aishell example

游雁

2024-02-19 6a9c21a40882dc59676dc882954b290e0a867960

 examples/aishell/paraformer/run.sh

@@ -1,6 +1,6 @@
#!/usr/bin/env bash

. ./path.sh || exit 1;
workspace=`pwd`

# machines configuration
CUDA_VISIBLE_DEVICES="0,1"
@@ -39,7 +39,7 @@
valid_set=dev
test_sets="dev test"

asr_config=conf/train_asr_paraformer_conformer_12e_6d_2048_256.yaml
asr_config=train_asr_paraformer_conformer_12e_6d_2048_256.yaml
model_dir="baseline_$(basename "${asr_config}" .yaml)_${lang}_${token_type}_${tag}"

#inference_config=conf/decode_asr_transformer_noctc_1best.yaml
@@ -74,19 +74,21 @@
        utils/text2token.py -n 1 -s 1 ${feats_dir}/data/${x}/text > ${feats_dir}/data/${x}/text.org
        mv ${feats_dir}/data/${x}/text.org ${feats_dir}/data/${x}/text

        python funasr/datasets/audio_datasets/scp2jsonl.py \
        ++scp_file_list='["${feats_dir}/data/${x}/wav.scp", "${feats_dir}/data/${x}/text"]' \
        # convert wav.scp text to jsonl
        scp_file_list_arg="++scp_file_list='[\"${feats_dir}/data/${x}/wav.scp\",\"${feats_dir}/data/${x}/text\"]'"
        python ../../../funasr/datasets/audio_datasets/scp2jsonl.py \
        ++data_type_list='["source", "target"]' \
        ++jsonl_file_out=${feats_dir}/data/${x}/audio_datasets.jsonl
        ++jsonl_file_out=${feats_dir}/data/${x}/audio_datasets.jsonl \
        ${scp_file_list_arg}
    done
fi

if [ ${stage} -le 1 ] && [ ${stop_stage} -ge 1 ]; then
    echo "stage 1: Feature and CMVN Generation"
#    utils/compute_cmvn.sh --fbankdir ${feats_dir}/data/${train_set} --cmd "$train_cmd" --nj $nj --feats_dim ${feats_dim} --config_file "$asr_config" --scale 1.0
    python funasr/bin/compute_audio_cmvn.py \
    --config-path "/Users/zhifu/funasr1.0/examples/aishell/conf" \
    --config-name "train_asr_paraformer_conformer_12e_6d_2048_256.yaml" \
    python ../../../funasr/bin/compute_audio_cmvn.py \
    --config-path "${workspace}" \
    --config-name "${asr_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
@@ -116,16 +118,16 @@
if [ ${stage} -le 4 ] && [ ${stop_stage} -ge 4 ]; then
echo "stage 4: ASR Training"

torchrun \
--nnodes 1 \
--nproc_per_node ${gpu_num} \
funasr/bin/train.py \
--config-path "/Users/zhifu/funasr1.0/examples/aishell/conf" \
--config-name "train_asr_paraformer_conformer_12e_6d_2048_256.yaml" \
++train_data_set_list="${feats_dir}/data/${train_set}/audio_datasets.jsonl" \
++cmvn_file="${feats_dir}/data/${train_set}/am.mvn" \
++token_list="${token_list}" \
++output_dir="${exp_dir}/exp/${model_dir}"
  torchrun \
  --nnodes 1 \
  --nproc_per_node ${gpu_num} \
  ../../../funasr/bin/train.py \
  --config-path "${workspace}" \
  --config-name "${asr_config}" \
  ++train_data_set_list="${feats_dir}/data/${train_set}/audio_datasets.jsonl" \
  ++cmvn_file="${feats_dir}/data/${train_set}/am.mvn" \
  ++token_list="${token_list}" \
  ++output_dir="${exp_dir}/exp/${model_dir}"
fi

#