update the minimum size of audio

huangmingming

2023-01-30 adcee8828ef5d78b575043954deb662a35e318f7

 egs/aishell/paraformerbert/run.sh

@@ -8,7 +8,7 @@
count=1
gpu_inference=true  # Whether to perform gpu decoding, set false for cpu decoding
# for gpu decoding, inference_nj=ngpu*njob; for cpu decoding, inference_nj=njob
njob=1
njob=5
train_cmd=utils/run.pl
infer_cmd=utils/run.pl

@@ -148,17 +148,17 @@
    cp ${feat_dev_dir}/speech_shape ${feat_dev_dir}/text_shape ${feat_dev_dir}/text_shape.char ${feats_dir}/asr_stats_fbank_zh_char/dev
fi

if ! "${skip_extract_embed}"; then
    local/extract_embeds.sh \
        --bert_model_root ${bert_model_root} \
        --bert_model_name ${bert_model_name} \
        --raw_dataset_path ${feats_dir}
fi

# Training Stage
world_size=$gpu_num  # run on one machine
if [ ${stage} -le 3 ] && [ ${stop_stage} -ge 3 ]; then
    echo "stage 3: Training"
    if ! "${skip_extract_embed}"; then
        echo "extract embeddings..."
        local/extract_embeds.sh \
            --bert_model_root ${bert_model_root} \
            --bert_model_name ${bert_model_name} \
            --raw_dataset_path ${feats_dir}
    fi
    mkdir -p ${exp_dir}/exp/${model_dir}
    mkdir -p ${exp_dir}/exp/${model_dir}/log
    INIT_FILE=${exp_dir}/exp/${model_dir}/ddp_init
@@ -192,6 +192,7 @@
                --resume true \
                --output_dir ${exp_dir}/exp/${model_dir} \
                --config $asr_config \
                --allow_variable_data_keys true \
                --input_size $feats_dim \
                --ngpu $gpu_num \
                --num_worker_count $count \
@@ -199,8 +200,7 @@
                --dist_init_method $init_method \
                --dist_world_size $world_size \
                --dist_rank $rank \
                --allow_variable_data_keys true \
                --local_rank $local_rank 1> $exp_dir/log/train.log.$i 2>&1
                --local_rank $local_rank 1> ${exp_dir}/exp/${model_dir}/log/train.log.$i 2>&1
        } &
        done
        wait
@@ -235,7 +235,7 @@
        fi
        ${infer_cmd} --gpu "${_ngpu}" --max-jobs-run "${_nj}" JOB=1:"${_nj}" "${_logdir}"/asr_inference.JOB.log \
            python -m funasr.bin.asr_inference_launch \
                --batch_size 100 \
                --batch_size 1 \
                --ngpu "${_ngpu}" \
                --njob ${njob} \
                --gpuid_list ${gpuid_list} \