def predict(input_file):
model = HMMPredictor(feature_type = "crp",
model_path = os.path.join(MODEL_DIR, "uniformcrp"),
lda = None,
window_size = 4410,
variance_filter = 0.16,
min_frames = 4,
plot_variance = False,
frame_split = None,
group_filter = 4,
max_count_filter = True)
predictions = model.run(input_file)
if predictions is not None and len(predictions):
return chord.decode(int(predictions[0]));
else:
return ""
# same as model = default_crp()
model = HMMPredictor(feature_type = "crp",
model_path = os.path.join(MODEL_DIR, "uniformcrp"),
lda = None,
window_size = 4410,
variance_filter = 0.16,
min_frames = 4,
plot_variance = False,
frame_split = None,
group_filter = 4,
max_count_filter = True)
predictions = model.run(input_file)
if predictions is not None and len(predictions):
print chord.decode(int(predictions[0]));
else:
print ""
# parameters can be changed here and parts of the model rerun.
#
# update number of frames to group:
#
# predictions.frame_split = 7
# predictions.process_features()
# predictions.predict()
#
# update group filter:
#
# predictions.group_filter = 5
# predictions.predict()
