Movatterモバイル変換

[0]ホーム
URL:

Skip to content

Navigation Menu

Sign in
Appearance settings

Search code, repositories, users, issues, pull requests...

Provide feedback

We read every piece of feedback, and take your input very seriously.

Saved searches

Use saved searches to filter your results more quickly

 Sign up
Appearance settings

Commit916ed2b

Browse files
[NPU] add npu result (#2569)
Motivationadd NPU results.Modificationadd docs/en/device/npu.md and docs/zh_cn/device/npu.md that accompaniesthe submission results.---------Co-authored-by: Miao Zheng <76149310+MeowZheng@users.noreply.github.com>
1 parentb2577e0 commit916ed2b

File tree

2 files changed

+72
-0
lines changed

2 files changed

+72
-0
lines changed

‎docs/en/device/npu.md‎

Lines changed: 36 additions & 0 deletions
Original file line numberDiff line numberDiff line change
@@ -0,0 +1,36 @@
1+
#NPU (HUAWEI Ascend)
2+
3+
##Usage
4+
5+
Please refer to the[building documentation of MMCV](https://mmcv.readthedocs.io/en/latest/get_started/build.html#build-mmcv-full-on-ascend-npu-machine) to install MMCV on NPU devices
6+
7+
Here we use 4 NPUs on your computer to train the model with the following command:
8+
9+
```shell
10+
bash tools/dist_train.sh configs/deeplabv3/deeplabv3_r50-d8_4xb2-40k_cityscapes-512x1024.py 4
11+
```
12+
13+
Also, you can use only one NPU to train the model with the following command:
14+
15+
```shell
16+
python tools/train.py configs/deeplabv3/deeplabv3_r50-d8_4xb2-40k_cityscapes-512x1024.py
17+
```
18+
19+
##Models Results
20+
21+
| Model| mIoU| Config| Download|
22+
| :-----------------:| :---:| :------------------------------------------------------------------------------------------------------------------------------------| :------------------------------------------------------------------------------------------------------------------------------------------|
23+
|[deeplabv3](<>)| 78.85|[config](https://github.com/wangjiangben-hw/mmsegmentation/blob/master/configs/deeplabv3/deeplabv3_r50-d8_512x1024_40k_cityscapes.py)|[log](https://download.openmmlab.com/mmsegmentation/v0.5/device/npu/deeplabv3_r50-d8_4xb2-40k_cityscapes-512x1024_20230115_205626.json)|
24+
|[deeplabv3plus](<>)| 79.23|[config](https://github.com/wangjiangben-hw/mmsegmentation/blob/master/configs/deeplabv3plus_r50-d8_4xb2-40k_cityscapes-512x1024)|[log](https://download.openmmlab.com/mmsegmentation/v0.5/device/npu/deeplabv3plus_r50-d8_4xb2-40k_cityscapes-512x1024_20230116_043450.json)|
25+
|[hrnet](<>)| 78.1|[config](https://github.com/wangjiangben-hw/mmsegmentation/blob/master/configs/hrnet/fcn_hr18_4xb2-40k_cityscapes-512x1024.py)|[log](https://download.openmmlab.com/mmsegmentation/v0.5/device/npu/fcn_hr18_4xb2-40k_cityscapes-512x1024_20230116_215821.json)|
26+
|[fcn](<>)| 74.15|[config](https://github.com/wangjiangben-hw/mmsegmentation/blob/master/configs/fcn/fcn_r50-d8_4xb2-40k_cityscapes-512x1024.py)|[log](https://download.openmmlab.com/mmsegmentation/v0.5/device/npu/fcn_r50-d8_4xb2-40k_cityscapes-512x1024_20230111_083014.json)|
27+
|[icnet](<>)| 69.25|[config](https://github.com/wangjiangben-hw/mmsegmentation/blob/master/configs/icnet/icnet_r50-d8_4xb2-80k_cityscapes-832x832.py)|[log](https://download.openmmlab.com/mmsegmentation/v0.5/device/npu/icnet_r50-d8_4xb2-80k_cityscapes-832x832_20230119_002929.json)|
28+
|[pspnet](<>)| 77.21|[config](https://github.com/wangjiangben-hw/mmsegmentation/blob/master/configs/pspnet/pspnet_r50b-d8_4xb2-80k_cityscapes-512x1024.py)|[log](https://download.openmmlab.com/mmsegmentation/v0.5/device/npu/pspnet_r50b-d8_4xb2-80k_cityscapes-512x1024_20230114_042721.json)|
29+
|[unet](<>)| 68.86|[config](https://github.com/wangjiangben-hw/mmsegmentation/blob/master/configs/unet/unet-s5-d16_fcn_4xb4-160k_cityscapes-512x1024.py)|[log](https://download.openmmlab.com/mmsegmentation/v0.5/device/npu/unet-s5-d16_fcn_4xb4-160k_cityscapes-512x1024_20230129_224750.json)|
30+
|[upernet](<>)| 77.81|[config](https://github.com/wangjiangben-hw/mmsegmentation/blob/master/configs/upernet/upernet_r50_4xb2-40k_cityscapes-512x1024.py)|[log](https://download.openmmlab.com/mmsegmentation/v0.5/device/npu/upernet_r50_4xb2-40k_cityscapes-512x1024_20230129_014634.json)|
31+
32+
**Notes:**
33+
34+
- If not specially marked, the results on NPU with amp are the basically same as those on the GPU with FP32.
35+
36+
**All above models are provided by Huawei Ascend group.**

‎docs/zh_cn/device/npu.md‎

Lines changed: 36 additions & 0 deletions
Original file line numberDiff line numberDiff line change
@@ -0,0 +1,36 @@
1+
#NPU (华为 昇腾)
2+
3+
##使用方法
4+
5+
请参考[MMCV 的安装文档](https://mmcv.readthedocs.io/en/latest/get_started/build.html#build-mmcv-full-on-ascend-npu-machine) 来安装 NPU 版本的 MMCV。
6+
7+
以下展示单机四卡场景的运行指令:
8+
9+
```shell
10+
bash tools/dist_train.sh configs/deeplabv3/deeplabv3_r50-d8_4xb2-40k_cityscapes-512x1024.py 4
11+
```
12+
13+
以下展示单机单卡下的运行指令:
14+
15+
```shell
16+
python tools/train.py configs/deeplabv3/deeplabv3_r50-d8_4xb2-40k_cityscapes-512x1024.py
17+
```
18+
19+
##模型验证结果
20+
21+
| Model| mIoU| Config| Download|
22+
| :-----------------:| :---:| :------------------------------------------------------------------------------------------------------------------------------------| :------------------------------------------------------------------------------------------------------------------------------------------|
23+
|[deeplabv3](<>)| 78.85|[config](https://github.com/wangjiangben-hw/mmsegmentation/blob/master/configs/deeplabv3/deeplabv3_r50-d8_512x1024_40k_cityscapes.py)|[log](https://download.openmmlab.com/mmsegmentation/v0.5/device/npu/deeplabv3_r50-d8_4xb2-40k_cityscapes-512x1024_20230115_205626.json)|
24+
|[deeplabv3plus](<>)| 79.23|[config](https://github.com/wangjiangben-hw/mmsegmentation/blob/master/configs/deeplabv3plus_r50-d8_4xb2-40k_cityscapes-512x1024)|[log](https://download.openmmlab.com/mmsegmentation/v0.5/device/npu/deeplabv3plus_r50-d8_4xb2-40k_cityscapes-512x1024_20230116_043450.json)|
25+
|[hrnet](<>)| 78.1|[config](https://github.com/wangjiangben-hw/mmsegmentation/blob/master/configs/hrnet/fcn_hr18_4xb2-40k_cityscapes-512x1024.py)|[log](https://download.openmmlab.com/mmsegmentation/v0.5/device/npu/fcn_hr18_4xb2-40k_cityscapes-512x1024_20230116_215821.json)|
26+
|[fcn](<>)| 74.15|[config](https://github.com/wangjiangben-hw/mmsegmentation/blob/master/configs/fcn/fcn_r50-d8_4xb2-40k_cityscapes-512x1024.py)|[log](https://download.openmmlab.com/mmsegmentation/v0.5/device/npu/fcn_r50-d8_4xb2-40k_cityscapes-512x1024_20230111_083014.json)|
27+
|[icnet](<>)| 69.25|[config](https://github.com/wangjiangben-hw/mmsegmentation/blob/master/configs/icnet/icnet_r50-d8_4xb2-80k_cityscapes-832x832.py)|[log](https://download.openmmlab.com/mmsegmentation/v0.5/device/npu/icnet_r50-d8_4xb2-80k_cityscapes-832x832_20230119_002929.json)|
28+
|[pspnet](<>)| 77.21|[config](https://github.com/wangjiangben-hw/mmsegmentation/blob/master/configs/pspnet/pspnet_r50b-d8_4xb2-80k_cityscapes-512x1024.py)|[log](https://download.openmmlab.com/mmsegmentation/v0.5/device/npu/pspnet_r50b-d8_4xb2-80k_cityscapes-512x1024_20230114_042721.json)|
29+
|[unet](<>)| 68.86|[config](https://github.com/wangjiangben-hw/mmsegmentation/blob/master/configs/unet/unet-s5-d16_fcn_4xb4-160k_cityscapes-512x1024.py)|[log](https://download.openmmlab.com/mmsegmentation/v0.5/device/npu/unet-s5-d16_fcn_4xb4-160k_cityscapes-512x1024_20230129_224750.json)|
30+
|[upernet](<>)| 77.81|[config](https://github.com/wangjiangben-hw/mmsegmentation/blob/master/configs/upernet/upernet_r50_4xb2-40k_cityscapes-512x1024.py)|[log](https://download.openmmlab.com/mmsegmentation/v0.5/device/npu/upernet_r50_4xb2-40k_cityscapes-512x1024_20230129_014634.json)|
31+
32+
**注意:**
33+
34+
- 如果没有特别标记，NPU 上的使用混合精度训练的结果与使用 FP32 的 GPU 上的结果相同。
35+
36+
**以上模型结果由华为昇腾团队提供**

0 commit comments

Comments
 (0)
[8]ページ先頭
©2009-2025 Movatter.jp