Skip to content

Commit 02c0531

Browse files
nepeplwunepeplwu
authored
Add npu model list (PaddlePaddle#758)
1 parent 9a7ab51 commit 02c0531

File tree

4 files changed

+125
-3
lines changed

4 files changed

+125
-3
lines changed

‎README.md

Lines changed: 29 additions & 1 deletion
Original file line numberDiff line numberDiff line change
@@ -67,7 +67,9 @@ conda create -n paddlemix python=3.10 -y
6767
conda activate paddlemix
6868
```
6969
### 3. 安装PaddlePaddle
70-
#### 方法 1: 一键安装（推荐）
70+
71+
#### 方法 1: 一键安装（GPU/CPU推荐）
72+
7173
- CUDA 11.x或12.3
7274
- PaddlePaddle 3.0.0b1
7375
```
@@ -77,6 +79,30 @@ sh build_paddle_env.sh
7779
#### 方法 2: 手动安装
7880
关于PaddlePaddle安装的详细教程请查看[Installation](https://www.paddlepaddle.org.cn/install/quick?docurl=/documentation/docs/zh/develop/install/pip/linux-pip.html)。
7981

82+
#### 方法 3：昇腾环境安装
83+
84+
当前 PaddleMIX 支持昇腾 910B 芯片（更多型号还在支持中，如果您有其他型号的相关需求，请提交issue告知我们），昇腾驱动版本为 23.0.3。考虑到环境差异性，我们推荐使用飞桨官方提供的标准镜像完成环境准备。
85+
86+
* 参考如下命令启动容器，ASCEND_RT_VISIBLE_DEVICES 指定可见的 NPU 卡号
87+
88+
```shell
89+
docker run -it --name paddle-npu-dev -v $(pwd):/work \
90+
--privileged --network=host --shm-size=128G -w=/work \
91+
-v /usr/local/Ascend/driver:/usr/local/Ascend/driver \
92+
-v /usr/local/bin/npu-smi:/usr/local/bin/npu-smi \
93+
-v /usr/local/dcmi:/usr/local/dcmi \
94+
-e ASCEND_RT_VISIBLE_DEVICES="0,1,2,3,4,5,6,7" \
95+
registry.baidubce.com/device/paddle-npu:cann80T13-ubuntu20-$(uname -m)-gcc84-py39 /bin/bash
96+
```
97+
98+
* 在容器内安装飞桨
99+
100+
```shell
101+
# 注意需要先安装飞桨 cpu 版本
102+
python3.9 -m pip install paddlepaddle==3.0.0.dev20240520 -i https://www.paddlepaddle.org.cn/packages/nightly/cpu/
103+
python3.9 -m pip install paddle_custom_npu==3.0.0.dev20240719 -i https://www.paddlepaddle.org.cn/packages/nightly/npu/
104+
```
105+
80106

81107
### 4. 安装依赖
82108

@@ -211,6 +237,8 @@ sh build_env.sh
211237

212238
更多模型能力，可参考[模型能力矩阵](./paddlemix/examples/README.md)
213239

240+
昇腾910B支持的模型列表，可参考[昇腾模型列表](./docs/hardware_support/npu_model_list.md)
241+
214242
## 社区交流
215243

216244
- 微信扫描二维码并填写问卷，即可加入交流群与众多社区开发者以及官方团队深度交流。

‎README_EN.md

Lines changed: 28 additions & 2 deletions
Original file line numberDiff line numberDiff line change
@@ -47,7 +47,7 @@ https://github.com/PaddlePaddle/PaddleMIX/assets/29787866/8d32722a-e307-46cb-a8c
4747

4848

4949

50-
## Installation
50+
## Installation
5151
### 1. Clone the PaddleMIX repository
5252
```
5353
git clone https://github.com/PaddlePaddle/PaddleMIX
@@ -73,7 +73,7 @@ For detailed instructions on installing PaddlePaddle, please refer to [Installat
7373

7474
### 4. Install dependencies
7575

76-
#### Method 1: One-click installation (recommended)
76+
#### Method 1: One-click installation (recommended on GPU / CPU)
7777

7878
Run the following command to automatically install all necessary dependencies:
7979
```
@@ -88,6 +88,29 @@ Detailed [installation]((https://www.paddlepaddle.org.cn/install/quick?docurl=/d
8888
8989
> If you wish to train and infer using **bf16**, please use a GPU that supports **bf16**, such as the A100.
9090
91+
#### Method 3: For Ascend 910B
92+
93+
PaddleMIX currently supports the Ascend 910B chip (more chip types are still being supported. If you have related needs for other chip types, please submit a question and let us know). Our recommended driver version is 23.0.3. Considering the differences in environments, we recommend using docker image provided by PaddlePaddle to complete the environment preparation.
94+
95+
* Refer to the following command to start the container, ASCEND_RT_VISIBLE_DEVICES specifies the visible NPU card number
96+
97+
```shell
98+
docker run -it --name paddle-npu-dev -v $(pwd):/work \
99+
--privileged --network=host --shm-size=128G -w=/work \
100+
-v /usr/local/Ascend/driver:/usr/local/Ascend/driver \
101+
-v /usr/local/bin/npu-smi:/usr/local/bin/npu-smi \
102+
-v /usr/local/dcmi:/usr/local/dcmi \
103+
-e ASCEND_RT_VISIBLE_DEVICES="0,1,2,3,4,5,6,7" \
104+
registry.baidubce.com/device/paddle-npu:cann80T13-ubuntu20-$(uname -m)-gcc84-py39 /bin/bash
105+
```
106+
107+
* Install PaddlePaddle in the container
108+
109+
```shell
110+
python3.9 -m pip install paddlepaddle==3.0.0.dev20240520 -i https://www.paddlepaddle.org.cn/packages/nightly/cpu/
111+
python3.9 -m pip install paddle_custom_npu==3.0.0.dev20240719 -i https://www.paddlepaddle.org.cn/packages/nightly/npu/
112+
```
113+
91114

92115
## Tutorial
93116

@@ -184,6 +207,9 @@ Detailed [installation]((https://www.paddlepaddle.org.cn/install/quick?docurl=/d
184207
</table>
185208

186209
For more information on additional model capabilities, please refer to the [Model Capability Matrix](./paddlemix/examples/README.md).
210+
211+
For the list of models supported on Ascend 910B, please refer to [Ascend Model List](./docs/hardware_support/npu_model_list_en.md)
212+
187213
## LICENSE
188214

189215
This repository is licensed under the [Apache 2.0 license](LICENSE)

‎docs/hardware_support/npu_model_list.md

Lines changed: 34 additions & 0 deletions
Original file line numberDiff line numberDiff line change
@@ -0,0 +1,34 @@
1+
# PaddleMIX模型列表（昇腾）
2+
3+
<table align="center">
4+
<tbody>
5+
<tr align="center" valign="center">
6+
<td>
7+
<b>多模态理解</b>
8+
</td>
9+
<td>
10+
<b>多模态生成</b>
11+
</td>
12+
</tr>
13+
<tr valign="top">
14+
<td>
15+
<ul>
16+
</ul>
17+
<li><b>图文预训练</b></li>
18+
<ul>
19+
<li><a href="../../paddlemix/examples/llava">LLaVA-1.6</a></li>
20+
<li><a href="../../paddlemix/examples/internvl2">InternVL2</a></li>
21+
</ul>
22+
</td>
23+
<td>
24+
<ul>
25+
</ul>
26+
<li><b>文生图</b></li>
27+
<ul>
28+
<li><a href="../../ppdiffusers/examples/stable_diffusion">Stable Diffusion</a></li>
29+
<li><a href="../../ppdiffusers/examples/dreambooth/README_sd3.md">Stable Diffusion 3 (SD3)</a></li>
30+
</ul>
31+
</td>
32+
</tr>
33+
</tbody>
34+
</table>

‎docs/hardware_support/npu_model_list_en.md

Lines changed: 34 additions & 0 deletions
Original file line numberDiff line numberDiff line change
@@ -0,0 +1,34 @@
1+
# PaddleMIX model list（Ascend）
2+
3+
<table align="center">
4+
<tbody>
5+
<tr align="center" valign="center">
6+
<td>
7+
<b>Multi-modal Large Language Models</b>
8+
</td>
9+
<td>
10+
<b>Diffusion-based Models</b>
11+
</td>
12+
</tr>
13+
<tr valign="top">
14+
<td>
15+
<ul>
16+
</ul>
17+
<li><b>Image-Text Pre-training</b></li>
18+
<ul>
19+
<li><a href="../../paddlemix/examples/llava">LLaVA-1.6</a></li>
20+
<li><a href="../../paddlemix/examples/internvl2">InternVL2</a></li>
21+
</ul>
22+
</td>
23+
<td>
24+
<ul>
25+
</ul>
26+
<li><b>Text-to-Image</b></li>
27+
<ul>
28+
<li><a href="../../ppdiffusers/examples/stable_diffusion">Stable Diffusion</a></li>
29+
<li><a href="../../ppdiffusers/examples/dreambooth/README_sd3.md">Stable Diffusion 3 (SD3)</a></li>
30+
</ul>
31+
</td>
32+
</tr>
33+
</tbody>
34+
</table>

0 commit comments

Comments
 (0)