A lightweight yet powerful audio-to-MIDI converter with pitch bend detection
Python audiolightweightmachine-learningmidimusicpitch-detectionpolyphonicpythontranscriptiontypescript
Default predict() to the default model path from __init__.py.
David Rubinstein6ec8658f25da
diff --git a/basic_pitch/inference.py b/basic_pitch/inference.pyindex 7d79cf8..94b4ab9 100644--- a/basic_pitch/inference.py+++ b/basic_pitch/inference.py@@ -24,7 +24,7 @@ import pathlibfrom typing import Any, Dict, Iterable, List, Optional, Sequence, Tuple, Union, cast-from basic_pitch import CT_PRESENT, ONNX_PRESENT, TF_PRESENT, TFLITE_PRESENT+from basic_pitch import CT_PRESENT, ICASSP_2022_MODEL_PATH, ONNX_PRESENT, TF_PRESENT, TFLITE_PRESENTtry:import tensorflow as tf@@ -413,7 +413,7 @@ def save_note_events(def predict(audio_path: Union[pathlib.Path, str],- model_or_model_path: Union[Model, pathlib.Path, str],+ model_or_model_path: Union[Model, pathlib.Path, str] = ICASSP_2022_MODEL_PATH,onset_threshold: float = 0.5,frame_threshold: float = 0.3,minimum_note_length: float = 127.70,