-
Notifications
You must be signed in to change notification settings - Fork 3
/
train.py
320 lines (264 loc) · 12.5 KB
/
train.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
261
262
263
264
265
266
267
268
269
270
271
272
273
274
275
276
277
278
279
280
281
282
283
284
285
286
287
288
289
290
291
292
293
294
295
296
297
298
299
300
301
302
303
304
305
306
307
308
309
310
311
312
313
314
315
316
317
318
319
320
import os
import gc
import csv
import time
import shutil
import yaml
import numpy as np
from datetime import datetime
from sklearn.metrics import mean_absolute_error, mean_squared_error
import torch
from torch.optim import AdamW
import torch.nn.functional as F
from torch.utils.tensorboard import SummaryWriter
import torch.multiprocessing
torch.multiprocessing.set_sharing_strategy('file_system')
from utils import adjust_learning_rate
class Normalizer(object):
"""Normalize a Tensor and restore it later. """
def __init__(self, tensor):
"""tensor is taken as a sample to calculate the mean and std"""
self.mean = torch.mean(tensor)
self.std = torch.std(tensor)
def norm(self, tensor):
return (tensor - self.mean) / self.std
def denorm(self, normed_tensor):
return normed_tensor * self.std + self.mean
def state_dict(self):
return {'mean': self.mean,
'std': self.std}
def load_state_dict(self, state_dict):
self.mean = state_dict['mean']
self.std = state_dict['std']
class Trainer(object):
def __init__(self, config):
self.config = config
self.device = self._get_device()
if 'SPICE' in self.config['dataset']['data_dir']:
from dataset.dataset_spice import SPICEWrapper
self.dataset = SPICEWrapper(**self.config['dataset'])
self.prefix = 'spice'
if self.config['model']['name'] == 'SE3Transformer':
from dataset.dataset_spice_dgl import SPICEWrapper
self.dataset = SPICEWrapper(
**self.config['dataset'], cutoff=self.config['model']['cutoff'],
max_num_neighbors=self.config['model']['max_num_neighbors']
)
elif 'ANI-1x' in self.config['dataset']['data_dir']:
from dataset.dataset_ani1x import ANI1XWrapper
self.dataset = ANI1XWrapper(**self.config['dataset'])
self.prefix = 'ani1x'
if self.config['model']['name'] == 'SE3Transformer':
from dataset.dataset_ani1x_dgl import ANI1XWrapper
self.dataset = ANI1XWrapper(
**self.config['dataset'], cutoff=self.config['model']['cutoff'],
max_num_neighbors=self.config['model']['max_num_neighbors']
)
elif 'ANI-1' in self.config['dataset']['data_dir']:
from dataset.dataset_ani1 import ANI1Wrapper
self.dataset = ANI1Wrapper(**self.config['dataset'])
self.prefix = 'ani1'
if self.config['model']['name'] == 'SE3Transformer':
from dataset.dataset_ani1_dgl import ANI1Wrapper
self.dataset = ANI1Wrapper(
**self.config['dataset'], cutoff=self.config['model']['cutoff'],
max_num_neighbors=self.config['model']['max_num_neighbors']
)
elif 'iso17' in self.config['dataset']['data_dir']:
from dataset.dataset_iso17 import ISO17Wrapper
self.dataset = ISO17Wrapper(**self.config['dataset'])
self.prefix = 'iso17'
if self.config['model']['name'] == 'SE3Transformer':
from dataset.dataset_iso17_dgl import ISO17Wrapper
self.dataset = ISO17Wrapper(
**self.config['dataset'], cutoff=self.config['model']['cutoff'],
max_num_neighbors=self.config['model']['max_num_neighbors']
)
elif 'MD22' in self.config['dataset']['data_dir']:
from dataset.dataset_md22 import MD22Wrapper
self.dataset = MD22Wrapper(**self.config['dataset'])
self.prefix = self.config['dataset']['data_dir'].split('/')[-1]
if self.config['model']['name'] == 'SE3Transformer':
from dataset.dataset_md22_dgl import MD22Wrapper
self.dataset = MD22Wrapper(
**self.config['dataset'], cutoff=self.config['model']['cutoff'],
max_num_neighbors=self.config['model']['max_num_neighbors']
)
else:
raise NotImplementedError('Undefined dataset!')
if self.config['model']['name'] == 'TorchMD-Net':
self.model_prefix = 'torchmdnet'
elif self.config['model']['name'] == 'EGNN':
self.model_prefix = 'egnn'
elif self.config['model']['name'] == 'SphereNet':
self.model_prefix = 'spherenet'
elif self.config['model']['name'] == 'SchNet':
self.model_prefix = 'schnet'
elif self.config['model']['name'] == 'SE3Transformer':
self.model_prefix = 'se3transformer'
else:
raise NotImplementedError('Undefined model!')
dir_name = '_'.join([datetime.now().strftime('%b%d_%H-%M-%S'), self.prefix, self.model_prefix])
self.log_dir = os.path.join('runs', dir_name)
self.writer = SummaryWriter(log_dir=self.log_dir)
def _get_device(self):
if torch.cuda.is_available() and self.config['gpu'] != 'cpu':
device = self.config['gpu']
else:
device = 'cpu'
print("Running on:", device)
return device
@staticmethod
def _save_config_file(ckpt_dir):
if not os.path.exists(ckpt_dir):
os.makedirs(ckpt_dir)
shutil.copy('./config.yaml', os.path.join(ckpt_dir, 'config.yaml'))
def loss_fn(self, model, data):
if self.config['model']['name'] == 'SE3Transformer':
pred_e, __ = model(data, self.device)
y = data.y.to(self.device)
loss = F.mse_loss(
pred_e, self.normalizer.norm(y), reduction='mean'
)
else:
data = data.to(self.device)
pred_e, __ = model(data.x, data.pos, data.batch)
loss = F.mse_loss(
pred_e, self.normalizer.norm(data.y), reduction='mean'
)
return pred_e, loss
def train(self):
train_loader, valid_loader, test_loader = self.dataset.get_data_loaders()
labels = []
for i, d in enumerate(train_loader):
labels.append(d.y)
if i % 5000 == 0:
print('normalizing', i)
labels = torch.cat(labels)
self.normalizer = Normalizer(labels)
print(self.normalizer.mean, self.normalizer.std, labels.shape)
del labels
gc.collect() # free memory
if self.config['model']['name'] == 'TorchMD-Net':
from models.torchmdnet import TorchMD_ET
model = TorchMD_ET(**self.config["model"])
elif self.config['model']['name'] == 'EGNN':
from models.egnn import EGNN
model = EGNN(**self.config["model"])
elif self.config['model']['name'] == 'SchNet':
from models.schnet import SchNetWrap
model = SchNetWrap(**self.config["model"], auto_grad=False)
elif self.config['model']['name'] == 'SE3Transformer':
from models.se3transformer import SE3Transformer
model = SE3Transformer(**self.config["model"], auto_grad=False)
self._load_weights(model)
model = model.to(self.device)
if type(self.config['lr']) == str: self.config['lr'] = eval(self.config['lr'])
if type(self.config['min_lr']) == str: self.config['min_lr'] = eval(self.config['min_lr'])
if type(self.config['weight_decay']) == str: self.config['weight_decay'] = eval(self.config['weight_decay'])
optimizer = AdamW(
model.parameters(), self.config['lr'],
weight_decay=self.config['weight_decay'],
)
ckpt_dir = os.path.join(self.writer.log_dir, 'checkpoints')
self._save_config_file(ckpt_dir)
n_iter = 0
valid_n_iter = 0
best_valid_loss = np.inf
for epoch_counter in range(self.config['epochs']):
for bn, data in enumerate(train_loader):
adjust_learning_rate(optimizer, epoch_counter + bn / len(train_loader), self.config)
__, loss = self.loss_fn(model, data)
optimizer.zero_grad()
loss.backward()
optimizer.step()
if n_iter % self.config['log_every_n_steps'] == 0:
self.writer.add_scalar('loss', loss.item(), global_step=n_iter)
self.writer.add_scalar('lr', optimizer.param_groups[0]['lr'], global_step=n_iter)
print(epoch_counter, bn, 'loss', loss.item())
torch.cuda.empty_cache()
n_iter += 1
gc.collect() # free memory
torch.cuda.empty_cache()
# validate the model
valid_rmse = self._validate(model, valid_loader)
self.writer.add_scalar('valid_rmse', valid_rmse, global_step=valid_n_iter)
print('Validation', epoch_counter, 'valid rmse', valid_rmse)
if valid_rmse < best_valid_loss:
best_valid_loss = valid_rmse
torch.save(model.state_dict(), os.path.join(ckpt_dir, 'model.pth'))
valid_n_iter += 1
start_time = time.time()
self._test(model, test_loader)
print('test duration:', time.time() - start_time)
def _load_weights(self, model):
try:
state_dict = torch.load(os.path.join(self.config['load_model'], 'model.pth'), map_location=self.device)
model.load_state_dict(state_dict)
print("Loaded pre-trained model with success.")
except FileNotFoundError:
print("Pre-trained weights not found. Training from scratch.")
return model
def _validate(self, model, valid_loader):
predictions, labels = [], []
model.eval()
for bn, data in enumerate(valid_loader):
pred_e, __ = self.loss_fn(model, data)
pred_e = self.normalizer.denorm(pred_e)
if self.config['model']['name'] == 'SE3Transformer':
y = data.y.to(self.device)
else:
y = data.y
if self.device == 'cpu':
predictions.extend(pred_e.flatten().detach().numpy())
labels.extend(y.flatten().numpy())
else:
predictions.extend(pred_e.flatten().cpu().detach().numpy())
labels.extend(y.cpu().flatten().numpy())
torch.cuda.empty_cache()
gc.collect() # free memory
model.train()
return mean_squared_error(labels, predictions, squared=False)
def _test(self, model, test_loader):
model_path = os.path.join(self.log_dir, 'checkpoints', 'model.pth')
state_dict = torch.load(model_path, map_location=self.device)
model.load_state_dict(state_dict)
print("Loaded {} with success.".format(model_path))
predictions, labels, smiles = [], [], []
model.eval()
for bn, data in enumerate(test_loader):
pred_e, __ = self.loss_fn(model, data)
pred_e = self.normalizer.denorm(pred_e)
if self.model_prefix == 'se3transformer':
label = data.y.to(self.device)
else:
label = data.y
smiles.extend(data.smi)
if 'ani1' in self.prefix:
if self.model_prefix == 'se3transformer':
self_energy = data.self_energy.to(self.device)
pred_e += self_energy
label += self_energy
else:
pred_e += data.self_energy
label += data.self_energy
if self.device == 'cpu':
predictions.extend(pred_e.flatten().detach().numpy())
labels.extend(label.flatten().numpy())
else:
predictions.extend(pred_e.flatten().cpu().detach().numpy())
labels.extend(label.cpu().flatten().numpy())
torch.cuda.empty_cache()
gc.collect() # free memory
rmse = mean_squared_error(labels, predictions, squared=False)
mae = mean_absolute_error(labels, predictions)
with open(os.path.join(self.log_dir, 'results.csv'), mode='w', newline='') as csv_file:
csv_writer = csv.writer(csv_file, delimiter=',', quotechar='"', quoting=csv.QUOTE_MINIMAL)
for i in range(len(labels)):
csv_writer.writerow([smiles[i], predictions[i], labels[i]])
csv_writer.writerow([rmse, mae])
if __name__ == "__main__":
config = yaml.load(open("config.yaml", "r"), Loader=yaml.FullLoader)
print(config)
trainer = Trainer(config)
trainer.train()