[Example] Add STAFNet Model for Air Quality Prediction #982

dylan-yin · 2024-08-26T11:11:39Z

PR types

PR changes

Describe

paddle-bot · 2024-08-26T11:11:44Z

Thanks for your contribution!

CLAassistant · 2024-08-26T11:11:44Z

Thank you for your submission! We really appreciate it. Like many open source projects, we ask that you sign our Contributor License Agreement before we can accept your contribution.
_{You have signed the CLA already but the status is still pending? Let us recheck it.}

HydrogenSulfate

项目文件看起来不完整，缺少多个必要的文件
按照如下步骤提交代码：

HydrogenSulfate · 2024-08-27T03:05:12Z

examples/demo/conf/stafnet.yaml

+hydra:
+  run:
+    # dynamic output directory according to running time and override name
+    dir: outputs_chip_heat/${now:%Y-%m-%d}/${now:%H-%M-%S}/${hydra.job.override_dirname}
+  job:
+    name: ${mode} # name of logfile
+    chdir: false # keep current working directory unchanged
+    config:
+      override_dirname:
+        exclude_keys:
+          - TRAIN.checkpoint_path
+          - TRAIN.pretrained_model_path
+          - EVAL.pretrained_model_path
+          - mode
+          - output_dir
+          - log_freq
+  callbacks:
+    init_callback:
+      _target_: ppsci.utils.callbacks.InitCallback
+  sweep:
+    # output directory for multirun
+    dir: ${hydra.run.dir}
+    subdir: ./


Suggested change

hydra:

run:

# dynamic output directory according to running time and override name

dir: outputs_chip_heat/${now:%Y-%m-%d}/${now:%H-%M-%S}/${hydra.job.override_dirname}

job:

name: ${mode} # name of logfile

chdir: false # keep current working directory unchanged

config:

override_dirname:

exclude_keys:

- TRAIN.checkpoint_path

- TRAIN.pretrained_model_path

- EVAL.pretrained_model_path

- mode

- output_dir

- log_freq

callbacks:

init_callback:

_target_: ppsci.utils.callbacks.InitCallback

sweep:

# output directory for multirun

dir: ${hydra.run.dir}

subdir: ./

defaults:

- ppsci_default

- TRAIN: train_default

- TRAIN/ema: ema_default

- TRAIN/swa: swa_default

- EVAL: eval_default

- INFER: infer_default

- hydra/job/config/override_dirname/exclude_keys: exclude_keys_default

- _self_

hydra:

run:

# dynamic output directory according to running time and override name

dir: outputs_chip_heat/${now:%Y-%m-%d}/${now:%H-%M-%S}/${hydra.job.override_dirname}

job:

name: ${mode} # name of logfile

chdir: false # keep current working directory unchanged

callbacks:

init_callback:

_target_: ppsci.utils.callbacks.InitCallback

sweep:

# output directory for multirun

dir: ${hydra.run.dir}

subdir: ./

HydrogenSulfate · 2024-08-27T03:05:39Z

examples/demo/conf/stafnet.yaml

+hydra:
+  run:
+    # dynamic output directory according to running time and override name
+    dir: outputs_chip_heat/${now:%Y-%m-%d}/${now:%H-%M-%S}/${hydra.job.override_dirname}


dir: 后面的名字改成正确的名字

HydrogenSulfate · 2024-08-27T03:05:48Z

examples/demo/conf/stafnet.yaml

+output_dir: ${hydra:run.dir}
+log_freq: 20
+
+STAFNet_DATA_PATH: "/data4/home/hyin/workspace/SATFNet/data/2020-2023_new/train_data.pkl" # 


使用相对路径

HydrogenSulfate · 2024-08-28T08:48:31Z

docs/zh/examples/STAFNet.md

@@ -0,0 +1,1029 @@
+# STAFNet
+


文档开头请加上模型训练和评估命令，参考

PaddleScience/docs/zh/examples/allen_cahn.md

Lines 1 to 23 in 5d710f6

# Allen-Cahn

<a href="https://aistudio.baidu.com/projectdetail/7927786" class="md-button md-button--primary" style>AI Studio快速体验</a>

=== "模型训练命令"

``` sh

# linux

wget -nc https://paddle-org.bj.bcebos.com/paddlescience/datasets/AllenCahn/allen_cahn.mat -P ./dataset/

# windows

# curl https://paddle-org.bj.bcebos.com/paddlescience/datasets/AllenCahn/allen_cahn.mat --create-dirs -o ./dataset/allen_cahn.mat

python allen_cahn_piratenet.py

```

=== "模型评估命令"

``` sh

# linux

wget -nc https://paddle-org.bj.bcebos.com/paddlescience/datasets/AllenCahn/allen_cahn.mat -P ./dataset/

# windows

# curl https://paddle-org.bj.bcebos.com/paddlescience/datasets/AllenCahn/allen_cahn.mat --create-dirs -o ./dataset/allen_cahn.mat

python allen_cahn_piratenet.py mode=eval EVAL.pretrained_model_path=https://paddle-org.bj.bcebos.com/paddlescience/models/AllenCahn/allen_cahn_piratenet_pretrained.pdparams

```

HydrogenSulfate · 2024-08-28T08:48:48Z

docs/zh/examples/STAFNet.md

+
+模型的总体结构如图所示：
+
+![image-20240530165151443](C:\Users\pal\AppData\Roaming\Typora\typora-user-images\image-20240530165151443.png)


请将图片上传到本PR

HydrogenSulfate · 2024-08-28T08:48:57Z

docs/zh/examples/STAFNet.md

+
+STAFNet 包含三个模块，分别将空间信息、气象信息和历史信息融合到空气质量特征表征中。首先模型的输入：过去T个时刻的**空气质量**数据和**气象**数据，使用两个空间感知 GNN（SAGNN），利用监测站之间的空间关系分别提取空气质量和气象信息。然后，跨图融合注意（CGF）将气象信息融合到空气质量表征中。最后，我们采用 TimesNet 模型来描述空气质量序列的时间动态，并生成多步骤预测。这一推理过程可表述如下，
+
+![image-20240531173333183](C:\Users\pal\AppData\Roaming\Typora\typora-user-images\image-20240531173333183.png)


请将图片上传到本PR

HydrogenSulfate · 2024-08-28T08:49:02Z

docs/zh/examples/STAFNet.md

+
+所有空气质量和气象观测数据每小时记录一次。数据集的收集时间为 2021 年 1 月 24 日至 2023 年 1 月 19 日，按 9:1的比例将数据分为训练集和测试集。空气质量观测数据来自国家城市空气质量实时发布平台，气象观测数据来自中国气象局。数据集的具体细节如下表所示，
+
+<div>			<!--块级封装-->     <center>	<!--将图片和文字居中-->     <img src="C:\Users\pal\AppData\Roaming\Typora\typora-user-images\image-20240530104042194.png" alt="image-20240530104042194" style="zoom: 25%;" />     <br>		<!--换行-->     北京空气质量数据集	<!--标题-->     </center> </div>


请将图片上传到本PR

HydrogenSulfate · 2024-08-28T08:49:19Z

docs/zh/examples/STAFNet.md

+
+数据加载代码如下：
+
+```python


代码块请使用引用语法，参考

PaddleScience/docs/zh/examples/allen_cahn.md

Lines 136 to 142 in 5d710f6

``` py linenums="94"

--8<--

examples/allen_cahn/allen_cahn_piratenet.py:94:110

--8<--

```

HydrogenSulfate · 2024-08-28T08:51:57Z

examples/demo/output_example/train.log

训练日志可以不用上传，需要上传的是训练完成保存下来的模型参数

HydrogenSulfate · 2024-08-28T08:52:02Z

examples/demo/outputs/2024-08-24/14-16-45/demo.log

同理，不用上传

HydrogenSulfate · 2024-08-28T08:53:17Z

examples/demo/outputs_chip_heat/2024-08-24/14-18-21/train.log

同理，不用上传

HydrogenSulfate · 2024-08-28T08:53:57Z

ppsci/data/dataset/__init__.py

请不要修改或删除已有代码，建议将当前代码与上游代码进行合并，再提交代码

Add files via upload

6300a58

Merge branch 'develop' into dev_model

45ab3f3

dylan-yin changed the title ~~Add files via upload~~ Add files STAFNet Aug 26, 2024

dylan-yin changed the title ~~Add files STAFNet~~ Add files STAFNet Model for Air Quality Prediction Aug 26, 2024

Add files via upload

c229d7e

paddle-bot bot added the contributor label Aug 26, 2024

HydrogenSulfate changed the title ~~Add files STAFNet Model for Air Quality Prediction~~ [Example] Add STAFNet Model for Air Quality Prediction Aug 27, 2024

HydrogenSulfate requested changes Aug 27, 2024

View reviewed changes

dylan-yin added 2 commits August 27, 2024 12:25

Add files via upload

d14dbc1

Add files via upload

c7940d5

luotao1 self-assigned this Aug 27, 2024

HydrogenSulfate requested changes Aug 28, 2024

View reviewed changes

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

[Example] Add STAFNet Model for Air Quality Prediction #982

[Example] Add STAFNet Model for Air Quality Prediction #982

dylan-yin commented Aug 26, 2024

paddle-bot bot commented Aug 26, 2024

CLAassistant commented Aug 26, 2024 •

edited

Loading

HydrogenSulfate left a comment

HydrogenSulfate Aug 27, 2024

HydrogenSulfate Aug 27, 2024

HydrogenSulfate Aug 27, 2024

HydrogenSulfate Aug 28, 2024

HydrogenSulfate Aug 28, 2024

HydrogenSulfate Aug 28, 2024

HydrogenSulfate Aug 28, 2024

HydrogenSulfate Aug 28, 2024

HydrogenSulfate Aug 28, 2024

HydrogenSulfate Aug 28, 2024

HydrogenSulfate Aug 28, 2024

HydrogenSulfate Aug 28, 2024

	# Allen-Cahn

	<a href="https://aistudio.baidu.com/projectdetail/7927786" class="md-button md-button--primary" style>AI Studio快速体验</a>

	=== "模型训练命令"

	``` sh
	# linux
	wget -nc https://paddle-org.bj.bcebos.com/paddlescience/datasets/AllenCahn/allen_cahn.mat -P ./dataset/
	# windows
	# curl https://paddle-org.bj.bcebos.com/paddlescience/datasets/AllenCahn/allen_cahn.mat --create-dirs -o ./dataset/allen_cahn.mat
	python allen_cahn_piratenet.py
	```

	=== "模型评估命令"

	``` sh
	# linux
	wget -nc https://paddle-org.bj.bcebos.com/paddlescience/datasets/AllenCahn/allen_cahn.mat -P ./dataset/
	# windows
	# curl https://paddle-org.bj.bcebos.com/paddlescience/datasets/AllenCahn/allen_cahn.mat --create-dirs -o ./dataset/allen_cahn.mat
	python allen_cahn_piratenet.py mode=eval EVAL.pretrained_model_path=https://paddle-org.bj.bcebos.com/paddlescience/models/AllenCahn/allen_cahn_piratenet_pretrained.pdparams
	```


		模型的总体结构如图所示：

		![image-20240530165151443](C:\Users\pal\AppData\Roaming\Typora\typora-user-images\image-20240530165151443.png)


		STAFNet 包含三个模块，分别将空间信息、气象信息和历史信息融合到空气质量特征表征中。首先模型的输入：过去T个时刻的空气质量数据和气象数据，使用两个空间感知 GNN（SAGNN），利用监测站之间的空间关系分别提取空气质量和气象信息。然后，跨图融合注意（CGF）将气象信息融合到空气质量表征中。最后，我们采用 TimesNet 模型来描述空气质量序列的时间动态，并生成多步骤预测。这一推理过程可表述如下，

		![image-20240531173333183](C:\Users\pal\AppData\Roaming\Typora\typora-user-images\image-20240531173333183.png)


		所有空气质量和气象观测数据每小时记录一次。数据集的收集时间为 2021 年 1 月 24 日至 2023 年 1 月 19 日，按 9:1的比例将数据分为训练集和测试集。空气质量观测数据来自国家城市空气质量实时发布平台，气象观测数据来自中国气象局。数据集的具体细节如下表所示，

		<div> <!--块级封装--> <center> <!--将图片和文字居中--> <img src="C:\Users\pal\AppData\Roaming\Typora\typora-user-images\image-20240530104042194.png" alt="image-20240530104042194" style="zoom: 25%;" /> <br> <!--换行--> 北京空气质量数据集 <!--标题--> </center> </div>


	``` py linenums="94"
	--8<--
	examples/allen_cahn/allen_cahn_piratenet.py:94:110
	--8<--
	```


		数据加载代码如下：

		```python

[Example] Add STAFNet Model for Air Quality Prediction #982

Are you sure you want to change the base?

[Example] Add STAFNet Model for Air Quality Prediction #982

Conversation

dylan-yin commented Aug 26, 2024

PR types

PR changes

Describe

paddle-bot bot commented Aug 26, 2024

CLAassistant commented Aug 26, 2024 • edited Loading

HydrogenSulfate left a comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

CLAassistant commented Aug 26, 2024 •

edited

Loading