-
Make a new file named
rttm_sd
and copy the rttms generated by sd model (sd/data/aishell4_test/dia_part/splited_rttm/
) into it. -
Generate the evaluation results of Speaker Independent task
python calcer_nospk.py path/to/beam24.decode_time_format nofe/or/fe
- Generate the evaluation results of Speaker Dependent task
python calcer_spk.py path/to/beam24.decode_time_format nofe/or/fe