Skip to content

Commit a244705

Browse files
HydrogenSulfateHydrogenSulfate
committed
update doc
1 parent 65f981b commit a244705

File tree

6 files changed

+97
-66
lines changed

6 files changed

+97
-66
lines changed

‎deploy/python_infer/base.py

+2-2
Original file line numberDiff line numberDiff line change
@@ -38,7 +38,7 @@ class Predictor:
3838
Args:
3939
pdmodel_path (Optional[str]): Path to the PaddlePaddle model file. Defaults to None.
4040
pdiparams_path (Optional[str]): Path to the PaddlePaddle model parameters file. Defaults to None.
41-
device (Literal["gpu", "cpu", "npu", "xpu"], optional): Device to use for inference. Defaults to "cpu".
41+
device (Literal["gpu", "cpu", "npu", "xpu", "sdaa"], optional): Device to use for inference. Defaults to "cpu".
4242
engine (Literal["native", "tensorrt", "onnx", "mkldnn"], optional): Inference engine to use. Defaults to "native".
4343
precision (Literal["fp32", "fp16", "int8"], optional): Precision to use for inference. Defaults to "fp32".
4444
onnx_path (Optional[str], optional): Path to the ONNX model file. Defaults to None.
@@ -54,7 +54,7 @@ def __init__(
5454
pdmodel_path: Optional[str] = None,
5555
pdiparams_path: Optional[str] = None,
5656
*,
57-
device: Literal["gpu", "cpu", "npu", "xpu"] = "cpu",
57+
device: Literal["gpu", "cpu", "npu", "xpu", "sdaa"] = "cpu",
5858
engine: Literal["native", "tensorrt", "onnx", "mkldnn"] = "native",
5959
precision: Literal["fp32", "fp16", "int8"] = "fp32",
6060
onnx_path: Optional[str] = None,

‎docs/zh/multi_device.md

+89-59
Original file line numberDiff line numberDiff line change
@@ -2,86 +2,116 @@
22

33
飞桨生态的繁荣离不开开发者和用户的贡献，我们非常欢迎为飞桨的多硬件适配贡献更多的模型。
44

5+
## 1. 硬件支持列表
6+
57
当前 PaddleScience 中对于各硬件适配模型的列表汇总如下
68

79
<p align="center"><b>数学(AI for Math)</b></p>
810

911
| 问题类型 | 案例名称 | 昆仑芯 | 海光 | 寒武纪 | 昇腾 | 燧原 | 天数 | 摩尔线程 | 沐曦 | 太初 |
1012
|-----|-----|-----|-----|-----|-----|-----|-----|-----|-----|-----|
11-
| 亥姆霍兹方程 | [SPINN(Helmholtz3D)](./examples/spinn.md) | |✅| | | | | | |✅|
12-
| 相场方程 | [Allen-Cahn](./examples/allen_cahn.md) | |✅| | | | | | |✅|
13-
| 微分方程 | [拉普拉斯方程](./examples/laplace2d.md) | |✅| | | | | | |✅|
14-
| 微分方程 | [伯格斯方程](./examples/deephpms.md) | |✅| | | | | | |✅|
15-
| 微分方程 | [非线性偏微分方程](./examples/pirbn.md) | |✅| | | | | | |✅|
16-
| 微分方程 | [洛伦兹方程](./examples/lorenz.md) | |✅| | | | | | |✅|
17-
| 微分方程 | [若斯叻方程](./examples/rossler.md) | |✅| | | | | | |✅|
18-
| 算子学习 | [DeepONet](./examples/deeponet.md) | |✅| | | | | | |✅|
19-
| 微分方程 | [梯度增强的物理知识融合 PDE 求解](https://github.com/PaddlePaddle/PaddleScience/blob/develop/examples/gpinn/poisson_1d.py) | |✅| | | | | | |✅|
20-
| 积分方程 | [沃尔泰拉积分方程](./examples/volterra_ide.md) | |✅| | | | | | |✅|
21-
| 微分方程 | [分数阶微分方程](https://github.com/PaddlePaddle/PaddleScience/blob/develop/examples/fpde/fractional_poisson_2d.py) | |✅| | | | | | |✅|
22-
| 光纤怪波 | [Optical rogue wave](./examples/nlsmb.md) | |✅| | | | | | |✅|
23-
| 域分解 | [XPINN](./examples/xpinns.md) | |✅| | | | | | |✅|
24-
| 布鲁塞尔扩散系统 | [3D-Brusselator](./examples/brusselator3d.md) | |✅| | | | | | |✅|
25-
| 符号回归 | [Transformer4SR](./examples/transformer4sr.md) | |✅| | | | | | |✅|
13+
| 亥姆霍兹方程 | [SPINN(Helmholtz3D)](./examples/spinn.md) | | | | | | | | | |
14+
| 相场方程 | [Allen-Cahn](./examples/allen_cahn.md) | | | | | | | | | |
15+
| 微分方程 | [拉普拉斯方程](./examples/laplace2d.md) | | | | | | | | | |
16+
| 微分方程 | [伯格斯方程](./examples/deephpms.md) | | | | | | | | | |
17+
| 微分方程 | [非线性偏微分方程](./examples/pirbn.md) | | | | | | | | | |
18+
| 微分方程 | [洛伦兹方程](./examples/lorenz.md) | | | | | | | | | |
19+
| 微分方程 | [若斯叻方程](./examples/rossler.md) | | | | | | | | | |
20+
| 算子学习 | [DeepONet](./examples/deeponet.md) | | | | | | | | | |
21+
| 微分方程 | [梯度增强的物理知识融合 PDE 求解](https://github.com/PaddlePaddle/PaddleScience/blob/develop/examples/gpinn/poisson_1d.py) | | | | | | | | | |
22+
| 积分方程 | [沃尔泰拉积分方程](./examples/volterra_ide.md) | | | | | | | | | |
23+
| 微分方程 | [分数阶微分方程](https://github.com/PaddlePaddle/PaddleScience/blob/develop/examples/fpde/fractional_poisson_2d.py) | | | | | | | | | |
24+
| 光纤怪波 | [Optical rogue wave](./examples/nlsmb.md) | | | | | | | | | |
25+
| 域分解 | [XPINN](./examples/xpinns.md) | | | | | | | | | |
26+
| 布鲁塞尔扩散系统 | [3D-Brusselator](./examples/brusselator3d.md) | | | | | | | | | |
27+
| 符号回归 | [Transformer4SR](./examples/transformer4sr.md) | | | | | | | | | |
2628

2729
<br>
2830
<p align="center"><b>技术科学(AI for Technology)</b></p>
2931

3032
| 问题类型 | 案例名称 | 昆仑芯 | 海光 | 寒武纪 | 昇腾 | 燧原 | 天数 | 摩尔线程 | 沐曦 | 太初 |
3133
|-----|-----|-----|-----|-----|-----|-----|-----|-----|-----|-----|
32-
| 汽车表面阻力预测 | [DrivAerNet](./examples/drivaernet.md) | |✅| | | | | | |✅|
33-
| 一维线性对流问题 | [1D 线性对流](./examples/adv_cvit.md) | |✅| | | | | | |✅|
34-
| 非定常不可压流体 | [2D 方腔浮力驱动流](./examples/ns_cvit.md) | |✅| | | | | | |✅|
35-
| 定常不可压流体 | [Re3200 2D 定常方腔流](./examples/ldc2d_steady.md) | |✅| | | | | | |✅|
36-
| 定常不可压流体 | [2D 达西流](./examples/darcy2d.md) | |✅| | | | | | |✅|
37-
| 定常不可压流体 | [2D 管道流](./examples/labelfree_DNN_surrogate.md) | |✅| | | | | | |✅|
38-
| 定常不可压流体 | [3D 颅内动脉瘤](./examples/aneurysm.md) | |✅| | | | | | |✅|
39-
| 定常不可压流体 | [任意 2D 几何体绕流](./examples/deepcfd.md) | |✅| | | | | | |✅|
40-
| 非定常不可压流体 | [2D 非定常方腔流](./examples/ldc2d_unsteady.md) | |✅| | | | | | |✅|
41-
| 非定常不可压流体 | [Re100 2D 圆柱绕流](./examples/cylinder2d_unsteady.md) | |✅| | | | | | |✅|
42-
| 非定常不可压流体 | [Re100~750 2D 圆柱绕流](./examples/cylinder2d_unsteady_transformer_physx.md) | |✅| | | | | | |✅|
43-
| 可压缩流体 | [2D 空气激波](./examples/shock_wave.md) | |✅| | | | | | |✅|
44-
| 飞行器设计 | [MeshGraphNets](https://aistudio.baidu.com/projectdetail/5322713) | |✅| | | | | | |✅|
45-
| 飞行器设计 | [火箭发动机真空羽流](https://aistudio.baidu.com/projectdetail/4486133) | |✅| | | | | | |✅|
46-
| 飞行器设计 | [Deep-Flow-Prediction](https://aistudio.baidu.com/projectdetail/5671596) | |✅| | | | | | |✅|
47-
| 通用流场模拟 | [气动外形设计](./examples/amgnet.md) | |✅| | | | | | |✅|
48-
| 流固耦合 | [涡激振动](./examples/viv.md) | |✅| | | | | | |✅|
49-
| 多相流 | [气液两相流](./examples/bubble.md) | |✅| | | | | | |✅|
50-
| 多相流 | [twophasePINN](https://aistudio.baidu.com/projectdetail/5379212) | |✅| | | | | | |✅|
51-
| 流场高分辨率重构 | [2D 湍流流场重构](./examples/tempoGAN.md) | |✅| | | | | | |✅|
52-
| 流场高分辨率重构 | [2D 湍流流场重构](https://aistudio.baidu.com/projectdetail/4493261?contributionType=1) | |✅| | | | | | |✅|
53-
| 流场高分辨率重构 | [基于Voronoi嵌入辅助深度学习的稀疏传感器全局场重建](https://aistudio.baidu.com/projectdetail/5807904) | |✅| | | | | | |✅|
54-
| 流场预测 | [Catheter](https://aistudio.baidu.com/projectdetail/5379212) | |✅| | | | | | |✅|
55-
| 求解器耦合 | [CFD-GCN](./examples/cfdgcn.md) | |✅| | | | | | |✅|
56-
| 受力分析 | [1D 欧拉梁变形](./examples/euler_beam.md) | |✅| | | | | | |✅|
57-
| 受力分析 | [2D 平板变形](./examples/biharmonic2d.md) | |✅| | | | | | |✅|
58-
| 受力分析 | [3D 连接件变形](./examples/bracket.md) | |✅| | | | | | |✅|
59-
| 受力分析 | [结构震动模拟](./examples/phylstm.md) | |✅| | | | | | |✅|
60-
| 受力分析 | [2D 弹塑性结构](./examples/epnn.md) | |✅| | | | | | |✅|
61-
| 受力分析和逆问题 | [3D 汽车控制臂变形](./examples/control_arm.md) | |✅| | | | | | |✅|
62-
| 受力分析和逆问题 | [3D 心脏仿真](./examples/heart.md) | |✅| | | | | | |✅|
63-
| 拓扑优化 | [2D 拓扑优化](./examples/topopt.md) | |✅| | | | | | |✅|
64-
| 热仿真 | [1D 换热器热仿真](./examples/heat_exchanger.md) | |✅| | | | | | |✅|
65-
| 热仿真 | [2D 热仿真](./examples/heat_pinn.md) | |✅| | | | | | |✅|
66-
| 热仿真 | [2D 芯片热仿真](./examples/chip_heat.md) | |✅| | | | | | |✅|
34+
| 汽车表面阻力预测 | [DrivAerNet](./examples/drivaernet.md) | | | | | | | | | |
35+
| 一维线性对流问题 | [1D 线性对流](./examples/adv_cvit.md) | | | | | | | | | |
36+
| 非定常不可压流体 | [2D 方腔浮力驱动流](./examples/ns_cvit.md) | | | | | | | | | |
37+
| 定常不可压流体 | [Re3200 2D 定常方腔流](./examples/ldc2d_steady.md) | | | | | | | | | |
38+
| 定常不可压流体 | [2D 达西流](./examples/darcy2d.md) | | | | | | | | | |
39+
| 定常不可压流体 | [2D 管道流](./examples/labelfree_DNN_surrogate.md) | | | | | | | | | |
40+
| 定常不可压流体 | [3D 颅内动脉瘤](./examples/aneurysm.md) | | | | | | | | | |
41+
| 定常不可压流体 | [任意 2D 几何体绕流](./examples/deepcfd.md) | | | | | | | | | |
42+
| 非定常不可压流体 | [2D 非定常方腔流](./examples/ldc2d_unsteady.md) | | | | | | | | | |
43+
| 非定常不可压流体 | [Re100 2D 圆柱绕流](./examples/cylinder2d_unsteady.md) | | | | | | | | | |
44+
| 非定常不可压流体 | [Re100~750 2D 圆柱绕流](./examples/cylinder2d_unsteady_transformer_physx.md) | | | | | | | | | |
45+
| 可压缩流体 | [2D 空气激波](./examples/shock_wave.md) | | | | | | | | | |
46+
| 飞行器设计 | [MeshGraphNets](https://aistudio.baidu.com/projectdetail/5322713) | | | | | | | | | |
47+
| 飞行器设计 | [火箭发动机真空羽流](https://aistudio.baidu.com/projectdetail/4486133) | | | | | | | | | |
48+
| 飞行器设计 | [Deep-Flow-Prediction](https://aistudio.baidu.com/projectdetail/5671596) | | | | | | | | | |
49+
| 通用流场模拟 | [气动外形设计](./examples/amgnet.md) | | | | | | | | | |
50+
| 流固耦合 | [涡激振动](./examples/viv.md) | | | | | | | | | |
51+
| 多相流 | [气液两相流](./examples/bubble.md) | | | | | | | | | |
52+
| 多相流 | [twophasePINN](https://aistudio.baidu.com/projectdetail/5379212) | | | | | | | | | |
53+
| 流场高分辨率重构 | [2D 湍流流场重构](./examples/tempoGAN.md) | | | | | | | | | |
54+
| 流场高分辨率重构 | [2D 湍流流场重构](https://aistudio.baidu.com/projectdetail/4493261?contributionType=1) | | | | | | | | | |
55+
| 流场高分辨率重构 | [基于Voronoi嵌入辅助深度学习的稀疏传感器全局场重建](https://aistudio.baidu.com/projectdetail/5807904) | | | | | | | | | |
56+
| 流场预测 | [Catheter](https://aistudio.baidu.com/projectdetail/5379212) | | | | | | | | | |
57+
| 求解器耦合 | [CFD-GCN](./examples/cfdgcn.md) | | | | | | | | | |
58+
| 受力分析 | [1D 欧拉梁变形](./examples/euler_beam.md) | | | | | | | | | |
59+
| 受力分析 | [2D 平板变形](./examples/biharmonic2d.md) | | | | | | | | | |
60+
| 受力分析 | [3D 连接件变形](./examples/bracket.md) | | | | | | | | | |
61+
| 受力分析 | [结构震动模拟](./examples/phylstm.md) | | | | | | | | | |
62+
| 受力分析 | [2D 弹塑性结构](./examples/epnn.md) | | | | | | | | | |
63+
| 受力分析和逆问题 | [3D 汽车控制臂变形](./examples/control_arm.md) | | | | | | | | | |
64+
| 受力分析和逆问题 | [3D 心脏仿真](./examples/heart.md) | | | | | | | | | |
65+
| 拓扑优化 | [2D 拓扑优化](./examples/topopt.md) | | | | | | | | | |
66+
| 热仿真 | [1D 换热器热仿真](./examples/heat_exchanger.md) | | | | | | | | | |
67+
| 热仿真 | [2D 热仿真](./examples/heat_pinn.md) | | | | | | | | | |
68+
| 热仿真 | [2D 芯片热仿真](./examples/chip_heat.md) | | | | | | | | | |
6769

6870
<br>
6971
<p align="center"><b>材料科学(AI for Material)</b></p>
7072

7173
| 问题类型 | 案例名称 | 昆仑芯 | 海光 | 寒武纪 | 昇腾 | 燧原 | 天数 | 摩尔线程 | 沐曦 | 太初 |
7274
|-----|-----|-----|-----|-----|-----|-----|-----|-----|-----|-----|
73-
| 材料设计 | [散射板设计(反问题)](./examples/hpinns.md) | |✅| | | | | | |✅|
75+
| 材料设计 | [散射板设计(反问题)](./examples/hpinns.md) | | | | | | | | | |
7476

7577
<br>
7678
<p align="center"><b>地球科学(AI for Earth Science)</b></p>
7779

7880
| 问题类型 | 案例名称 | 昆仑芯 | 海光 | 寒武纪 | 昇腾 | 燧原 | 天数 | 摩尔线程 | 沐曦 | 太初 |
7981
|-----|-----|-----|-----|-----|-----|-----|-----|-----|-----|-----|
80-
| 天气预报 | [Extformer-MoE 气象预报](./examples/extformer_moe.md) | |✅| | | | | | |✅|
81-
| 天气预报 | [FourCastNet 气象预报](./examples/fourcastnet.md) | |✅| | | | | | |✅|
82-
| 天气预报 | [NowCastNet 气象预报](./examples/nowcastnet.md) | |✅| | | | | | |✅|
83-
| 天气预报 | [GraphCast 气象预报](./examples/graphcast.md) | |✅| | | | | | |✅|
84-
| 大气污染物 | [UNet 污染物扩散](https://aistudio.baidu.com/projectdetail/5663515?channel=0&channelType=0&sUid=438690&shared=1&ts=1698221963752) | |✅| | | | | | |✅|
85-
| 天气预报 | [DGMR 气象预报](./examples/dgmr.md) | |✅| | | | | | |✅|
86-
| 地震波形反演 | [VelocityGAN 地震波形反演](./examples/velocity_gan.md) | |✅| | | | | | |✅|
87-
| 交通预测 | [TGCN 交通流量预测](./examples/tgcn.md) | |✅| | | | | | |✅|
82+
| 天气预报 | [Extformer-MoE 气象预报](./examples/extformer_moe.md) | | | | | | | | | |
83+
| 天气预报 | [FourCastNet 气象预报](./examples/fourcastnet.md) | | | | | | | | | |
84+
| 天气预报 | [NowCastNet 气象预报](./examples/nowcastnet.md) | | | | | | | | | |
85+
| 天气预报 | [GraphCast 气象预报](./examples/graphcast.md) | | | | | | | | | |
86+
| 大气污染物 | [UNet 污染物扩散](https://aistudio.baidu.com/projectdetail/5663515?channel=0&channelType=0&sUid=438690&shared=1&ts=1698221963752) | | | | | | | | | |
87+
| 天气预报 | [DGMR 气象预报](./examples/dgmr.md) | | | | | | | | | |
88+
| 地震波形反演 | [VelocityGAN 地震波形反演](./examples/velocity_gan.md) | | | | | | | | | |
89+
| 交通预测 | [TGCN 交通流量预测](./examples/tgcn.md) | | | | | | | | | |
90+
91+
## 2. 贡献指南
92+
93+
我们在公开的案例文档开头提供了基于 GPU 训练的参考精度和对应的预训练模型权重，如果需要在指定的硬件上运行，可以参考如下步骤：
94+
95+
1. 在案例开头位置添加一行代码，将飞桨运行设备设置为当前硬件设备
96+
97+
``` py hl_lines="3"
98+
import paddle
99+
100+
paddle.set_device("your_device_name")
101+
102+
# 原案例代码
103+
```
104+
105+
2. 按照案例文档步骤，准备好数据集，在指定硬件上进行全量训练，保存训练日志，记录最佳模型精度以及最佳模型权重，这些内容一般会在训练过程中，自动保存在案例文件夹下
106+
107+
3. 如果模型文档中提供了模型导出和推理命令，请按照模型导出和推理命令，验证在新硬件上模型导出和推理是否能够正常执行并对齐 GPU 的推理结果
108+
109+
4. 上述步骤完成后，可以在本文档(`docs/zh/multi_device.md`)的表格中，给对应模型在指定硬件上添加支持信息(✅)，然后提交 PR 到 PaddleScience
110+
111+
## 3. 更多文档
112+
113+
更多关于飞桨多硬件适配和使用的相关文档，可以参考：
114+
115+
* [飞桨使用指南](https://www.paddlepaddle.org.cn/documentation/docs/zh/develop/guides/index_cn.html)
116+
* [飞桨硬件支持](https://www.paddlepaddle.org.cn/documentation/docs/zh/develop/hardware_support/index_cn.html)
117+
* [PaddleCustomDevice](https://github.com/PaddlePaddle/PaddleCustomDevice)

‎docs/zh/user_guide.md

+1
Original file line numberDiff line numberDiff line change
@@ -373,6 +373,7 @@ PaddleScience 提供了多种推理配置组合，可通过命令行进行组合
373373
| CPU | ✅ | ✅| / | ✅ |
374374
| GPU | ✅ | ✅ | ✅ | / |
375375
| XPU | TODO | / | / | / |
376+
| SDAA | TODO | / | / | / |
376377
377378
接下来以 aneurysm 案例和 Linux x86_64 + TensorRT 8.6 GA + CUDA 11.6 软硬件环境为例，介绍如何使用不同的推理配置。
378379

‎ppsci/arch/tfnonet.py

+1-1
Original file line numberDiff line numberDiff line change
@@ -511,4 +511,4 @@ def __init__(
511511
)
512512
self.n_modes_height = n_modes_height
513513
self.n_modes_width = n_modes_width
514-
self.n_modes_height = n_modes_height
514+
self.n_modes_depth = n_modes_depth

‎ppsci/solver/solver.py

+2-2
Original file line numberDiff line numberDiff line change
@@ -83,7 +83,7 @@ class Solver:
8383
use_wandb (Optional[bool]): Whether use wandb to log data. Defaults to False.
8484
use_tbd (Optional[bool]): Whether use tensorboardX to log data. Defaults to False.
8585
wandb_config (Optional[Dict[str, str]]): Config dict of WandB. Defaults to None.
86-
device (Literal["cpu", "gpu", "xpu", None], optional): Runtime device. Defaults to None, which means use default device on current platform.
86+
device (Literal["cpu", "gpu", "xpu", "sdaa", None], optional): Runtime device. Defaults to None, which means use default device on current platform.
8787
equation (Optional[Dict[str, ppsci.equation.PDE]]): Equation dict. Defaults to None.
8888
geom (Optional[Dict[str, ppsci.geometry.Geometry]]): Geometry dict. Defaults to None.
8989
validator (Optional[Dict[str, ppsci.validate.Validator]]): Validator dict. Defaults to None.
@@ -145,7 +145,7 @@ def __init__(
145145
use_wandb: bool = False,
146146
use_tbd: bool = False,
147147
wandb_config: Optional[Mapping] = None,
148-
device: Literal["cpu", "gpu", "xpu", None] = None,
148+
device: Literal["cpu", "gpu", "xpu", "sdaa", None] = None,
149149
equation: Optional[Dict[str, ppsci.equation.PDE]] = None,
150150
geom: Optional[Dict[str, ppsci.geometry.Geometry]] = None,
151151
validator: Optional[Dict[str, ppsci.validate.Validator]] = None,

‎ppsci/utils/config.py

+2-2
Original file line numberDiff line numberDiff line change
@@ -212,7 +212,7 @@ class InferConfig(BaseModel):
212212
pdmodel_path: Optional[str] = None
213213
pdiparams_path: Optional[str] = None
214214
onnx_path: Optional[str] = None
215-
device: Literal["gpu", "cpu", "npu", "xpu"] = "cpu"
215+
device: Literal["gpu", "cpu", "npu", "xpu", "sdaa"] = "cpu"
216216
engine: Literal["native", "tensorrt", "onnx", "mkldnn"] = "native"
217217
precision: Literal["fp32", "fp16", "int8"] = "fp32"
218218
ir_optim: bool = True
@@ -305,7 +305,7 @@ class SolverConfig(BaseModel):
305305
use_tbd: bool = False
306306
wandb_config: Mapping = {}
307307
use_wandb: bool = False
308-
device: Literal["cpu", "gpu", "xpu", None] = None
308+
device: Literal["cpu", "gpu", "xpu", "sdaa", None] = None
309309
use_amp: bool = False
310310
amp_level: Literal["O0", "O1", "O2", "OD"] = "O1"
311311
to_static: bool = False

0 commit comments

Comments
 (0)