From c5fd99a075b45d6f4648b51f920c50dc17300015 Mon Sep 17 00:00:00 2001
From: 92homa1mst <92homa1mst@hft-stuttgart.de>
Date: Fri, 26 Feb 2021 10:43:18 +0100
Subject: [PATCH] Commit for generating prediction report
---
main.py | 60 ++++++++++++++++++++++++++++++---------------------------
1 file changed, 32 insertions(+), 28 deletions(-)
diff --git a/main.py b/main.py
index defe61c..d192f4c 100644
--- a/main.py
+++ b/main.py
@@ -37,31 +37,32 @@ def extract_features(trainingDataDir, trainingDataSubDirs):
create_csv_header()
# Looping over every file inside the subdirectories for feature extraction
for trainingDataSubDir in trainingDataSubDirs:
- for audio_file_name in os.listdir(trainingDataDir/f'{trainingDataSubDir}'):
- if audio_file_name.endswith(".wav"):
- audio_file = trainingDataDir/f'{trainingDataSubDir}/{audio_file_name}'
- print("Extracting Features from Directory "+trainingDataSubDir+" and file "+audio_file.name)
- y, sr = librosa.load(audio_file, mono=True)
- mfcc_features = librosa.feature.mfcc(y=y, sr=sr,
- n_mfcc=(constants.MFCC_RANGE_END - constants.MFCC_RANGE_START))
- to_append = f'{audio_file.name}'
- for mfcc_segment in mfcc_features:
- to_append += f' {np.mean(mfcc_segment)}'
- if trainingDataSubDir == constants.CAR:
- to_append += f' {constants.LIGHT_WEIGHT}'
- elif trainingDataSubDir == constants.BUS:
- to_append += f' {constants.MEDIUM_WEIGHT}'
- elif trainingDataSubDir == constants.TRUCK:
- to_append += f' {constants.HEAVY_WEIGHT}'
- elif trainingDataSubDir == constants.MOTORCYCLE:
- to_append += f' {constants.TWO_WHEELED}'
- elif trainingDataSubDir == constants.TRAM:
- to_append += f' {constants.RAIL_BOUND}'
-
- file = open(constants.FEATURES_CSV_NAME, 'a', newline='')
- with file:
- writer = csv.writer(file)
- writer.writerow(to_append.split())
+ if os.path.isdir(trainingDataSubDir):
+ for audio_file_name in os.listdir(trainingDataDir/f'{trainingDataSubDir}'):
+ if audio_file_name.endswith(".wav"):
+ audio_file = trainingDataDir/f'{trainingDataSubDir}/{audio_file_name}'
+ print("Extracting Features from Directory "+trainingDataSubDir+" and file "+audio_file.name)
+ y, sr = librosa.load(audio_file, mono=True)
+ mfcc_features = librosa.feature.mfcc(y=y, sr=sr,
+ n_mfcc=(constants.MFCC_RANGE_END - constants.MFCC_RANGE_START))
+ to_append = f'{audio_file.name}'
+ for mfcc_segment in mfcc_features:
+ to_append += f' {np.mean(mfcc_segment)}'
+ if trainingDataSubDir == constants.CAR:
+ to_append += f' {constants.LIGHT_WEIGHT}'
+ elif trainingDataSubDir == constants.BUS:
+ to_append += f' {constants.MEDIUM_WEIGHT}'
+ elif trainingDataSubDir == constants.TRUCK:
+ to_append += f' {constants.HEAVY_WEIGHT}'
+ elif trainingDataSubDir == constants.MOTORCYCLE:
+ to_append += f' {constants.TWO_WHEELED}'
+ elif trainingDataSubDir == constants.TRAM:
+ to_append += f' {constants.RAIL_BOUND}'
+
+ file = open(constants.FEATURES_CSV_NAME, 'a', newline='')
+ with file:
+ writer = csv.writer(file)
+ writer.writerow(to_append.split())
def preprocessing_csv_data():
@@ -121,17 +122,20 @@ def train_and_save_model(compiledModel, X_train, y_train, X_test, y_test):
callbacks=[tensorboard_callback])
# Saving the trained model to avoid re-training
- #print(training_history)
compiledModel.save(constants.TRAINED_MODEL)
return training_history
-def predict(X_test, y_test):
+def predict(X_test, y_test, encoder):
print("Predictions.....")
final_predictions = np.argmax(compiled_model.predict(X_test), axis=-1)
target_names = [constants.LIGHT_WEIGHT, constants.MEDIUM_WEIGHT, constants.HEAVY_WEIGHT, constants.TWO_WHEELED,
constants.RAIL_BOUND]
+ test_labels = encoder.inverse_transform(y_test)
+ output_labels = encoder.inverse_transform(final_predictions)
print(classification_report(y_test, final_predictions, target_names=target_names))
+ df = pd.DataFrame({"Actual Output": test_labels, "Predicted Output": output_labels})
+ df.to_csv(r'prediction_report.csv', index=False)
# Changing Directory to Training Dataset Folder
@@ -145,4 +149,4 @@ X_input_features = normalize_data(processed_features_data)
X_train_data, X_test_data, y_train_data, y_test_data = train_test_data_split(X_input_features, target_audio_labels)
compiled_model = create_and_compile_model()
model_training_history = train_and_save_model(compiled_model, X_train_data, y_train_data, X_test_data, y_test_data)
-predict(X_test_data, y_test_data)
+predict(X_test_data, y_test_data, encoder_object)
--
GitLab