[Example] Add STAFNet Model for Air Quality Prediction by dylan-yin · Pull Request #1070 · PaddlePaddle/PaddleScience

dylan-yin · 2025-02-07T13:02:36Z

PR types

PR changes

Describe

CLAassistant · 2025-02-07T13:02:43Z

All committers have signed the CLA.

HydrogenSulfate

整体项目请使用pre-commit格式化一边

HydrogenSulfate · 2025-02-10T03:17:57Z

@@ -0,0 +1,136 @@
+hydra:


配置文件开头请加上以下字段：

PaddleScience/examples/ldc/conf/ldc_2d_Re3200_piratenet.yaml

Lines 1 to 9 in fad6927

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_

HydrogenSulfate · 2025-02-10T03:18:07Z

+    config:
+      override_dirname:
+        exclude_keys:
+          - TRAIN.checkpoint_path
+          - TRAIN.pretrained_model_path
+          - EVAL.pretrained_model_path
+          - mode
+          - output_dir
+          - log_freq


这个可以删了

HydrogenSulfate · 2025-02-10T03:18:45Z

+STAFNet_DATA_PATH: "/data6/home/yinhang2021/workspace/SATFNet/data/2020-2023_new/train_data.pkl" # 
+DATASET:
+  label_keys: ["label"]
+  data_dir: "/data6/home/yinhang2021/workspace/SATFNet/data/2020-2023_new/train_data.pkl"
+  STAFNet_DATA_args: {
+            "data_dir": "/data6/home/yinhang2021/workspace/SATFNet/data/2020-2023_new/train_data.pkl",
+            "batch_size": 1,
+            "shuffle": True,
+            "num_workers": 0,
+            "training": True
+        }
+
+
+
+# "data_dir": "data/2020-2023_new/train_data.pkl",
+#             "batch_size": 32,
+#             "shuffle": True,
+#             "num_workers": 0,
+#             "training": True
+# model settings
+# MODEL: # 
+


建议改为相对路径，以./data/...开头即可

HydrogenSulfate · 2025-02-10T03:19:04Z

+# "data_dir": "data/2020-2023_new/train_data.pkl",
+#             "batch_size": 32,
+#             "shuffle": True,
+#             "num_workers": 0,
+#             "training": True
+# model settings
+# MODEL: # 


这个注释如果没用的可以删除

HydrogenSulfate · 2025-02-10T03:19:16Z

+  # configs: {
+  #         "task_name": "forecast",
+  #         "output_attention": False,
+  #         "seq_len": 72,
+  #         "label_len": 24,
+  #         "pred_len": 48,
+
+  #         "aq_gat_node_features" : 7,
+  #         "aq_gat_node_num": 35,
+
+  #         "mete_gat_node_features" : 7,
+  #         "mete_gat_node_num": 18,
+
+  #         "gat_hidden_dim": 32,
+  #         "gat_edge_dim": 3,
+  #         "gat_embed_dim": 32,
+
+  #         "e_layers": 1,
+  #         "enc_in": 7,
+  #         "dec_in": 7,
+  #         "c_out": 7,
+  #         "d_model": 16 ,
+  #         "embed": "fixed",
+  #         "freq": "t",
+  #         "dropout": 0.05,
+  #         "factor": 3,
+  #         "n_heads": 4,
+
+  #         "d_ff": 32 ,
+  #         "num_kernels": 6,
+  #         "top_k": 4
+  #     }


同上，如果没用可以删除

HydrogenSulfate · 2025-02-10T03:22:09Z

+
+
+
+


避免连续空行

HydrogenSulfate · 2025-02-10T03:22:59Z

+    # set random seed for reproducibility
+    ppsci.utils.misc.set_random_seed(42)
+    # set output directory
+    OUTPUT_DIR = "./output_example"
+    # initialize logger
+    logger.init_logger("ppsci", f"{OUTPUT_DIR}/train.log", "info")


这个可以删除，output_dir会由ppsci.utils.callbacks.InitCallback自动创建:

PaddleScience/ppsci/utils/callbacks.py

Lines 90 to 96 in fad6927

logger.init_logger(

"ppsci",

osp.join(full_cfg.output_dir, f"{full_cfg.mode}.log")

if full_cfg.output_dir and full_cfg.mode not in ["export", "infer"]

else None,

full_cfg.log_level,

)

HydrogenSulfate · 2025-02-10T03:23:20Z

+from typing import Tuple
+
+class Inception_Block_V1(paddle.nn.Layer):
+


冗余的空行请删除，下同

HydrogenSulfate · 2025-02-12T05:31:51Z

+output_dir: ${hydra:run.dir}
+log_freq: 20
+# dataset setting
+STAFNet_DATA_PATH: "/data6/home/yinhang2021/dataset/chongqing_1921/train_data.pkl" # 


这里的路径是否能改成相对路径？比如 ./dataset/train_data.pkl，其余的路径字段也是，建议改为相对路径，并去掉用户名

STAFNet_DATA_PATH是否应该放到DATASET字段下？

HydrogenSulfate · 2025-02-12T05:33:02Z

+
+
+MODEL:
+  input_keys: ["aq_train_data","mete_train_data",]


Suggested change

input_keys: ["aq_train_data","mete_train_data",]

input_keys: [aq_train_data, mete_train_data]

HydrogenSulfate · 2025-02-12T05:33:06Z

+
+MODEL:
+  input_keys: ["aq_train_data","mete_train_data",]
+  output_keys: ["label"]


Suggested change

output_keys: ["label"]

output_keys: [label]

HydrogenSulfate · 2025-02-12T05:33:23Z

+  checkpoint_path: null 
+
+EVAL:
+  eval_data_path: "/data6/home/yinhang2021/dataset/chongqing_1921/val_data.pkl"


Suggested change

eval_data_path: "/data6/home/yinhang2021/dataset/chongqing_1921/val_data.pkl"

eval_data_path: ./dataset/val_data.pkl

HydrogenSulfate · 2025-02-12T05:34:06Z

+STAFNet_DATA_PATH: "/data6/home/yinhang2021/dataset/chongqing_1921/train_data.pkl" # 
+DATASET:
+  label_keys: ["label"]
+  data_dir: "/data6/home/yinhang2021/dataset/chongqing_1921/train_data.pkl"


data_dir为什么是具体文件路径而不是某个文件夹路径?

此处的路径是否跟STAFNet_DATA_PATH重复了？

HydrogenSulfate · 2025-02-12T05:37:13Z

+        cfg.TRAIN.epochs,
+        ITERS_PER_EPOCH,
+        eval_during_train=cfg.TRAIN.eval_during_train,
+        seed=cfg.seed,


Suggested change

seed=cfg.seed,

HydrogenSulfate · 2025-02-12T05:37:30Z

+    """
+    Validate after training an epoch
+
+    :param epoch: Integer, current training epoch.
+    :return: A log that contains information about validation
+    """


Suggested change

"""

Validate after training an epoch

:param epoch: Integer, current training epoch.

:return: A log that contains information about validation

"""

HydrogenSulfate · 2025-02-12T05:37:52Z

+        "sampler": {
+            "name": "BatchSampler",
+            "drop_last": False,
+            "shuffle": True,
+        },


Suggested change

"sampler": {

"name": "BatchSampler",

"drop_last": False,

"shuffle": True,

},

HydrogenSulfate · 2025-02-12T05:38:01Z

+    # set random seed for reproducibility
+    ppsci.utils.misc.set_random_seed(42)
+    # set output directory
+    OUTPUT_DIR = "./output_example"
+    # initialize logger
+    logger.init_logger("ppsci", f"{OUTPUT_DIR}/train.log", "info")


Suggested change

# set random seed for reproducibility

ppsci.utils.misc.set_random_seed(42)

# set output directory

OUTPUT_DIR = "./output_example"

# initialize logger

logger.init_logger("ppsci", f"{OUTPUT_DIR}/train.log", "info")

HydrogenSulfate · 2025-02-12T05:38:38Z

+    OUTPUT_DIR = "./output_example"
+    # initialize logger
+    logger.init_logger("ppsci", f"{OUTPUT_DIR}/train.log", "info")
+    multiprocessing.set_start_method("spawn")


这句代码是什么作用？paddle的多卡训练不需要这样吧？

我这边如果不加 multiprocessing.set_start_method("spawn")，会出现cuda error（3）

我这边如果不加 multiprocessing.set_start_method("spawn")，会出现cuda error（3）

这是多卡训练时才会出现的吗？你的训练命令是什么呢？如果按照我们文档里给的数据并行命令，也会报错吗？

liaoxin2 · 2025-06-06T03:24:32Z

+
+````
+``` sh
+python stafnet.py TRAIN_DIR="Your train dataset path" eval_data_path="Your evaluate dataset path"


python stafnet.py DATASET.data_dir="Your train dataset path" EVAL.eval_data_path="Your evaluate dataset path"

liaoxin2 · 2025-06-06T03:30:23Z

+
+````
+``` sh
+python stafnet.py mode=eval EVAL.pretrained_model_path="https://paddle-org.bj.bcebos.com/paddlescience/models/stafnet/stafnet.pdparams" EVAL.pretrained_model_path="https://paddle-org.bj.bcebos.com/paddlescience/datasets/stafnet/val_data.pkl"


wget -nc https://paddle-org.bj.bcebos.com/paddlescience/datasets/stafnet/val_data.pkl -P ./dataset/
python stafnet.py mode=eval EVAL.pretrained_model_path="https://paddle-org.bj.bcebos.com/paddlescience/models/stafnet/stafnet.pdparams"

liaoxin2 · 2025-06-06T03:32:33Z

+
+针对空气质量预测提出了许多研究。早期的方法侧重于学习单个观测站观测数据的时间模式，而放弃了观测站之间的空间关系。最近，由于图神经网络（GNN）在处理非欧几里得图结构方面的有效性，越来越多的方法采用 GNN 来模拟空间依赖关系。这些方法将车站位置作为上下文特征，隐含地建立空间依赖关系模型，没有充分利用车站位置和车站之间关系所包含的宝贵空间信息。此外，现有的时空 GNN 缺乏在错位图中融合多个特征的能力。因此，大多数方法都需要额外的插值算法，以便在早期阶段将气象特征与 AQ 特征进行对齐和连接。这种方法消除了空气质量站和气象站之间的空间和结构信息，还可能引入噪声导致误差累积。此外，在空气质量预测中利用多周期性的问题仍未得到探索。
+
+该案例研究时空图网络网络在空气质量预测方向上的应用。


去掉一个‘网络’

liaoxin2 · 2025-06-06T03:40:58Z

+```
+--8<--
+examples/stafnet/stafnet.py:11
+--8<--
+```


代码不显示，参考其他案例文档，在第一行添加py linenums="11" title="examples/stafnet/stafnet.py"，下面的代码块均修改一下

liaoxin2 · 2025-06-06T03:43:14Z

+
+### 3.7 评估器构建
+
+在训练过程中通常会按一定轮数间隔，用验证集(测试集)评估当前模型的训练情况，因此使用 `ppsci.validate.SupervisedValidator` 构建评估器，构建过程与 [约束构建](https://github.com/PaddlePaddle/PaddleScience/blob/develop/docs/zh/examples/cfdgcn.md#34) 类似，只需把数据目录改为测试集的目录，并在配置文件中设置 `EVAL.batch_size=1` 即可。


构建过程与 [3.6 约束构建](#36) 类似

liaoxin2 · 2025-06-06T03:44:48Z

+
+## 5. 参考资料
+
+- [STAFNet: Spatiotemporal-Aware Fusion Network for Air Quality Prediction]([STAFNet: Spatiotemporal-Aware Fusion Network for Air Quality Prediction | SpringerLink](https://link.springer.com/chapter/10.1007/978-3-031-78186-5_22))


[STAFNet: Spatiotemporal-Aware Fusion Network for Air Quality Prediction](https://link.springer.com/chapter/10.1007/978-3-031-78186-5_22)

liaoxin2 · 2025-06-06T05:09:11Z

+  output_attention: True
+  seq_len: 72
+  pred_len: 48
+  aq_gat_node_features: 7
+  aq_gat_node_num: 35
+  mete_gat_node_features: 7
+  mete_gat_node_num: 18
+  gat_hidden_dim: 32
+  gat_edge_dim: 3
+  e_layers: 2
+  enc_in: 7
+  dec_in: 7
+  c_out: 7
+  d_model: 32
+  embed: "fixed"
+  freq: "t"
+  dropout: 0.05
+  factor: 3
+  n_heads: 4
+  d_ff: 64
+  num_kernels: 6
+  top_k: 4


output_attention: true
seq_len: 72
pred_len: 48
aq_gat_node_features: 7
aq_gat_node_num: 35
mete_gat_node_features: 7
mete_gat_node_num: 18
gat_hidden_dim: 32
gat_edge_dim: 3
e_layers: 1
enc_in: 7
dec_in: 7
c_out: 7
d_model: 16
embed: fixed
freq: t
dropout: 0.05
factor: 3
n_heads: 4
d_ff: 32
num_kernels: 6
top_k: 4

liaoxin2 · 2025-06-06T05:09:29Z

+  checkpoint_path: null 
+
+EVAL:
+  eval_data_path: ./dataset/new_val_data.pkl


./dataset/val_data.pkl

liaoxin2 · 2025-06-06T05:15:37Z

+from omegaconf import DictConfig
+import hydra
+import paddle
+from ppsci.data.dataset.stafnet_dataset import gat_lstmcollate_fn


from ppsci.data.dataset.stafnet_dataset import gat_lstmcollate_fn去掉吧，并没有用到

HydrogenSulfate

所有提交的代码请使用pre-commit格式化一下

HydrogenSulfate · 2025-06-06T05:30:50Z

+DATASET:
+  label_keys: [label]
+  data_dir: ./dataset/train_data.pkl
+


Suggested change

HydrogenSulfate · 2025-06-06T05:30:59Z

+  pretrained_model_path: null
+  compute_metric_by_batch: false
+  eval_with_no_grad: true
+  batch_size: 32


文件末尾换行

liaoxin2 · 2025-06-09T02:25:51Z

+``` sh
+wget -nc https://paddle-org.bj.bcebos.com/paddlescience/datasets/stafnet/val_data.pkl -P ./dataset/
+python stafnet.py mode=eval EVAL.pretrained_model_path="https://paddle-org.bj.bcebos.com/paddlescience/models/stafnet/stafnet.pdparams"
+python stafnet.py mode=eval EVAL.pretrained_model_path="https://paddle-org.bj.bcebos.com/paddlescience/models/stafnet/stafnet.pdparams" EVAL.pretrained_model_path="https://paddle-org.bj.bcebos.com/paddlescience/datasets/stafnet/val_data.pkl"


去掉第21行

liaoxin2 · 2025-06-09T11:26:57Z

+--8<--
+examples/stafnet/stafnet.py:11
+--8<--


插入的代码还需再检查一下，插入的代码和文字说明不太对应

liaoxin2 · 2025-06-09T11:33:38Z

+
+### 3.7 评估器构建
+
+在训练过程中通常会按一定轮数间隔，用验证集(测试集)评估当前模型的训练情况，因此使用 `ppsci.validate.SupervisedValidator` 构建评估器，构建过程与 [约束构建](https://github.com/PaddlePaddle/PaddleScience/blob/develop/docs/zh/examples/stafnet.md#36) 类似，只需把数据目录改为测试集的目录，并在配置文件中设置 `EVAL.batch_size=1` 即可。


构建过程与 [3.6 约束构建](#36) 类似

liaoxin2 · 2025-06-11T06:37:09Z

+
+```py linenums="10" title="examples/stafnet/stafnet.py"
+--8<--
+examples/stafnet/stafnet.py:10


examples/stafnet/stafnet.py:10:10

liaoxin2 · 2025-06-11T07:21:15Z

+
+## 4. 完整代码
+
+```python py linenums="1" title="examples/stafnet/stafnet.py"


去掉python

liaoxin2 · 2025-06-12T07:25:50Z

+
+``` py linenums="62" title="examples/stafnet/stafnet.py"
+--8<--
+examples/stafnet/stafnet.py:62


examples/stafnet/stafnet.py:62:62

liaoxin2 · 2025-06-12T07:26:29Z

+
+``` py linenums="55" title="examples/stafnet/stafnet.py"
+--8<--
+examples/stafnet/stafnet.py:55


examples/stafnet/stafnet.py:55:55

HydrogenSulfate · 2025-06-16T06:39:51Z

+
+import numpy as np
+import paddle
+from pgl.nn.conv import GATv2Conv


因为pgl不是pdsci的依赖库，所以还麻烦按照以下方式修改导入方式，避免影响ppsci其它代码

HydrogenSulfate · 2025-06-16T06:40:47Z

还麻烦在docs/zh/api/data/dataset.md加上这个类

HydrogenSulfate · 2025-06-16T06:40:59Z

还麻烦在arch.md加上这个类

dylan-yin added 6 commits February 7, 2025 20:26

Add files via upload

f79a3f9

Add files via upload

68b23d1

Update __init__.py

d9d2b54

Add files via upload

bfa3e69

Merge branch 'PaddlePaddle:develop' into dev_model

2d9dc85

Update __init__.py

57dc7c2

paddle-bot Bot added the contributor label Feb 7, 2025

HydrogenSulfate reviewed Feb 10, 2025

View reviewed changes

dylan-yin added 5 commits February 11, 2025 23:12

Merge branch 'PaddlePaddle:develop' into dev_model

fa1cdee

Update demo.py

ab1ae03

Update stafnet.yaml

d257a49

Update stafnet.py

b43c7f5

Update stafnet_dataset.py

757477a

HydrogenSulfate changed the title ~~Add STAFNet Model for Air Quality Prediction~~ [Example] Add STAFNet Model for Air Quality Prediction Feb 12, 2025

HydrogenSulfate reviewed Feb 12, 2025

View reviewed changes

dylan-yin added 12 commits February 14, 2025 21:39

Merge branch 'PaddlePaddle:develop' into dev_model

2b46497

Update stafnet.yaml

711cd36

Update demo.py

a79ad1d

Update stafnet.yaml

af96434

Update demo.py

86a9c0b

Merge branch 'PaddlePaddle:develop' into dev_model

e2d8b60

Merge branch 'PaddlePaddle:develop' into dev_model

66eeb9b

Delete examples/demo directory

792796a

Add files via upload

db2f093

Add files via upload

862cbfb

Add files via upload

726a026

Add files via upload

5ccc908

liaoxin2 reviewed Jun 6, 2025

View reviewed changes

HydrogenSulfate reviewed Jun 6, 2025

View reviewed changes

dylan-yin added 4 commits June 7, 2025 10:10

Add files via upload

3332954

Add files via upload

f5557b9

Add files via upload

878cc52

Add files via upload

c2e88bd

liaoxin2 reviewed Jun 10, 2025

View reviewed changes

dylan-yin added 3 commits June 10, 2025 17:03

Update README.md

bc748cd

Add files via upload

c7d4426

Update stafnet.md

a7b2eb1

liaoxin2 reviewed Jun 12, 2025

View reviewed changes

dylan-yin added 2 commits June 12, 2025 15:33

Update stafnet.md

c025ab2

Merge branch 'develop' into dev_model

ed22102

HydrogenSulfate reviewed Jun 13, 2025

View reviewed changes

Comment thread docs/zh/examples/stafnet.md Outdated

Comment thread examples/stafnet/stafnet.py Outdated

Comment thread examples/stafnet/stafnet.py Outdated

HydrogenSulfate added 3 commits June 13, 2025 13:12

Update docs/zh/examples/stafnet.md

8d2433d

Update examples/stafnet/stafnet.py

3a87833

Update examples/stafnet/stafnet.py

88c2957

HydrogenSulfate reviewed Jun 16, 2025

View reviewed changes

dylan-yin added 9 commits June 16, 2025 18:13

Update stafnet.py

ec92f69

Update dataset.md

454729f

Update arch.md

b1b4ad3

Update arch.md

9c658dc

Update arch.md

9195278

Merge branch 'develop' into dev_model

3e35a0d

Update stafnet_dataset.py

803eb28

Update stafnet.py

8599724

Update stafnet_dataset.py

59cd4e0

dylan-yin force-pushed the dev_model branch from 7bb6fc7 to 59cd4e0 Compare June 23, 2025 13:33

dylan-yin and others added 3 commits June 23, 2025 21:34

Merge branch 'PaddlePaddle:develop' into dev_model

84f91bf

style: Apply automatic formatting via pre-commit

5039b14

Add files via upload

d38fbaa

HydrogenSulfate closed this Jan 5, 2026

	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_

	logger.init_logger(
	"ppsci",
	osp.join(full_cfg.output_dir, f"{full_cfg.mode}.log")
	if full_cfg.output_dir and full_cfg.mode not in ["export", "infer"]
	else None,
	full_cfg.log_level,
	)

		from typing import Tuple

		class Inception_Block_V1(paddle.nn.Layer):

	input_keys: ["aq_train_data","mete_train_data",]
	input_keys: [aq_train_data, mete_train_data]

	eval_data_path: "/data6/home/yinhang2021/dataset/chongqing_1921/val_data.pkl"
	eval_data_path: ./dataset/val_data.pkl


		针对空气质量预测提出了许多研究。早期的方法侧重于学习单个观测站观测数据的时间模式，而放弃了观测站之间的空间关系。最近，由于图神经网络（GNN）在处理非欧几里得图结构方面的有效性，越来越多的方法采用 GNN 来模拟空间依赖关系。这些方法将车站位置作为上下文特征，隐含地建立空间依赖关系模型，没有充分利用车站位置和车站之间关系所包含的宝贵空间信息。此外，现有的时空 GNN 缺乏在错位图中融合多个特征的能力。因此，大多数方法都需要额外的插值算法，以便在早期阶段将气象特征与 AQ 特征进行对齐和连接。这种方法消除了空气质量站和气象站之间的空间和结构信息，还可能引入噪声导致误差累积。此外，在空气质量预测中利用多周期性的问题仍未得到探索。

		该案例研究时空图网络网络在空气质量预测方向上的应用。


		### 3.7 评估器构建

		在训练过程中通常会按一定轮数间隔，用验证集(测试集)评估当前模型的训练情况，因此使用 `ppsci.validate.SupervisedValidator` 构建评估器，构建过程与 [约束构建](https://github.com/PaddlePaddle/PaddleScience/blob/develop/docs/zh/examples/cfdgcn.md#34) 类似，只需把数据目录改为测试集的目录，并在配置文件中设置 `EVAL.batch_size=1` 即可。


		## 5. 参考资料

		- [STAFNet: Spatiotemporal-Aware Fusion Network for Air Quality Prediction]([STAFNet: Spatiotemporal-Aware Fusion Network for Air Quality Prediction \| SpringerLink](https://link.springer.com/chapter/10.1007/978-3-031-78186-5_22))


		## 4. 完整代码

		```python py linenums="1" title="examples/stafnet/stafnet.py"

Conversation

dylan-yin commented Feb 7, 2025

PR types

PR changes

Describe

Uh oh!

CLAassistant commented Feb 7, 2025 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

HydrogenSulfate left a comment

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Uh oh!

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

liaoxin2 Jun 6, 2025 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Choose a reason for hiding this comment

Uh oh!

liaoxin2 Jun 6, 2025 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

HydrogenSulfate left a comment

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

CLAassistant commented Feb 7, 2025 •

edited

Loading

liaoxin2 Jun 6, 2025 •

edited

Loading

liaoxin2 Jun 6, 2025 •

edited

Loading