ValueError: Cannot feed value of shape (1, 0, 257) for Tensor 'speech_mix_batch_inputs:0', which has shape '(1, 256, 257)'

Hello,
    I find the line 250 of utils.py - "inputslist[g][0][:,-s_len:-s_len+seq_len,:]" may be wrong. Because  "-s_len+seq_len" may greater than 0, which makes "inp_dict[speech_mix[g]]" none.