Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Jump to bottom

[Docs] Replace NPU with Ascend in installation documentation #2469

Merged
merged 14 commits into from
Dec 7, 2022
Merged

[Docs] Replace NPU with Ascend in installation documentation #2469

Show file tree
Hide file tree
Changes from all commits
Commits
Show all changes
14 commits
Select commit Hold shift + click to select a range
0bc804f
update NPU --> ASCEND
wangjiangben-hw Dec 5, 2022
c3a9c9d
update docs
wangjiangben-hw Dec 5, 2022
62c6119
update
wangjiangben-hw Dec 5, 2022
5f38cfd
update docs
wangjiangben-hw Dec 5, 2022
f808320
Update docs/en/get_started/build.md
wangjiangben-hw Dec 6, 2022
94c1d1e
Update docs/zh_cn/get_started/build.md
wangjiangben-hw Dec 6, 2022
6adc79b
Update docs/zh_cn/get_started/build.md
wangjiangben-hw Dec 7, 2022
ee5e282
Update docs/zh_cn/get_started/build.md
wangjiangben-hw Dec 7, 2022
e122f39
Update docs/zh_cn/get_started/build.md
wangjiangben-hw Dec 7, 2022
0faf717
Update docs/en/get_started/build.md
wangjiangben-hw Dec 7, 2022
79ed0c2
Update docs/en/get_started/build.md
wangjiangben-hw Dec 7, 2022
8ca9d27
Update docs/en/get_started/build.md
wangjiangben-hw Dec 7, 2022
196df22
Update docs/en/get_started/build.md
zhouzaida Dec 7, 2022
046dcda
Update docs/zh_cn/get_started/build.md
zhouzaida Dec 7, 2022
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
48 changes: 48 additions & 0 deletions docs/en/get_started/build.md
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -327,3 +327,51 @@ Firstly, you need to apply for an IPU cloud machine, see [here](https://www.grap
1. Build MMCV

2. Use pip to install sdk according to [IPU PyTorch document](https://docs.graphcore.ai/projects/poptorch-user-guide/en/latest/installation.html). Also, you need to apply for machine and sdk to Graphcore.

### Build mmcv-full on Ascend NPU machine

Before building mmcv-full, `torch_npu` should be installed. See the complete installation tutorial [PyTorch Installation Guide](https://gitee.com/ascend/pytorch/blob/master/docs/en/PyTorch%20Installation%20Guide/PyTorch%20Installation%20Guide.md)

#### Option 1: Install mmcv-full with pip

The Ascend compiled version of mmcv-full is already supported when the version of mmcv >= 1.7.0, we can pip install directly

```bash
pip install mmcv-full -f https://download.openmmlab.com/mmcv/dist/ascend/torch1.8.0/index.html
```

#### Option 2: Build mmcv-full NPU (Ascend) from Source

- Pull the source code

```bash
git pull https://github.com/open-mmlab/mmcv/tree/master
```

- Build

```bash
MMCV_WITH_OPS=1 MAX_JOBS=8 FORCE_NPU=1 python setup.py build_ext
```

- Install

```bash
MMCV_WITH_OPS=1 FORCE_NPU=1 python setup.py develop
```

#### Test Case

```python
import torch
import torch_npu
from mmcv.ops import softmax_focal_loss

# Init tensor to the NPU
x = torch.randn(3, 10).npu()
y = torch.tensor([1, 5, 3]).npu()
w = torch.ones(10).float().npu()

output = softmax_focal_loss(x, y, 2.0, 0.25, w, 'none')
print(output)
```
116 changes: 58 additions & 58 deletions docs/en/understand_mmcv/ops.md
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -2,61 +2,61 @@

We implement common ops used in detection, segmentation, etc.

| 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 | | √ | | | |
| 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 | | √ | | | |
35 changes: 21 additions & 14 deletions docs/zh_cn/get_started/build.md
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -344,41 +344,48 @@ mmcv-full 有两个版本：

### 在昇腾 NPU 机器编译 mmcv-full

#### 1. 安装 torch_npu
在编译 mmcv-full 前，需要安装 torch_npu，完整安装教程详见 [PyTorch 安装指南](https://gitee.com/ascend/pytorch/blob/master/docs/zh/PyTorch%E5%AE%89%E8%A3%85%E6%8C%87%E5%8D%97/PyTorch%E5%AE%89%E8%A3%85%E6%8C%87%E5%8D%97.md#pytorch%E5%AE%89%E8%A3%85%E6%8C%87%E5%8D%97)

- torch_npu 完整安装教程详见 [PyTorch安装指南](https://gitee.com/ascend/pytorch/blob/master/docs/zh/PyTorch%E5%AE%89%E8%A3%85%E6%8C%87%E5%8D%97/PyTorch%E5%AE%89%E8%A3%85%E6%8C%87%E5%8D%97.md#pytorch%E5%AE%89%E8%A3%85%E6%8C%87%E5%8D%97)
#### 选项 1: 使用 pip 安装 Ascend 编译版本的 mmcv-full

#### 2. 编译 MMCV NPU
Ascend 编译版本的 mmcv-full 在 mmcv >= 1.7.0 时已经支持直接 pip 安装

拉取 [MMCV 源码](https://github.com/open-mmlab/mmcv/tree/master) 进行编译安装
```bash
pip install mmcv-full -f https://download.openmmlab.com/mmcv/dist/ascend/torch1.8.0/index.html
```

#### 选项 2: 使用 NPU 设备源码编译安装 mmcv-full

- 拉取 [MMCV 源码](https://github.com/open-mmlab/mmcv/tree/master)

```bash
git pull https://github.com/open-mmlab/mmcv/tree/master
```

编译：
- 编译

```bash
MMCV_WITH_OPS=1 MAX_JOBS=8 FORCE_NPU=1 python setup.py build_ext
```

安装：
- 安装

```bash
MMCV_WITH_OPS=1 FORCE_NPU=1 python setup.py develop
```

验证：
#### 验证

```python
import torch
import torch_npu
from mmcv.ops import softmax_focal_loss

# Init tensor to the NPU
x = torch.randn(3, 10).npu()
x.requires_grad = True
y = torch.tensor([1, 5, 3]).npu()
w = torch.ones(10).float().npu()
output = softmax_focal_loss(x, y, 2.0, 0.25, w, 'none')
```

使用说明：

```{note}
在 tensor 末尾加上 .npu() 就可以调用到 npu 算子
output = softmax_focal_loss(x, y, 2.0, 0.25, w, 'none')
print(output)
```
116 changes: 58 additions & 58 deletions docs/zh_cn/understand_mmcv/ops.md
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -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 | 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 | | √ | | | |
| 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 | | √ | | | |