@@ -78,112 +78,112 @@ def test_create_fft_sampled():
7878 assert result .exit_code == 0
7979
8080
81- def test_normalization ():
82- from mnist_cnn .scripts .calculate_normalization import main
83- import pandas as pd
84- from dl_framework .data import get_bundles , open_fft_pair , do_normalisation
85- import re
86- import torch
87-
88- data_path = "./tests/build/mnist"
89- out_path = "./tests/build/mnist/normalization_factors.csv"
90-
91- runner = CliRunner ()
92- options = [data_path , out_path ]
93- result = runner .invoke (main , options )
94- print (traceback .print_exception (* result .exc_info ))
95-
96- assert result .exit_code == 0
97-
98- factors = pd .read_csv (out_path )
99-
100- assert (
101- factors .keys ()
102- == ["train_mean_c0" , "train_std_c0" , "train_mean_c1" , "train_std_c1" , ]
103- ).all ()
104- assert ~ np .isnan (factors .values ).all ()
105- assert ~ np .isinf (factors .values ).all ()
106- assert (factors .values != 0 ).all ()
107-
108- bundle_paths = get_bundles (data_path )
109- bundle_paths = [
110- path for path in bundle_paths if re .findall ("fft_bundle_samp_train" , path .name )
111- ]
112-
113- bundles = [open_fft_pair (bund ) for bund in bundle_paths ]
114-
115- a = np .stack ((bundles [0 ][0 ][:, 0 ], bundles [0 ][0 ][:, 1 ]), axis = 1 )
116-
117- assert np .isclose (do_normalisation (torch .tensor (a ), factors ).mean (), 0 , atol = 1e-1 )
118- assert np .isclose (do_normalisation (torch .tensor (a ), factors ).std (), 1 , atol = 1e-1 )
119-
120-
121- def test_train_cnn ():
122- from mnist_cnn .scripts .train_cnn import main
123-
124- data_path = "./tests/build/mnist"
125- path_model = "./tests/build/mnist/test.model"
126- arch = "UNet_denoise"
127- norm_path = "./tests/build/mnist/normalization_factors.csv"
128- epochs = "5"
129- lr = "1e-3"
130- lr_type = "mse"
131- bs = "2"
132-
133- runner = CliRunner ()
134- options = [
135- data_path ,
136- path_model ,
137- arch ,
138- norm_path ,
139- epochs ,
140- lr ,
141- lr_type ,
142- bs ,
143- "-fourier" ,
144- False ,
145- "-pretrained" ,
146- False ,
147- "-inspection" ,
148- False ,
149- "-test" ,
150- True ,
151- ]
152- result = runner .invoke (main , options )
153- print (traceback .print_exception (* result .exc_info ))
154-
155- assert result .exit_code == 0
156-
157-
158- def test_find_lr ():
159- from mnist_cnn .scripts .find_lr import main
160-
161- data_path = "./tests/build/mnist"
162- arch = "UNet_denoise"
163- norm_path = "./tests/build/mnist/normalization_factors.csv"
164- lr_type = "mse"
165-
166- runner = CliRunner ()
167- options = [
168- data_path ,
169- arch ,
170- data_path ,
171- lr_type ,
172- norm_path ,
173- "-max_iter" ,
174- "400" ,
175- "-min_lr" ,
176- "1e-6" ,
177- "-max_lr" ,
178- "1e-1" ,
179- "-fourier" ,
180- False ,
181- "-pretrained" ,
182- False ,
183- "-save" ,
184- True ,
185- ]
186- result = runner .invoke (main , options )
187- print (traceback .print_exception (* result .exc_info ))
188-
189- assert result .exit_code == 0
81+ # def test_normalization():
82+ # from mnist_cnn.scripts.calculate_normalization import main
83+ # import pandas as pd
84+ # from dl_framework.data import get_bundles, open_fft_pair, do_normalisation
85+ # import re
86+ # import torch
87+
88+ # data_path = "./tests/build/mnist"
89+ # out_path = "./tests/build/mnist/normalization_factors.csv"
90+
91+ # runner = CliRunner()
92+ # options = [data_path, out_path]
93+ # result = runner.invoke(main, options)
94+ # print(traceback.print_exception(*result.exc_info))
95+
96+ # assert result.exit_code == 0
97+
98+ # factors = pd.read_csv(out_path)
99+
100+ # assert (
101+ # factors.keys()
102+ # == ["train_mean_c0", "train_std_c0", "train_mean_c1", "train_std_c1", ]
103+ # ).all()
104+ # assert ~np.isnan(factors.values).all()
105+ # assert ~np.isinf(factors.values).all()
106+ # assert (factors.values != 0).all()
107+
108+ # bundle_paths = get_bundles(data_path)
109+ # bundle_paths = [
110+ # path for path in bundle_paths if re.findall("fft_bundle_samp_train", path.name)
111+ # ]
112+
113+ # bundles = [open_fft_pair(bund) for bund in bundle_paths]
114+
115+ # a = np.stack((bundles[0][0][:, 0], bundles[0][0][:, 1]), axis=1)
116+
117+ # assert np.isclose(do_normalisation(torch.tensor(a), factors).mean(), 0, atol=1e-1)
118+ # assert np.isclose(do_normalisation(torch.tensor(a), factors).std(), 1, atol=1e-1)
119+
120+
121+ # def test_train_cnn():
122+ # from mnist_cnn.scripts.train_cnn import main
123+
124+ # data_path = "./tests/build/mnist"
125+ # path_model = "./tests/build/mnist/test.model"
126+ # arch = "UNet_denoise"
127+ # norm_path = "./tests/build/mnist/normalization_factors.csv"
128+ # epochs = "5"
129+ # lr = "1e-3"
130+ # lr_type = "mse"
131+ # bs = "2"
132+
133+ # runner = CliRunner()
134+ # options = [
135+ # data_path,
136+ # path_model,
137+ # arch,
138+ # norm_path,
139+ # epochs,
140+ # lr,
141+ # lr_type,
142+ # bs,
143+ # "-fourier",
144+ # False,
145+ # "-pretrained",
146+ # False,
147+ # "-inspection",
148+ # False,
149+ # "-test",
150+ # True,
151+ # ]
152+ # result = runner.invoke(main, options)
153+ # print(traceback.print_exception(*result.exc_info))
154+
155+ # assert result.exit_code == 0
156+
157+
158+ # def test_find_lr():
159+ # from mnist_cnn.scripts.find_lr import main
160+
161+ # data_path = "./tests/build/mnist"
162+ # arch = "UNet_denoise"
163+ # norm_path = "./tests/build/mnist/normalization_factors.csv"
164+ # lr_type = "mse"
165+
166+ # runner = CliRunner()
167+ # options = [
168+ # data_path,
169+ # arch,
170+ # data_path,
171+ # lr_type,
172+ # norm_path,
173+ # "-max_iter",
174+ # "400",
175+ # "-min_lr",
176+ # "1e-6",
177+ # "-max_lr",
178+ # "1e-1",
179+ # "-fourier",
180+ # False,
181+ # "-pretrained",
182+ # False,
183+ # "-save",
184+ # True,
185+ # ]
186+ # result = runner.invoke(main, options)
187+ # print(traceback.print_exception(*result.exc_info))
188+
189+ # assert result.exit_code == 0
0 commit comments