x_test, y_test, test_num = read_labels_batch_out(
labels_path, (224, 224), (224, 224),
root_images_folder,
batch_size=batch_size,
iteration_num=i,
random_horizontal_flip=False)
if (test_num == 0):
break
predicts = model.predict(x_test, batch_size=batch_size)
for j, value in enumerate(predicts):
predict = np.argmax(value)
truth_val = y_test[j]
confusion_mat.AddValueToConfusionMat(val=predict, label_val=truth_val)
file_writer.Write("[[" + str(j + (batch_size * i)) + "]] " +
str(truth_val) + " {" + str(predict) + "}")
print(
"====================================================================")
file_writer.Write(
"====================================================================")
print("result of iteration number : " + str(i))
file_writer.Write("result of iteration number : " + str(i))
print(" ")
file_writer.Write(" ")
print("---Show Current Score---\n")
file_writer.Write("---Show Current Score---\n")
print(confusion_mat.PrintConfusionMat())
file_writer.Write(confusion_mat.PrintConfusionMat())
print("---Images Means is Set to Zero---")
print("Mean-subtracted values:", zip("BGR", np.array([0, 0, 0])))
else:
mu = np.load(means_images_npy_path).mean(1).mean(1)
transformer.set_mean('data', mu) #Means From .npy Files
print("---Images Means is Set by .npy Files---")
print("Mean-subtracted values:", zip("BGR", mu))
net.blobs['data'].reshape(batch_size, image_color_channel, image_size_h,
image_size_w)
######################################################################################################
######################################################################################################
###################################################################################################
file_writer.Write(
"====================================================================")
print("====================================================================")
print("for each layer, show the output shape")
file_writer.Write("for each layer, show the output shape")
for layer_name, blob in net.blobs.items():
print(layer_name + "\t" + str(blob.data.shape))
file_writer.Write(layer_name + "\t" + str(blob.data.shape))
print("====================================================================")
file_writer.Write(
"====================================================================")
print(
"The param shapes typically have the form (output_channels, input_channels, filter_height, filter_width) (for the weights) and the 1-dimensional shape (output_channels,) (for the biases)."
)
file_writer.Write(
"The param shapes typically have the form (output_channels, input_channels, filter_height, filter_width) (for the weights) and the 1-dimensional shape (output_channels,) (for the biases)."