     |        |
-| RoIAlign                     | √   | √    | √   |     |        |
-| RoIAwarePool3d               |     | √    | √   |     |        |
-| SAConv2d                     |     | √    |     |     |        |
-| SigmoidFocalLoss             |     | √    | √   |     | √      |
-| SoftmaxFocalLoss             |     | √    |     |     | √      |
-| SoftNMS                      |     | √    |     |     |        |
-| Sparse Convolution           |     | √    |     |     |        |
-| Synchronized BatchNorm       |     | √    |     |     |        |
-| ThreeInterpolate             |     | √    |     |     |        |
-| ThreeNN                      |     | √    | √   |     |        |
-| TINShift                     |     | √    | √   |     |        |
-| UpFirDn2d                    |     | √    |     |     |        |
-| Voxelization                 | √   | √    |     |     |        |
-| PrRoIPool                    |     | √    |     |     |        |
+| Device                       | CPU | CUDA | MLU | MPS | NPU |
+| ---------------------------- | --- | ---- | --- | --- | --- |
+| ActiveRotatedFilter          | √   | √    |     |     |     |
+| AssignScoreWithK             |     | √    |     |     |     |
+| BallQuery                    |     | √    |     |     |     |
+| BBoxOverlaps                 |     | √    | √   | √   |     |
+| BorderAlign                  |     | √    |     |     |     |
+| BoxIouRotated                | √   | √    |     |     |     |
+| BoxIouQuadri                 | √   | √    |     |     |     |
+| CARAFE                       |     | √    | √   |     |     |
+| ChamferDistance              |     | √    |     |     |     |
+| CrissCrossAttention          |     | √    |     |     |     |
+| ContourExpand                | √   |      |     |     |     |
+| ConvexIoU                    |     | √    |     |     |     |
+| CornerPool                   |     | √    |     |     |     |
+| Correlation                  |     | √    |     |     |     |
+| Deformable Convolution v1/v2 | √   | √    |     |     |     |
+| Deformable RoIPool           |     | √    | √   |     |     |
+| DiffIoURotated               |     | √    |     |     |     |
+| DynamicScatter               |     | √    |     |     |     |
+| FurthestPointSample          |     | √    |     |     |     |
+| FurthestPointSampleWithDist  |     | √    |     |     |     |
+| FusedBiasLeakyrelu           |     | √    |     |     | √   |
+| GatherPoints                 |     | √    |     |     |     |
+| GroupPoints                  |     | √    |     |     |     |
+| Iou3d                        |     | √    | √   |     |     |
+| KNN                          |     | √    |     |     |     |
+| MaskedConv                   |     | √    | √   |     | √   |
+| MergeCells                   |     | √    |     |     |     |
+| MinAreaPolygon               |     | √    |     |     |     |
+| ModulatedDeformConv2d        | √   | √    | √   |     | √   |
+| MultiScaleDeformableAttn     |     | √    | √   |     |     |
+| NMS                          | √   | √    | √   |     | √   |
+| NMSRotated                   | √   | √    |     |     |     |
+| NMSQuadri                    | √   | √    |     |     |     |
+| PixelGroup                   | √   |      |     |     |     |
+| PointsInBoxes                | √   | √    |     |     |     |
+| PointsInPolygons             |     | √    |     |     |     |
+| PSAMask                      | √   | √    | √   |     |     |
+| RotatedFeatureAlign          | √   | √    |     |     |     |
+| RoIPointPool3d               |     | √    | √   |     |     |
+| RoIPool                      |     | √    | √   |     |     |
+| RoIAlignRotated              | √   | √    | √   |     |     |
+| RiRoIAlignRotated            |     | √    |     |     |     |
+| RoIAlign                     | √   | √    | √   |     |     |
+| RoIAwarePool3d               |     | √    | √   |     |     |
+| SAConv2d                     |     | √    |     |     |     |
+| SigmoidFocalLoss             |     | √    | √   |     | √   |
+| SoftmaxFocalLoss             |     | √    |     |     | √   |
+| SoftNMS                      |     | √    |     |     |     |
+| Sparse Convolution           |     | √    |     |     |     |
+| Synchronized BatchNorm       |     | √    |     |     |     |
+| ThreeInterpolate             |     | √    |     |     |     |
+| ThreeNN                      |     | √    | √   |     |     |
+| TINShift                     |     | √    | √   |     |     |
+| UpFirDn2d                    |     | √    |     |     |     |
+| Voxelization                 | √   | √    |     |     |     |
+| PrRoIPool                    |     | √    |     |     |     |

docs/zh_cn/understand_mmcv/ops.md

@@ -2,61 +2,61 @@
 
 MMCV 提供了检测、分割等任务中常用的算子
 
-| Device                       | CPU | CUDA | MLU | MPS | Ascend |
-| ---------------------------- | --- | ---- | --- | --- | ------ |
-| ActiveRotatedFilter          | √   | √    |     |     |        |
-| AssignScoreWithK             |     | √    |     |     |        |
-| BallQuery                    |     | √    |     |     |        |
-| BBoxOverlaps                 |     | √    | √   | √   |        |
-| BorderAlign                  |     | √    |     |     |        |
-| BoxIouRotated                | √   | √    |     |     |        |
-| BoxIouQuadri                 | √   | √    |     |     |        |
-| CARAFE                       |     | √    | √   |     |        |
-| ChamferDistance              |     | √    |     |     |        |
-| CrissCrossAttention          |     | √    |     |     |        |
-| ContourExpand                | √   |      |     |     |        |
-| ConvexIoU                    |     | √    |     |     |        |
-| CornerPool                   |     | √    |     |     |        |
-| Correlation                  |     | √    |     |     |        |
-| Deformable Convolution v1/v2 | √   | √    |     |     |        |
-| Deformable RoIPool           |     | √    | √   |     |        |
-| DiffIoURotated               |     | √    |     |     |        |
-| DynamicScatter               |     | √    |     |     |        |
-| FurthestPointSample          |     | √    |     |     |        |
-| FurthestPointSampleWithDist  |     | √    |     |     |        |
-| FusedBiasLeakyrelu           |     | √    |     |     |        |
-| GatherPoints                 |     | √    |     |     |        |
-| GroupPoints                  |     | √    |     |     |        |
-| Iou3d                        |     | √    | √   |     |        |
-| KNN                          |     | √    |     |     |        |
-| MaskedConv                   |     | √    | √   |     | √      |
-| MergeCells                   |     | √    |     |     |        |
-| MinAreaPolygon               |     | √    |     |     |        |
-| ModulatedDeformConv2d        | √   | √    | √   |     | √      |
-| MultiScaleDeformableAttn     |     | √    | √   |     |        |
-| NMS                          | √   | √    | √   |     | √      |
-| NMSRotated                   | √   | √    |     |     |        |
-| NMSQuadri                    | √   | √    |     |     |        |
-| PixelGroup                   | √   |      |     |     |        |
-| PointsInBoxes                | √   | √    |     |     |        |
-| PointsInPolygons             |     | √    |     |     |        |
-| PSAMask                      | √   | √    | √   |     |        |
-| RotatedFeatureAlign          | √   | √    |     |     |        |
-| RoIPointPool3d               |     | √    | √   |     |        |
-| RoIPool                      |     | √    | √   |     |        |
-| RoIAlignRotated              | √   | √    | √   |     |        |
-| RiRoIAlignRotated            |     | √    |     |     |        |
-| RoIAlign                     | √   | √    | √   |     |        |
-| RoIAwarePool3d               |     | √    | √   |     |        |
-| SAConv2d                     |     | √    |     |     |        |
-| SigmoidFocalLoss             |     | √    | √   |     | √      |
-| SoftmaxFocalLoss             |     | √    |     |     | √      |
-| SoftNMS                      |     | √    |     |     |        |
-| Sparse Convolution           |     | √    |     |     |        |
-| Synchronized BatchNorm       |     | √    |     |     |        |
-| ThreeInterpolate             |     | √    |     |     |        |
-| ThreeNN                      |     | √    | √   |     |        |
-| TINShift                     |     | √    | √   |     |        |
-| UpFirDn2d                    |     | √    |     |     |        |
-| Voxelization                 | √   | √    |     |     |        |
-| PrRoIPool                    |     | √    |     |     |        |
+| Device                       | CPU | CUDA | MLU | MPS | NPU |
+| ---------------------------- | --- | ---- | --- | --- | --- |
+| ActiveRotatedFilter          | √   | √    |     |     |     |
+| AssignScoreWithK             |     | √    |     |     |     |
+| BallQuery                    |     | √    |     |     |     |
+| BBoxOverlaps                 |     | √    | √   | √   |     |
+| BorderAlign                  |     | √    |     |     |     |
+| BoxIouRotated                | √   | √    |     |     |     |
+| BoxIouQuadri                 | √   | √    |     |     |     |
+| CARAFE                       |     | √    | √   |     |     |
+| ChamferDistance              |     | √    |     |     |     |
+| CrissCrossAttention          |     | √    |     |     |     |
+| ContourExpand                | √   |      |     |     |     |
+| ConvexIoU                    |     | √    |     |     |     |
+| CornerPool                   |     | √    |     |     |     |
+| Correlation                  |     | √    |     |     |     |
+| Deformable Convolution v1/v2 | √   | √    |     |     |     |
+| Deformable RoIPool           |     | √    | √   |     |     |
+| DiffIoURotated               |     | √    |     |     |     |
+| DynamicScatter               |     | √    |     |     |     |
+| FurthestPointSample          |     | √    |     |     |     |
+| FurthestPointSampleWithDist  |     | √    |     |     |     |
+| FusedBiasLeakyrelu           |     | √    |     |     | √   |
+| GatherPoints                 |     | √    |     |     |     |
+| GroupPoints                  |     | √    |     |     |     |
+| Iou3d                        |     | √    | √   |     |     |
+| KNN                          |     | √    |     |     |     |
+| MaskedConv                   |     | √    | √   |     | √   |
+| MergeCells                   |     | √    |     |     |     |
+| MinAreaPolygon               |     | √    |     |     |     |
+| ModulatedDeformConv2d        | √   | √    | √   |     | √   |
+| MultiScaleDeformableAttn     |     | √    | √   |     |     |
+| NMS                          | √   | √    | √   |     | √   |
+| NMSRotated                   | √   | √    |     |     |     |
+| NMSQuadri                    | √   | √    |     |     |     |
+| PixelGroup                   | √   |      |     |     |     |
+| PointsInBoxes                | √   | √    |     |     |     |
+| PointsInPolygons             |     | √    |     |     |     |
+| PSAMask                      | √   | √    | √   |     |     |
+| RotatedFeatureAlign          | √   | √    |     |     |     |
+| RoIPointPool3d               |     | √    | √   |     |     |
+| RoIPool                      |     | √    | √   |     |     |
+| RoIAlignRotated              | √   | √    | √   |     |     |
+| RiRoIAlignRotated            |     | √    |     |     |     |
+| RoIAlign                     | √   | √    | √   |     |     |
+| RoIAwarePool3d               |     | √    | √   |     |     |
+| SAConv2d                     |     | √    |     |     |     |
+| SigmoidFocalLoss             |     | √    | √   |     | √   |
+| SoftmaxFocalLoss             |     | √    |     |     | √   |
+| SoftNMS                      |     | √    |     |     |     |
+| Sparse Convolution           |     | √    |     |     |     |
+| Synchronized BatchNorm       |     | √    |     |     |     |
+| ThreeInterpolate             |     | √    |     |     |     |
+| ThreeNN                      |     | √    | √   |     |     |
+| TINShift                     |     | √    | √   |     |     |
+| UpFirDn2d                    |     | √    |     |     |     |
+| Voxelization                 | √   | √    |     |     |     |
+| PrRoIPool                    |     | √    |     |     |     |

mmcv/ops/csrc/pytorch/npu/fused_bias_leakyrelu_npu.cpp

@@ -0,0 +1,60 @@
+#include "pytorch_npu_helper.hpp"
+
+using namespace NPU_NAME_SPACE;
+using namespace std;
+
+Tensor fused_bias_leakyrelu_op_impl(const Tensor &input, const Tensor &bias,
+                                    const Tensor &refer, int act, int grad,
+                                    float alpha, float scale);
+
+Tensor fused_bias_leakyrelu_npu(const Tensor &input, const Tensor &bias,
+                                const Tensor &refer, int act, int grad,
+                                float alpha, float scale)
+{
+  at::Tensor py = at::empty_like(input);
+  // forward
+  if (grad == 0)
+  {
+    auto input_size = input.sizes();
+    int input_length = input_size.size();
+    c10::SmallVector<int64_t, SIZE> input_size_tmp;
+    input_size_tmp = array_to_small_vector(input_size);
+    if (input_length > 1)
+    {
+      for (int i = 0; i < input_length; i++)
+      {
+        if (i != 1)
+        {
+          input_size_tmp[i] = 1;
+        }
+      }
+    }
+    at::Tensor bias_tmp = at::reshape(bias, input_size_tmp);
+    at::Tensor bias_ = at_npu::native::NPUNativeFunctions::npu_broadcast(
+        bias_tmp, input.sizes());
+    OpCommand cmd;
+    cmd.Name("FusedBiasLeakyRelu")
+        .Input(input)
+        .Input(bias_)
+        .Output(py)
+        .Attr("scale", scale)
+        .Attr("negative_slope", alpha)
+        .Run();
+  }
+
+  // backward
+  if (grad == 1)
+  {
+    OpCommand cmd;
+    cmd.Name("FusedBiasLeakyReluGrad")
+        .Input(input)
+        .Input(refer)
+        .Output(py)
+        .Attr("scale", scale)
+        .Attr("negative_slope", alpha)
+        .Run();
+  }
+  return py;
+}
+
+REGISTER_NPU_IMPL(fused_bias_leakyrelu_op_impl, fused_bias_leakyrelu_npu);

mmcv/ops/fused_bias_leakyrelu.py

@@ -258,7 +258,7 @@ def fused_bias_leakyrelu(input: torch.Tensor,
         torch.Tensor: Feature map after non-linear activation.
     """
 
-    if not input.is_cuda:
+    if not input.is_cuda and input.device.type != 'npu':
         return bias_leakyrelu_ref(input, bias, negative_slope, scale)
 
     return FusedBiasLeakyReLUFunction.apply(input, bias.to(input.dtype),