#!/usr/bin/env bash
# Set bash to 'debug' mode, it will exit on :
# -e 'error', -u 'undefined variable', -o ... 'error in pipeline', -x 'print commands',
set -e
set -u
set -o pipefail

train_set="cgn_train"
valid_set="cgn_valid"
test_sets="cgn_test"

nbpe=20000
s2t_config=conf/train_cgn.yaml
inference_config=conf/decode_s2t_nl.yaml

./s2t.sh \
    --s2t_task s2t_wadp \
    --stage 11 \
    --stop_stage 13 \
    --use_lm false \
    --ngpu 1 \
    --nj 1 \
    --gpu_inference true \
    --inference_nj 1 \
    --feats_type raw \
    --audio_format flac.ark \
    --expdir /espnet/egs2/owsm_v1/s2t1/cgn \
    --token_type bpe \
    --nbpe ${nbpe} \
    --s2t_config "${s2t_config}" \
    --inference_config "${inference_config}" \
    --feats_normalize global_mvn \
    --s2t_args "--model_conf extract_feats_in_collect_stats=false" \
    --s2t_stats_dir /exp/s2t_stats_raw_bpe20000 \
    --train_set "${train_set}" \
    --valid_set "${valid_set}" \
    --test_sets "${test_sets}" \
    --bpe_train_text "dump/raw/${train_set}/text" \
    --bpe_nlsyms data/nlsyms.txt \
    --lm_train_text "dump/raw/${train_set}/text" "$@"