from utilsTrain import generator from modelLib import dice_coef import h5py import numpy as np batchSize = 1 h5path = "..\\out\\train_true_256.h5" h5file = h5py.File(h5path, "r") n = h5file["image"].shape[0] a = np.arange(n) train_index, test_index = train_test_split(a, test_size=0.2, random_state=42) trainGen = generator(h5file, train_index, batchSize) testGen = generator(h5file, test_index, batchSize) model = basic_unet() bestModelPath = '..\\out\\weights\\UNET_01-loss--1.375.hdf5' model.load_weights(bestModelPath) v = next(trainGen) x = v[0] y_true = v[1] y_pred = model.predict(x) y_true = y_true.flatten()
nTotal = db["RNASeq"].shape[0]
nFeat = db["RNASeq"].shape[1]
n_classes = 33
X = np.arange(nTotal)
y = db["label"][...]
skf = StratifiedKFold(n_splits=5)
skf.get_n_splits(X, y)
kdx = 0
for train_index, test_index in skf.split(X, y):
kdx += 1
train_generator = generator(db, train_index, batch_size=32)
test_generator = generator(db, test_index, batch_size=32)
model = makeModel(modelName)
model.compile(loss='categorical_crossentropy',
optimizer='adam',
metrics=['accuracy'])
print("\nCross Validation Fold : %02d \n" % kdx)
check1 = ModelCheckpoint(os.path.join(
weightsFolder, modelName + "_fold_%02d" % kdx +
"_{epoch:02d}-loss-{val_loss:.3f}.hdf5"),
monitor='val_loss',
save_best_only=True,
mode='auto')
nTotal = db["RNASeq"].shape[0]
nFeat = db["RNASeq"].shape[1]
n_classes = 33
X = np.arange(nTotal)
y = db["label"][...]
X_train, X_test, y_train, y_test = train_test_split(X,
y,
stratify=y,
test_size=0.25,
random_state=42)
train_generator = generator(db, X_train, batch_size=32)
test_generator = generator(db, X_test, batch_size=32)
model = makeModel(modelName)
model.compile(loss='categorical_crossentropy',
optimizer='adam',
metrics=['accuracy'])
check1 = ModelCheckpoint(os.path.join(
weightsFolder, modelName + "_{epoch:02d}-loss-{val_loss:.3f}.hdf5"),
monitor='val_loss',
save_best_only=True,
mode='auto')
check2 = ModelCheckpoint(bestModelPath,
monitor='val_loss',