Created by: VictorBebnev

Hi! Does Deep Speech support calculations using FP16 ? We are interested in this problem, because Nvidia v100 shows more TFLOPS with FP16 than FP32. Thanks.