4 ms·inference is not usually slow for classical ML (save lightgbm)by bbstats 7mo agoinference is not usually slow for classical ML (save lightgbm)