generated by char-rnn (original karpathy), random samples during all training phases for datasets drinksonus, exwe, arglaaa
sample_exwe_0334_cv_fm_lstm_256_3L_03_lm_lstm_epoch22.26_1.6623.wav_tr.mp3
聲道：1
類型：mp3
neural、network、recurrent、char-rnn、generative
神經、網絡、復發、炭RNN、生成的
資料來源：freesound ( https://freesound.org )