개정판 072a26bd
issue #1336: package test
| DTI_PID/WebServer/run.py | ||
|---|---|---|
| 1 | 1 |
from app import app |
| 2 | 2 |
|
| 3 | 3 |
if __name__ == '__main__': |
| 4 |
app.run(port=8080, debug=True, host='0.0.0.0') |
|
| 5 |
#app.run(debug=False) |
|
| 4 |
#app.run(port=8080, debug=False)#, host='0.0.0.0') |
|
| 5 |
app.run(debug=False) |
|
| DTI_PID/WebServer/symbol_training/src/doftech_dataset.py | ||
|---|---|---|
| 51 | 51 |
img_path = self.img_path_list[index] |
| 52 | 52 |
self.img_list.append(cv2.imread(img_path)) |
| 53 | 53 |
|
| 54 |
''' |
|
| 54 | 55 |
if self.is_training==True: |
| 55 | 56 |
if self.use_rescale == True: |
| 56 | 57 |
new_image_list, new_label_list = self.rescale_func([self.img_list[-1]], [self.anno_list[-1]]) |
| 57 | 58 |
|
| 58 | 59 |
self.img_list.extend(new_image_list) |
| 59 | 60 |
self.anno_list.extend(new_label_list) |
| 60 |
''' |
|
| 61 |
''' |
|
| 62 |
|
|
| 61 | 63 |
if self.is_training==True: |
| 62 | 64 |
if self.use_rescale == True: |
| 63 | 65 |
new_image_list, new_label_list = self.rescale_func(self.img_list, self.anno_list) |
| 64 | 66 |
|
| 65 | 67 |
self.img_list.extend(new_image_list) |
| 66 | 68 |
self.anno_list.extend(new_label_list) |
| 67 |
''' |
|
| 69 |
|
|
| 68 | 70 |
|
| 69 | 71 |
def rescale_func(self, img_list, anno_list): |
| 70 | 72 |
#scale_array = [0.7, 0.8, 0.9, 1.0, 1.1, 1.2, 1.3] |
| DTI_PID/WebServer/symbol_training/train.py | ||
|---|---|---|
| 88 | 88 |
learning_rate_schedule = {"0": 1e-5, "5": 1e-4,
|
| 89 | 89 |
"80": 1e-5, "110": 1e-6} |
| 90 | 90 |
|
| 91 |
training_params = {"batch_size": opt.batch_size,
|
|
| 91 |
training_params = {"batch_size": 1,#opt.batch_size,
|
|
| 92 | 92 |
"shuffle": True, |
| 93 | 93 |
"drop_last": True, |
| 94 | 94 |
"collate_fn": custom_collate_fn} |
| ... | ... | |
| 217 | 217 |
best_epoch = epoch |
| 218 | 218 |
print("SAVE MODEL")
|
| 219 | 219 |
# for debug for each loss |
| 220 |
torch.save(model.state_dict(), os.path.join(opt.saved_path, name + "_only_params_" + str(save_count) + "_" + "{:.5f}".format(best_loss) + ".pth"))
|
|
| 220 |
#torch.save(model.state_dict(), os.path.join(opt.saved_path, name + "_only_params_" + str(save_count) + "_" + "{:.5f}".format(best_loss) + ".pth"))
|
|
| 221 | 221 |
#torch.save(model, os.path.join(opt.saved_path, name + "_whole_model_" + str(save_count) + "_" + "{:.5f}".format(best_loss) + ".pth"))
|
| 222 | 222 |
# save |
| 223 | 223 |
torch.save(model.state_dict(), os.path.join(opt.saved_path, name + "_only_params.pth")) |
| ... | ... | |
| 225 | 225 |
else: |
| 226 | 226 |
save_count += 1 |
| 227 | 227 |
# for debug for each loss |
| 228 |
torch.save(model.state_dict(), os.path.join(opt.saved_path, name + "_only_params_" + str(save_count) + "_" + "{:.5f}".format(te_loss) + ".pth"))
|
|
| 228 |
#torch.save(model.state_dict(), os.path.join(opt.saved_path, name + "_only_params_" + str(save_count) + "_" + "{:.5f}".format(te_loss) + ".pth"))
|
|
| 229 | 229 |
#torch.save(model, os.path.join(opt.saved_path, name + "_whole_model_" + str(save_count) + "_" + "{:.5f}".format(te_loss) + ".pth"))
|
| 230 | 230 |
|
| 231 | 231 |
# Early stopping |
내보내기 Unified diff
