m = np.eye(256, 256, dtype=np.float64)
m = np.eye(256, 256, dtype=np.uint8) * 255
m = np.concatenate((np.atleast_3d(m), np.zeros((256, 256, 2), dtype=np.uint8)), axis=2)
logger.report_image(
    "image", 
    "image color red", 
    iteration=0, 
    image=m
)

logger.report_media(
    'audio', 'audio wav', iteration=0,
    local_path='/path/to/audio.wav'
)

This was how I was doing it as shown in the doc examples, btw, it works if I report some scalars before reporting any media. But if I report the media first, then the webUI doesnt show the media, but it gets uploaded to the fileserver.