-
Notifications
You must be signed in to change notification settings - Fork 0
/
Copy pathteecnet_exp_2_expressiveness.py
260 lines (216 loc) · 10.2 KB
/
teecnet_exp_2_expressiveness.py
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
import os
import shutil
import time
import numpy as np
import torch
import wandb
from torch.utils.tensorboard import SummaryWriter
import torch.optim as optim
from sklearn.metrics import r2_score
# from torch_geometric.nn import global_mean_pool
from torch_geometric.loader import DataLoader
from tqdm import tqdm
import matplotlib.pyplot as plt
import h5py
from utils import train_test_split, initialize_model, initialize_dataset, parse_args, load_yaml
def plot_edge_attributes(edge_index, edge_attr, pos):
num_edges = edge_index.shape[1]
x_values = []
y_values = []
edge_values = []
for i in range(num_edges):
start_node = edge_index[0][i]
end_node = edge_index[1][i]
x_start, y_start = pos[start_node]
x_end, y_end = pos[end_node]
x_center = (x_start + x_end) / 2
y_center = (y_start + y_end) / 2
edge_value = edge_attr[i].item()
x_values.append(x_center)
y_values.append(y_center)
edge_values.append(edge_value)
x_values = np.array(x_values).squeeze()
y_values = np.array(y_values).squeeze()
edge_values = np.array(edge_values).squeeze()
X, Y = np.meshgrid(np.unique(x_values), np.unique(y_values))
edge_value = np.zeros((len(np.unique(x_values)), len(np.unique(y_values))))
for i in range(len(x_values)):
x_index = np.where(X == x_values[i])
y_index = np.where(Y == y_values[i])
edge_value[x_index, y_index] = edge_values[i]
fig = plt.figure(figsize=(8, 6))
plt.contourf(X, Y, edge_value)
plt.title('Edge Attributes')
plt.colorbar()
return fig
def visualize_prediction(data, model, epoch, **kwargs):
model.eval()
x, edge_index, edge_attr = data.x, data.edge_index, data.edge_attr
x = x.to(kwargs['device'])
edge_index = edge_index.to(kwargs['device'])
edge_attr = edge_attr.to(kwargs['device'])
pred = model(x, edge_index, edge_attr).detach().cpu().numpy()
# pred = model(x, edge_index).detach().cpu().numpy()
x = data.pos[:, 0].detach().cpu().numpy()
y = data.pos[:, 1].detach().cpu().numpy()
# x = data.pos[:, 0].detach().cpu().numpy()
# y = data.pos[:, 1].detach().cpu().numpy()
x_values = np.unique(x)
y_values = np.unique(y)
temp_grid = pred.squeeze().reshape(len(x_values), len(y_values))
# temp_grid = np.sqrt(temp_grid[:, 0] ** 2 + temp_grid[:, 1] ** 2).reshape(len(x_values), len(y_values))
fig = plt.figure(figsize=(8, 6))
plt.contourf(x_values, y_values, temp_grid, levels=100, cmap='jet')
# plt.contourf(x_values, y_values, temp_grid)
plt.colorbar()
# plt.title('Velocity Contour Plot')
# plt.xlabel('x')
# plt.ylabel('y')
# plt.axes('off')
# remove the axis
plt.axis('off')
wandb.log({"Prediction": wandb.Image(fig)})
# plt.savefig("figures/pred_{}.png".format(epoch))
# writer.add_figure("Prediction", fig, epoch)
plt.close(fig)
temp_grid_true = data.y.cpu().detach().numpy().squeeze().reshape(len(x_values), len(y_values))
# temp_grid_true = np.sqrt(temp_grid_true[:, 0] ** 2 + temp_grid_true[:, 1] ** 2).reshape(len(x_values), len(y_values))
fig = plt.figure(figsize=(8, 6))
plt.contourf(x_values, y_values, temp_grid_true, levels=np.linspace(0, 1, 100), cmap='jet')
# plt.contourf(x_values, y_values, temp_grid_true)
# limit the three figures to have the same colorbar
plt.colorbar()
# plt.axes('off')
# remove the axis
plt.axis('off')
wandb.log({"True": wandb.Image(fig)})
# plt.title('Velocity Contour Plot')
# plt.xlabel('x')
# plt.ylabel('y')
# writer.add_figure("True", fig, epoch)
plt.close(fig)
temp_grid_error = np.abs(temp_grid - temp_grid_true)
fig = plt.figure(figsize=(8, 6))
plt.contourf(x_values, y_values, temp_grid_error, levels=np.linspace(0, 1, 100), cmap='jet')
# plt.contourf(x_values, y_values, temp_grid_error)
plt.colorbar()
# plt.axes('off')
# remove the axis
plt.axis('off')
wandb.log({"Error": wandb.Image(fig)})
# plt.title('Velocity Error Map')
# plt.xlabel('x')
# plt.ylabel('y')
# writer.add_figure("Error", fig, epoch)
plt.close(fig)
x_low = data.pos[:, 0].detach().cpu().numpy()
y_low = data.pos[:, 1].detach().cpu().numpy()
x_values_low = np.unique(x_low)
y_values_low = np.unique(y_low)
# temp_grid_low = data.x.detach().cpu().numpy().squeeze().reshape(len(x_values_low), len(y_values_low))
temp_grid_low = data.x[:, 0].detach().cpu().numpy().squeeze().reshape(len(x_values_low), len(y_values_low))
# temp_grid_low = np.sqrt(temp_grid_low[:, 0] ** 2 + temp_grid_low[:, 1] ** 2).reshape(len(x_values_low), len(y_values_low))
fig = plt.figure(figsize=(8, 6))
# plt.contourf(x_values_low, y_values_low, temp_grid_low, levels=np.linspace(0, 1, 100), cmap="RdBu_r")
plt.contourf(x_values, y_values, temp_grid_low, levels=np.linspace(0, 1, 100), cmap="jet")
# plt.contourf(x_values, y_values, temp_grid_low)
plt.colorbar()
# plt.axes('off')
# remove the axis
plt.axis('off')
# plt.title('Velocity Contour Map')
# plt.xlabel('x')
# plt.ylabel('y')
# writer.add_figure("Low Resolution", fig, epoch)
plt.close(fig)
# kernel_k = model.kernel_out.weight_k.detach().cpu().numpy().squeeze()
# kernel_op = model.kernel_out.weight_op.detach().cpu().numpy().squeeze()
# fig_k = plot_edge_attributes(edge_index, kernel_k, data.pos)
# writer.add_figure("Kernel_k", fig_k, epoch)
# plt.close(fig_k)
# fig_op = plot_edge_attributes(edge_index, kernel_op, data.pos)
# writer.add_figure("Kernel_op", fig_op, epoch)
# plt.close(fig_op)
model.train()
def train(model, dataset, log_dir, model_dir):
device = torch.device('cuda' if torch.cuda.is_available() else 'cpu')
# model = initialize_model(type='NeuralOperator', in_channel=1, out_channel=1, width=64, ker_width=512, depth=6).to(device)
model = model.to(device)
print('The model has {} parameters'.format(sum(p.numel() for p in model.parameters() if p.requires_grad)))
optimizer = optim.SGD(model.parameters(), lr=0.001, momentum=0.9, weight_decay=5e-4)
scheduler = optim.lr_scheduler.StepLR(optimizer, step_size=50, gamma=0.5)
# train_dataset, test_dataset = train_test_split(dataset, 0.8)
train_dataset = dataset[:int(len(dataset) * 0.9)]
test_dataset = dataset[int(len(dataset) * 0.9):]
train_loader = DataLoader(train_dataset, batch_size=4, shuffle=True)
test_loader = DataLoader(test_dataset, batch_size=6, shuffle=False)
# select one sample from the test dataset to visualize
test_data = test_dataset[10]
os.makedirs(model_dir, exist_ok=True)
t1 = time.time()
for epoch in range(1):
model.train()
loss_all = 0
accuracy_all = 0
i_sample = 0
for data in train_loader:
model.train()
# i_sample += 1
# if i_sample > 160:
# break
data = data.to(device)
x, edge_index, edge_attr = data.x, data.edge_index, data.edge_attr
optimizer.zero_grad()
out = model(x, edge_index, edge_attr)
# out = model(x, edge_index)
loss = torch.nn.functional.mse_loss(out, data.y)
loss_by_sample = torch.nn.functional.mse_loss(out, data.y, reduction='none')
r2_accuracy = r2_score(data.y.cpu().detach().numpy(), out.cpu().detach().numpy())
loss.backward()
wandb.log({"loss": loss, "r2_accuracy": r2_accuracy})
wandb.log({"loss_by_sample": wandb.Histogram(loss_by_sample.cpu().detach().numpy())})
# writer.add_scalar('Loss/train', loss, i_sample)
# writer.add_scalar('Accuracy/train', r2_accuracy, i_sample)
loss_all += loss.item()
accuracy_all += r2_accuracy
optimizer.step()
# wandb.log({"loss_train": loss_all / len(train_loader), "r2_accuracy_train": accuracy_all / len(train_loader)})
visualize_prediction(test_data, model, epoch, device=device)
model.eval()
with torch.no_grad():
loss_all_test = []
for data in test_loader:
data = data.to(device)
x, edge_index, edge_attr = data.x, data.edge_index, data.edge_attr
out = model(x, edge_index, edge_attr)
# out = model(x, edge_index)
if data.y.dim() == 1:
data.y = data.y.unsqueeze(-1)
loss = torch.nn.functional.mse_loss(out, data.y)
loss_all_test.append(loss.item())
wandb.log({"loss_test": loss_all_test})
scheduler.step()
print('Epoch: {:02d}, Loss: {:.4f}'.format(epoch, loss_all / len(train_loader)))
# writer.add_scalar('Loss/test', loss_all / len(test_loader), epoch)
# torch.save(model.state_dict(), 'test_cases/burger/CFDError/{}/model_{}.pt'.format(sim_start_time, epoch))
# print('Epoch: {:02d}, Loss: {:.4f}'.format(epoch, loss_all / len(test_loader)))
t2 = time.time()
print('Training time: {:.4f} s'.format(t2 - t1))
torch.save(model.state_dict(), '{}/model.pt'.format(model_dir))
if __name__ == '__main__':
# from args get model type, dataset type and testing configs
args = parse_args()
config_file = args.config
# load config
config = load_yaml(config_file)
# initialize wandb
wandb.init(project="teecnet_exp_2_expressiveness", config=config)
res = config["train_res"]
# # delete the processed dataset
# if os.path.exists(os.path.join(config["dataset_root"], "processed")):
# shutil.rmtree(os.path.join(config["dataset_root"], "processed"))
dataset = initialize_dataset(dataset=config["dataset_type"], root=config["dataset_root"], res_low=res[0], res_high=res[1], pre_transform='interpolate_high')
model = initialize_model(type=config["model_type"], in_channel=config["in_channel"], width=config["width"], out_channel=config["in_channel"], num_layers=config["num_layers"], retrieve_weight=True, num_powers=config["num_powers"])
log_dir = os.path.join(config["log_dir"], config["model_type"], config["dataset_type"], "res_{}_{}".format(res[0], res[1]))
model_dir = os.path.join(config["model_dir"], config["model_type"], "res_{}_{}".format(res[0], res[1]))
train(model, dataset, log_dir, model_dir)