Reproducing Whisper-Style Training Using an Open-Source Toolkit and Publicly Available Data