examples/aishell/transformer/run.sh
@@ -15,9 +15,9 @@ nj=32 inference_device="cuda" #"cpu" inference_checkpoint="model.pt" inference_checkpoint="model.pt.avg10" inference_scp="wav.scp" inference_batch_size=32 inference_batch_size=1 # data raw_data=../raw_data @@ -78,7 +78,7 @@ --config-name "${config}" \ ++train_data_set_list="${feats_dir}/data/${train_set}/audio_datasets.jsonl" \ ++cmvn_file="${feats_dir}/data/${train_set}/cmvn.json" \ ++dataset_conf.num_workers=$nj fi token_list=${feats_dir}/data/${lang}_token_list/$token_type/tokens.txt
