def noisy(model_config, dataset):
csv_file_name = model_config[
"estimates_path"] + os.path.sep + "noisy" + "_" + dataset + ".csv"
with open(csv_file_name, 'w') as csv_file:
csv_writer = csv.writer(csv_file, delimiter=',')
csv_writer.writerow(
["target_file", "input_file", "pesq", "lsd", "ssnr", "audio_len"])
# Get test set
if dataset == "VCTK":
_, test = Datasets.get_VCTK(
model_config["preprocessed_data_path"] + "/VCTK_8k_DBE",
model_config["input"], model_config["target"])
elif dataset == "DAPS":
_, test = Datasets.get_DAPS(
model_config["preprocessed_data_path"] + "/DAPS_8k_DBE",
model_config["input"], model_config["target"])
for sample in test:
# Evaluate
input_file_name = sample[model_config["input"]]
print("Test file " + input_file_name)
target_file_name = sample[model_config["target"]]
pesq, lsd, ssnr, audio_len = Metrics.Eval(
target_file_name, input_file_name, model_config['expected_sr'])
csv_writer.writerow([
target_file_name, input_file_name, pesq, lsd, ssnr, audio_len
])
print('PESQ:{:.3f} LSD:{:.3f} SSNR:{:.3f}'.format(pesq, lsd, ssnr))
results_df = pd.read_csv(csv_file_name,
usecols=["ssnr", "lsd", "pesq", "audio_len"])
pesq, lsd, ssnr = (results_df[["pesq", "lsd", "ssnr"]].multiply(
results_df["audio_len"], axis=0)).sum() / sum(results_df["audio_len"])
print('Results -> PESQ:{:.3f} LSD:{:.3f} SSNR:{:.3f}'.format(
pesq, lsd, ssnr))
def evaluate_dataset(model_config, dataset, experiment_id, load_model):
'''
For a given input file, saves the prediction made by a given model.
:param model_config: Model configuration
:param dataset: Dataset to evaluate
:param experiment_id: ID of the experiment
:param load_model: Model checkpoint path
'''
output_path = model_config[
"estimates_path"] + os.path.sep + experiment_id + "_" + dataset
csv_file_name = model_config[
"estimates_path"] + os.path.sep + experiment_id + "_" + dataset + ".csv"
with open(csv_file_name, 'w') as csv_file:
csv_writer = csv.writer(csv_file, delimiter=',')
csv_writer.writerow(
["target_file", "output_file", "pesq", "lsd", "ssnr", "audio_len"])
# Get test set
if dataset == "VCTK":
_, test = Datasets.get_VCTK(
model_config["preprocessed_data_path"] + "/VCTK_8k_DBE",
model_config["input"], model_config["target"])
elif dataset == "DAPS":
_, test = Datasets.get_DAPS(
model_config["preprocessed_data_path"] + "/DAPS_8k_DBE",
model_config["input"], model_config["target"])
for sample in test:
file = sample[model_config["input"]].split("/")[-1]
print("Producing estimate for file " + file)
track_audio, sr = Utils.load(sample[model_config["input"]],
sr=model_config['expected_sr'],
mono=True)
prediction_audio = predict(track_audio, model_config,
load_model) # Get estimate
prediction_file_name = os.path.join(output_path,
file) + "_prediction.wav"
# Save estimate as audio file
if not os.path.exists(output_path):
os.makedirs(output_path)
librosa.output.write_wav(prediction_file_name, prediction_audio,
sr)
# Evaluate
target_file_name = sample[model_config["target"]]
pesq, lsd, ssnr, audio_len = Metrics.Eval(
target_file_name, prediction_file_name,
model_config['expected_sr'])
csv_writer.writerow([
target_file_name, prediction_file_name, pesq, lsd, ssnr,
audio_len
])
print('PESQ:{:.3f} LSD:{:.3f} SSNR:{:.3f}'.format(pesq, lsd, ssnr))
results_df = pd.read_csv(csv_file_name,
usecols=["pesq", "lsd", "ssnr", "audio_len"])
pesq, lsd, ssnr = (results_df[["pesq", "lsd", "ssnr"]].multiply(
results_df["audio_len"], axis=0)).sum() / sum(results_df["audio_len"])
print('Results -> PESQ:{:.3f} LSD:{:.3f} SSNR:{:.3f}'.format(
pesq, lsd, ssnr))
