run.py 14.7 KB
Newer Older
1
2
3
4
5
6
7
8
9
10
"""Brain Mapper Run File

Description:
-------------
This file contains all the relevant functions for running BrainMapper.
The network can be ran in one of these modes:
    - train
    - evaluate path
    - evaluate whole

11
12
TODO: Might be worth adding some information on uncertaintiy estimation, later down the line

13
14
15
16
17
18
19
20

Usage
-------------
In order to run the network, in the terminal, the user needs to pass it relevant arguments:
    - (TODO: ADD ARGUMENTS)

"""

21
22
import torch
from utils.data_utils import get_datasets
Andrei-Claudiu Roibu's avatar
Andrei-Claudiu Roibu committed
23
from BrainMapperUNet import BrainMapperUNet
24
import torch.utils.data as data
25
26
from solver import Solver
import os
27
28
from utils.data_logging_utils import LogWriter
import utils.data_evaluation_utils as evaluations
29
30
31
32
33
34
35
36
37
38
39
40
41
42

# Set the default floating point tensor type to FloatTensor

torch.set_default_tensor_type(torch.FloatTensor)

def load_data(data_parameters):
    """Dataset Loader

    This function loads the training and testing datasets.
    TODO: Will need to define if all the training data is loaded as bulk or individually!

    Args:
        data_parameters (dict): Dictionary containing relevant information for the datafiles.
        data_parameters = {
43
44
45
46
47
            'data_directory': 'path/to/directory'
            'train_data_file': 'training_data'
            'train_output_targets': 'training_targets'
            'test_data_file': 'testing_data'
            'test_target_file': 'testing_targets'
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
        }

    Returns:
        train_data (dataset object): Pytorch map-style dataset object, mapping indices to training data samples.
        test_data (dataset object): Pytorch map-style dataset object, mapping indices to testing data samples.

    Raises:
        None

    """
    print("Data is loading...")
    train_data, test_data = get_datasets(data_parameters)
    print("Data has loaded!")
    print("Training dataset size is {}".format(len(train_data)))
    print("Testing dataset size is {}".format(len(test_data)))

    return train_data, test_data
65

66
67
def train(data_parameters, training_parameters, network_parameters, misc_parameters):
    """Training Function
68
    
69
    This function trains a given model using the provided training data.
70
71
72
73
74
75
    Currently, the data loaded is set to have multiple sub-processes. 
    A high enough number of workers assures that CPU computations are efficiently managed, i.e. that the bottleneck is indeed the neural network's forward and backward operations on the GPU (and not data generation)
    Loader memory is also pinned, to speed up data transfer from CPU to GPU  by using the page-locked memory.
    Train data is also re-shuffled at each training epoch. 

    Args:
76
77
78
79
80
81
82
        data_parameters (dict): Dictionary containing relevant information for the datafiles.
        data_parameters = {
            'data_directory': 'path/to/directory'
            'train_data_file': 'training_data'
            'train_output_targets': 'training_targets'
            'test_data_file': 'testing_data'
            'test_target_file': 'testing_targets'
Andrei-Claudiu Roibu's avatar
Andrei-Claudiu Roibu committed
83
        }
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101

        training_parameters(dict): Dictionary containing relevant hyperparameters for training the network.
        training_parameters = {
            'training_batch_size': 5
            'test_batch_size: 5
            'use_pre_trained': False
            'pre_trained_path': 'pre_trained/path'
            'experiment_name': 'experiment_name'
            'learning_rate': 1e-4
            'optimizer_beta': (0.9, 0.999)
            'optimizer_epsilon': 1e-8
            'optimizer_weigth_decay': 1e-5
            'number_of_epochs': 10
            'loss_log_period': 50
            'learning_rate_scheduler_step_size': 3
            'learning_rate_scheduler_gamma': 1e-1
            'use_last_checkpoint': True
            'final_model_output_file': 'path/to/model'
102
        }
103
104
105

        network_parameters (dict): Contains information relevant parameters 
        network_parameters= {
106
107
108
109
110
111
112
113
114
115
116
117
118
            'kernel_heigth': 5
            'kernel_width': 5
            'kernel_classification': 1
            'input_channels': 1
            'output_channels': 64
            'convolution_stride': 1
            'dropout': 0.2
            'pool_kernel_size': 2
            'pool_stride': 2
            'up_mode': 'upconv'
            'number_of_classes': 1
        }

119
120
121
122
123
124
125
126
127
        misc_parameters (dict): Dictionary of aditional hyperparameters
        misc_parameters = {
            'save_model_directory': 'directory_name'
            'model_name': 'BrainMapper'
            'logs_directory': 'log-directory'
            'device': 1
            'experiments_directory': 'experiments-directory'
        }

128
129
130
131
132
133
134
135
136
137
138
    Returns:
        None

    Raises:
        None
    """

    train_data, test_data = load_data(data_parameters)

    train_loader = data.DataLoader(
        dataset= train_data,
139
        batch_size= training_parameters['training_batch_size'],
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
        shuffle= True,
        num_workers= 4,
        pin_memory= True
    )

    test_loader = data.DataLoader(
        dataset= test_data,
        batch_size= training_parameters['test_batch_size'],
        shuffle= False,
        num_workers= 4,
        pin_memory= True 
    )

    if training_parameters['use_pre_trained']:
        BrainMapperModel = torch.load(training_parameters['pre_trained_path'])
    else:
        BrainMapperModel = BrainMapperUNet(network_parameters)

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
    solver = Solver(model= BrainMapperModel,
                    device= misc_parameters['device'],
                    number_of_classes= network_parameters['number_of_classes'],
                    experiment_name= training_parameters['experiment_name'],
                    optimizer_arguments = {'lr': training_parameters['learning_rate'],
                                            'betas': training_parameters['optimizer_beta'],
                                            'eps': training_parameters['optimizer_epsilon'],
                                            'weight_decay': training_parameters['optimizer_weigth_decay']
                                            },
                    model_name = misc_parameters['model_name'],
                    number_epochs = training_parameters['number_of_epochs'],
                    loss_log_period = training_parameters['loss_log_period'],
                    learning_rate_scheduler_step_size = training_parameters['learning_rate_scheduler_step_size'],
                    learning_rate_scheduler_gamma = training_parameters['learning_rate_scheduler_gamma'],
                    use_last_checkpoint = training_parameters['use_last_checkpoint'],
                    experiment_directory = misc_parameters['experiments_directory'],
                    logs_directory = misc_parameters['logs_directory']
                    )

    solver.train(train_loader, test_loader)

    model_output_path = os.path.join(misc_parameters['save_model_directory'], training_parameters['final_model_output_file'])
    BrainMapperModel.save(model_output_path)

    print("Final Model Saved in: {}".format(model_output_path))
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
def evaluate_score(data_parameters, training_parameters, network_parameters, misc_parameters, evaluation_parameters):
    """Mapping Score Evaluator

    This function evaluates a given trained model by calculating the it's dice score prediction.

    Args:
        data_parameters (dict): Dictionary containing relevant information for the datafiles.
        data_parameters = {
            'data_directory': 'path/to/directory'
            'train_data_file': 'training_data'
            'train_output_targets': 'training_targets'
            'test_data_file': 'testing_data'
            'test_target_file': 'testing_targets'
        }

        training_parameters(dict): Dictionary containing relevant hyperparameters for training the network.
        training_parameters = {
            'training_batch_size': 5
            'test_batch_size: 5
            'use_pre_trained': False
            'pre_trained_path': 'pre_trained/path'
            'experiment_name': 'experiment_name'
            'learning_rate': 1e-4
            'optimizer_beta': (0.9, 0.999)
            'optimizer_epsilon': 1e-8
            'optimizer_weigth_decay': 1e-5
            'number_of_epochs': 10
            'loss_log_period': 50
            'learning_rate_scheduler_step_size': 3
            'learning_rate_scheduler_gamma': 1e-1
            'use_last_checkpoint': True
            'final_model_output_file': 'path/to/model'
        }

        network_parameters (dict): Contains information relevant parameters 
        network_parameters= {
            'kernel_heigth': 5
            'kernel_width': 5
            'kernel_classification': 1
            'input_channels': 1
            'output_channels': 64
            'convolution_stride': 1
            'dropout': 0.2
            'pool_kernel_size': 2
            'pool_stride': 2
            'up_mode': 'upconv'
            'number_of_classes': 1
        }
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
        misc_parameters (dict): Dictionary of aditional hyperparameters
        misc_parameters = {
            'save_model_directory': 'directory_name'
            'model_name': 'BrainMapper'
            'logs_directory': 'log-directory'
            'device': 1
            'experiments_directory': 'experiments-directory'
        }

        evaluation_parameters (dict): Dictionary of parameters useful during evaluation.
        evaluation_parameters = {
            'trained_model_path': 'path/to/model'
            'data_directory': 'path/to/data'
            'targets_directory': 'path/to/targets'
            'data_list': 'path/to/datalist.txt/
            'orientation': 'coronal'
            'saved_predictions_directory': 'directory-of-saved-predictions'
        }

    Returns:
        None

    Raises:
        None
    """

    logWriter = LogWriter(number_of_classes= network_parameters['number_of_classes'],
                        logs_directory= misc_parameters['logs_directory'], 
                        experiment_name= training_parameters['experiment_name']
                        )

    prediction_output_path = os.path.join(misc_parameters['experiments_directory'],
                                            training_parameters['experiment_name'],
                                            evaluation_parameters['saved_predictions_directory']
                                            )

    average_dice_score = evaluations.evaluate_dice_score(trained_model_path= evaluation_parameters['trained_model_path'],
                                                        number_of_classes= network_parameters['number_of_classes'],
                                                        data_directory= evaluation_parameters['data_directory'],
                                                        targets_directory= evaluation_parameters['targets_directory'],
                                                        data_list= evaluation_parameters['data_list'],
                                                        orientation= evaluation_parameters['orientation'],
                                                        prediction_output_path= prediction_output_path,
                                                        device= misc_parameters['device'],
277
                                                        LogWriter= logWriter
278
279
280
281
                                                        )

    logWriter.close()

282
283
def evaluate_mapping(mapping_evaluation_parameters):
    """Mapping Evaluator
284

285
286
    This function passes through the network an input and generates the rsfMRI outputs.
    This function allows the user to either use one or two or three paths.
287

288
289
290
291
    The convention for the different model paths is as follows:
    - model1: coronal
    - model2: axial
    - model3: saggital
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
321
322
323
324
325
326
327
328
329
330
331
332
333
334
335
336
337
338
339
340
341
342
343
344
345
346
347
348
349
350
351
352
353
354
355
356
357
358
359
360
361
362
363
364
365
    However, this convention can be changed either bellow or the settings file.

    Args:
        mapping_evaluation_parameters (dict): Dictionary of parameters useful during mapping evaluation.
        mapping_evaluation_parameters = {
            'trained_model1_path': 'path/to/model1'
            'trained_model2_path': 'path/to/model2'
            'trained_model3_path': 'path/to/model3'
            'data_directory': 'path/to/data'
            'data_list': 'path/to/datalist.txt/
            'orientation1': 'coronal'
            'orientation2': 'axial'
            'orientation3': 'sagittal'
            'prediction_output_path': 'directory-of-saved-predictions'
            'batch_size': 2
            'device': 0
            'exit_on_error': True
            'number_of_paths': 3
        }

    Returns:
        None

    Raises:
        None
    """
    trained_model1_path = mapping_evaluation_parameters['trained_model1_path']
    trained_model2_path = mapping_evaluation_parameters['trained_model2_path']
    trained_model3_path = mapping_evaluation_parameters['trained_model3_path']
    data_directory = mapping_evaluation_parameters['data_directory']
    data_list = mapping_evaluation_parameters['data_list']
    orientation1 = mapping_evaluation_parameters['orientation1']
    orientation2 = mapping_evaluation_parameters['orientation2']
    orientation3 =mapping_evaluation_parameters['orientation3']
    prediction_output_path = mapping_evaluation_parameters['prediction_output_path']
    batch_size = mapping_evaluation_parameters['batch_size']
    device= mapping_evaluation_parameters['device']
    exit_on_error = mapping_evaluation_parameters['exit_on_error']
    

    if mapping_evaluation_parameters['number_of_paths'] == 1:
        evaluations.evaluate_single_path(trained_model_path= trained_model1_path,
                        data_directory,
                        data_list,
                        orientation= orientation1,
                        prediction_output_path,
                        batch_size,
                        device= device,
                        exit_on_error= exit_on_error):
    elif mapping_evaluation_parameters['number_of_paths'] == 2:
        evaluations.evaluate_two_paths(trained_model1_path,
                        trained_model2_path,
                        data_directory,
                        data_list,
                        orientation1,
                        orientation2,
                        prediction_output_path,
                        batch_size,
                        device= device,
                        exit_on_error= exit_on_error)
    elif mapping_evaluation_parameters['number_of_paths'] == 3:
        evaluations.evaluate_all_paths(trained_model1_path,
                        trained_model2_path,
                        trained_model3_path,
                        data_directory,
                        data_list,
                        orientation1,
                        orientation2,
                        orientation3,
                        prediction_output_path,
                        batch_size,
                        device= device
                        exit_on_error= exit_on_error)
366
367

def delete_files():
368
369
    """ Function which clears contents (like experiments or logs)
    """
370
371
372
373
    pass

if __name__ == '__main__':
    pass