| | |
| | | model_dir="baseline_$(basename "${config}" .yaml)_${lang}_${token_type}_${tag}" |
| | | |
| | | |
| | | |
| | | if [ ${stage} -le -1 ] && [ ${stop_stage} -ge -1 ]; then |
| | | echo "stage -1: Data Download" |
| | | mkdir -p ${raw_data} |
| | |
| | | if [ ${stage} -le 2 ] && [ ${stop_stage} -ge 2 ]; then |
| | | echo "stage 2: Dictionary Preparation" |
| | | mkdir -p ${feats_dir}/data/${lang}_token_list/$token_type/ |
| | | |
| | | |
| | | echo "make a dictionary" |
| | | echo "<blank>" > ${token_list} |
| | | echo "<s>" >> ${token_list} |
| | |
| | | |
| | | for dset in ${test_sets}; do |
| | | |
| | | inference_dir="${exp_dir}/exp/${model_dir}/infer-${inference_checkpoint}/${dset}" |
| | | inference_dir="${exp_dir}/exp/${model_dir}/inference-${inference_checkpoint}/${dset}" |
| | | _logdir="${inference_dir}/logdir" |
| | | |
| | | mkdir -p "${_logdir}" |
| | |
| | | ++input="${_logdir}/keys.${JOB}.scp" \ |
| | | ++output_dir="${inference_dir}/${JOB}" \ |
| | | ++device="${inference_device}" \ |
| | | ++batch_size="${inference_batch_size}" |
| | | ++batch_size="${inference_batch_size}" &> ${_logdir}/log.${JOB}.txt |
| | | }& |
| | | |
| | | done |
| | |
| | | done |
| | | |
| | | echo "Computing WER ..." |
| | | cp ${inference_dir}/1best_recog/text ${inference_dir}/1best_recog/text.proc |
| | | cp ${data_dir}/text ${inference_dir}/1best_recog/text.ref |
| | | python utils/postprocess_text_zh.py ${inference_dir}/1best_recog/text ${inference_dir}/1best_recog/text.proc |
| | | python utils/postprocess_text_zh.py ${data_dir}/text ${inference_dir}/1best_recog/text.ref |
| | | python utils/compute_wer.py ${inference_dir}/1best_recog/text.ref ${inference_dir}/1best_recog/text.proc ${inference_dir}/1best_recog/text.cer |
| | | tail -n 3 ${inference_dir}/1best_recog/text.cer |
| | | done |
| | | |
| | | fi |
| | | fi |