[TTS]soft link for shell in example, add skip_copy_wave in norm stage of G… (#2851)

soft link for shell in example, add skip_copy_wave in norm stage of GANVocoders to save disk

examples/aishell3/tts3/path.sh

@@ -1,13 +0,0 @@
-#!/bin/bash
-export MAIN_ROOT=`realpath ${PWD}/../../../`
-
-export PATH=${MAIN_ROOT}:${MAIN_ROOT}/utils:${PATH}
-export LC_ALL=C
-
-export PYTHONDONTWRITEBYTECODE=1
-# Use UTF-8 in Python to avoid UnicodeDecodeError when LC_ALL=C
-export PYTHONIOENCODING=UTF-8
-export PYTHONPATH=${MAIN_ROOT}:${PYTHONPATH}
-
-MODEL=fastspeech2
-export BIN_DIR=${MAIN_ROOT}/paddlespeech/t2s/exps/${MODEL}

examples/aishell3/tts3/path.sh

@@ -0,0 +1 @@
+../../csmsc/tts3/path.sh

examples/aishell3/vc0/path.sh

@@ -1,13 +0,0 @@
-#!/bin/bash
-export MAIN_ROOT=`realpath ${PWD}/../../../`
-
-export PATH=${MAIN_ROOT}:${MAIN_ROOT}/utils:${PATH}
-export LC_ALL=C
-
-export PYTHONDONTWRITEBYTECODE=1
-# Use UTF-8 in Python to avoid UnicodeDecodeError when LC_ALL=C
-export PYTHONIOENCODING=UTF-8
-export PYTHONPATH=${MAIN_ROOT}:${PYTHONPATH}
-
-MODEL=tacotron2
-export BIN_DIR=${MAIN_ROOT}/paddlespeech/t2s/exps/${MODEL}

examples/aishell3/vc0/path.sh

@@ -0,0 +1 @@
+../../csmsc/tts0/path.sh

examples/aishell3/vc1/local/train.sh

@@ -1,13 +0,0 @@
-#!/bin/bash
-
-config_path=$1
-train_output_path=$2
-
-python3 ${BIN_DIR}/train.py \
-    --train-metadata=dump/train/norm/metadata.jsonl \
-    --dev-metadata=dump/dev/norm/metadata.jsonl \
-    --config=${config_path} \
-    --output-dir=${train_output_path} \
-    --ngpu=2 \
-    --phones-dict=dump/phone_id_map.txt \
-    --voice-cloning=True

examples/aishell3/vc1/local/train.sh

@@ -0,0 +1 @@
+../../vc0/local/train.sh

examples/aishell3/vc1/path.sh

@@ -1,13 +0,0 @@
-#!/bin/bash
-export MAIN_ROOT=`realpath ${PWD}/../../../`
-
-export PATH=${MAIN_ROOT}:${MAIN_ROOT}/utils:${PATH}
-export LC_ALL=C
-
-export PYTHONDONTWRITEBYTECODE=1
-# Use UTF-8 in Python to avoid UnicodeDecodeError when LC_ALL=C
-export PYTHONIOENCODING=UTF-8
-export PYTHONPATH=${MAIN_ROOT}:${PYTHONPATH}
-
-MODEL=fastspeech2
-export BIN_DIR=${MAIN_ROOT}/paddlespeech/t2s/exps/${MODEL}

examples/aishell3/vc1/path.sh

@@ -0,0 +1 @@
+../../csmsc/tts3/path.sh

examples/aishell3/vc2/local/synthesize.sh

@@ -1,20 +0,0 @@
-#!/bin/bash
-
-config_path=$1
-train_output_path=$2
-ckpt_name=$3
-
-python3 ${BIN_DIR}/../synthesize.py \
-    --am=fastspeech2_aishell3 \
-    --am_config=${config_path} \
-    --am_ckpt=${train_output_path}/checkpoints/${ckpt_name} \
-    --am_stat=dump/train/speech_stats.npy \
-    --voc=pwgan_aishell3 \
-    --voc_config=pwg_aishell3_ckpt_0.5/default.yaml \
-    --voc_ckpt=pwg_aishell3_ckpt_0.5/snapshot_iter_1000000.pdz \
-    --voc_stat=pwg_aishell3_ckpt_0.5/feats_stats.npy \
-    --test_metadata=dump/test/norm/metadata.jsonl \
-    --output_dir=${train_output_path}/test \
-    --phones_dict=dump/phone_id_map.txt \
-    --speaker_dict=dump/speaker_id_map.txt \
-    --voice-cloning=True

examples/aishell3/vc2/local/synthesize.sh

@@ -0,0 +1 @@
+../../vc1/local/synthesize.sh

examples/aishell3/vc2/local/train.sh

@@ -1,13 +0,0 @@
-#!/bin/bash
-
-config_path=$1
-train_output_path=$2
-
-python3 ${BIN_DIR}/train.py \
-    --train-metadata=dump/train/norm/metadata.jsonl \
-    --dev-metadata=dump/dev/norm/metadata.jsonl \
-    --config=${config_path} \
-    --output-dir=${train_output_path} \
-    --ngpu=2 \
-    --phones-dict=dump/phone_id_map.txt \
-    --voice-cloning=True

examples/aishell3/vc2/local/train.sh

@@ -0,0 +1 @@
+../../vc0/local/train.sh

examples/aishell3/vc2/path.sh

@@ -1,13 +0,0 @@
-#!/bin/bash
-export MAIN_ROOT=`realpath ${PWD}/../../../`
-
-export PATH=${MAIN_ROOT}:${MAIN_ROOT}/utils:${PATH}
-export LC_ALL=C
-
-export PYTHONDONTWRITEBYTECODE=1
-# Use UTF-8 in Python to avoid UnicodeDecodeError when LC_ALL=C
-export PYTHONIOENCODING=UTF-8
-export PYTHONPATH=${MAIN_ROOT}:${PYTHONPATH}
-
-MODEL=fastspeech2
-export BIN_DIR=${MAIN_ROOT}/paddlespeech/t2s/exps/${MODEL}

examples/aishell3/vc2/path.sh

@@ -0,0 +1 @@
+../../csmsc/tts3/path.sh

examples/aishell3/voc1/local/preprocess.sh

@@ -42,14 +42,18 @@ if [ ${stage} -le 3 ] && [ ${stop_stage} -ge 3 ]; then
     python3 ${BIN_DIR}/../normalize.py \
         --metadata=dump/train/raw/metadata.jsonl \
         --dumpdir=dump/train/norm \
-        --stats=dump/train/feats_stats.npy
+        --stats=dump/train/feats_stats.npy \
+        --skip-wav-copy
+
     python3 ${BIN_DIR}/../normalize.py \
         --metadata=dump/dev/raw/metadata.jsonl \
         --dumpdir=dump/dev/norm \
-        --stats=dump/train/feats_stats.npy
+        --stats=dump/train/feats_stats.npy \
+        --skip-wav-copy
     
     python3 ${BIN_DIR}/../normalize.py \
         --metadata=dump/test/raw/metadata.jsonl \
         --dumpdir=dump/test/norm \
-        --stats=dump/train/feats_stats.npy
+        --stats=dump/train/feats_stats.npy \
+        --skip-wav-copy
 fi

examples/aishell3/voc1/local/synthesize.sh

@@ -1,14 +0,0 @@
-#!/bin/bash
-
-config_path=$1
-train_output_path=$2
-ckpt_name=$3
-
-FLAGS_allocator_strategy=naive_best_fit \
-FLAGS_fraction_of_gpu_memory_to_use=0.01 \
-python3 ${BIN_DIR}/../synthesize.py \
-    --config=${config_path} \
-    --checkpoint=${train_output_path}/checkpoints/${ckpt_name} \
-    --test-metadata=dump/test/norm/metadata.jsonl \
-    --output-dir=${train_output_path}/test \
-    --generator-type=pwgan

examples/aishell3/voc1/local/synthesize.sh

@@ -0,0 +1 @@
+../../../csmsc/voc1/local/synthesize.sh

examples/aishell3/voc1/local/train.sh

@@ -1,13 +0,0 @@
-#!/bin/bash
-
-config_path=$1
-train_output_path=$2
-
-FLAGS_cudnn_exhaustive_search=true \
-FLAGS_conv_workspace_size_limit=4000 \
-python ${BIN_DIR}/train.py \
-    --train-metadata=dump/train/norm/metadata.jsonl \
-    --dev-metadata=dump/dev/norm/metadata.jsonl \
-    --config=${config_path} \
-    --output-dir=${train_output_path} \
-    --ngpu=1

examples/aishell3/voc1/local/train.sh

@@ -0,0 +1 @@
+../../../csmsc/voc1/local/train.sh

examples/aishell3/voc1/path.sh

@@ -1,13 +0,0 @@
-#!/bin/bash
-export MAIN_ROOT=`realpath ${PWD}/../../../`
-
-export PATH=${MAIN_ROOT}:${MAIN_ROOT}/utils:${PATH}
-export LC_ALL=C
-
-export PYTHONDONTWRITEBYTECODE=1
-# Use UTF-8 in Python to avoid UnicodeDecodeError when LC_ALL=C
-export PYTHONIOENCODING=UTF-8
-export PYTHONPATH=${MAIN_ROOT}:${PYTHONPATH}
-
-MODEL=parallelwave_gan
-export BIN_DIR=${MAIN_ROOT}/paddlespeech/t2s/exps/gan_vocoder/${MODEL}

examples/aishell3/voc1/path.sh

@@ -0,0 +1 @@
+../../csmsc/voc1/path.sh

examples/aishell3/voc5/local/preprocess.sh

@@ -1,55 +0,0 @@
-#!/bin/bash
-
-stage=0
-stop_stage=100
-
-config_path=$1
-
-if [ ${stage} -le 0 ] && [ ${stop_stage} -ge 0 ]; then
-    # get durations from MFA's result
-    echo "Generate durations.txt from MFA results ..."
-    python3 ${MAIN_ROOT}/utils/gen_duration_from_textgrid.py \
-        --inputdir=./aishell3_alignment_tone \
-        --output=durations.txt \
-        --config=${config_path}
-fi
-
-if [ ${stage} -le 1 ] && [ ${stop_stage} -ge 1 ]; then
-    # extract features
-    echo "Extract features ..."
-    python3 ${BIN_DIR}/../preprocess.py \
-        --rootdir=~/datasets/data_aishell3/ \
-        --dataset=aishell3 \
-        --dumpdir=dump \
-        --dur-file=durations.txt \
-        --config=${config_path} \
-        --cut-sil=True \
-        --num-cpu=20
-fi
-
-if [ ${stage} -le 2 ] && [ ${stop_stage} -ge 2 ]; then
-    # get features' stats(mean and std)
-    echo "Get features' stats ..."
-    python3 ${MAIN_ROOT}/utils/compute_statistics.py \
-        --metadata=dump/train/raw/metadata.jsonl \
-        --field-name="feats"
-fi
-
-if [ ${stage} -le 3 ] && [ ${stop_stage} -ge 3 ]; then
-    # normalize, dev and test should use train's stats
-    echo "Normalize ..."
-   
-    python3 ${BIN_DIR}/../normalize.py \
-        --metadata=dump/train/raw/metadata.jsonl \
-        --dumpdir=dump/train/norm \
-        --stats=dump/train/feats_stats.npy
-    python3 ${BIN_DIR}/../normalize.py \
-        --metadata=dump/dev/raw/metadata.jsonl \
-        --dumpdir=dump/dev/norm \
-        --stats=dump/train/feats_stats.npy
-    
-    python3 ${BIN_DIR}/../normalize.py \
-        --metadata=dump/test/raw/metadata.jsonl \
-        --dumpdir=dump/test/norm \
-        --stats=dump/train/feats_stats.npy
-fi

examples/aishell3/voc5/local/preprocess.sh

@@ -0,0 +1 @@
+../../voc1/local/preprocess.sh

examples/aishell3/voc5/local/synthesize.sh

@@ -1,14 +0,0 @@
-#!/bin/bash
-
-config_path=$1
-train_output_path=$2
-ckpt_name=$3
-
-FLAGS_allocator_strategy=naive_best_fit \
-FLAGS_fraction_of_gpu_memory_to_use=0.01 \
-python3 ${BIN_DIR}/../synthesize.py \
-    --config=${config_path} \
-    --checkpoint=${train_output_path}/checkpoints/${ckpt_name} \
-    --test-metadata=dump/test/norm/metadata.jsonl \
-    --output-dir=${train_output_path}/test \
-    --generator-type=hifigan

examples/aishell3/voc5/local/synthesize.sh

@@ -0,0 +1 @@
+../../../csmsc/voc5/local/synthesize.sh

examples/aishell3/voc5/local/train.sh

@@ -1,13 +0,0 @@
-#!/bin/bash
-
-config_path=$1
-train_output_path=$2
-
-FLAGS_cudnn_exhaustive_search=true \
-FLAGS_conv_workspace_size_limit=4000 \
-python ${BIN_DIR}/train.py \
-    --train-metadata=dump/train/norm/metadata.jsonl \
-    --dev-metadata=dump/dev/norm/metadata.jsonl \
-    --config=${config_path} \
-    --output-dir=${train_output_path} \
-    --ngpu=1

examples/aishell3/voc5/local/train.sh

@@ -0,0 +1 @@
+../../../csmsc/voc1/local/train.sh

examples/aishell3/voc5/path.sh

@@ -1,13 +0,0 @@
-#!/bin/bash
-export MAIN_ROOT=`realpath ${PWD}/../../../`
-
-export PATH=${MAIN_ROOT}:${MAIN_ROOT}/utils:${PATH}
-export LC_ALL=C
-
-export PYTHONDONTWRITEBYTECODE=1
-# Use UTF-8 in Python to avoid UnicodeDecodeError when LC_ALL=C
-export PYTHONIOENCODING=UTF-8
-export PYTHONPATH=${MAIN_ROOT}:${PYTHONPATH}
-
-MODEL=hifigan
-export BIN_DIR=${MAIN_ROOT}/paddlespeech/t2s/exps/gan_vocoder/${MODEL}

examples/aishell3/voc5/path.sh

@@ -0,0 +1 @@
+../../csmsc/voc5/path.sh

examples/aishell3_vctk/ernie_sat/local/synthesize.sh

@@ -1,25 +0,0 @@
-#!/bin/bash
-
-config_path=$1
-train_output_path=$2
-ckpt_name=$3
-
-stage=0
-stop_stage=0
-
-# hifigan
-if [ ${stage} -le 0 ] && [ ${stop_stage} -ge 0 ]; then
-    FLAGS_allocator_strategy=naive_best_fit \
-    FLAGS_fraction_of_gpu_memory_to_use=0.01 \
-    python3 ${BIN_DIR}/synthesize.py \
-        --erniesat_config=${config_path} \
-        --erniesat_ckpt=${train_output_path}/checkpoints/${ckpt_name} \
-        --erniesat_stat=dump/train/speech_stats.npy \
-        --voc=hifigan_aishell3 \
-        --voc_config=hifigan_aishell3_ckpt_0.2.0/default.yaml \
-        --voc_ckpt=hifigan_aishell3_ckpt_0.2.0/snapshot_iter_2500000.pdz \
-        --voc_stat=hifigan_aishell3_ckpt_0.2.0/feats_stats.npy \
-        --test_metadata=dump/test/norm/metadata.jsonl \
-        --output_dir=${train_output_path}/test \
-        --phones_dict=dump/phone_id_map.txt
-fi

examples/aishell3_vctk/ernie_sat/local/synthesize.sh

@@ -0,0 +1 @@
+../../../aishell3/ernie_sat/local/synthesize.sh

examples/aishell3_vctk/ernie_sat/local/train.sh

@@ -1,12 +0,0 @@
-#!/bin/bash
-
-config_path=$1
-train_output_path=$2
-
-python3 ${BIN_DIR}/train.py \
-    --train-metadata=dump/train/norm/metadata.jsonl \
-    --dev-metadata=dump/dev/norm/metadata.jsonl \
-    --config=${config_path} \
-    --output-dir=${train_output_path} \
-    --ngpu=8 \
-    --phones-dict=dump/phone_id_map.txt

examples/aishell3_vctk/ernie_sat/local/train.sh

@@ -0,0 +1 @@
+../../../aishell3/ernie_sat/local/train.sh

examples/aishell3_vctk/ernie_sat/path.sh

@@ -1,13 +0,0 @@
-#!/bin/bash
-export MAIN_ROOT=`realpath ${PWD}/../../../`
-
-export PATH=${MAIN_ROOT}:${MAIN_ROOT}/utils:${PATH}
-export LC_ALL=C
-
-export PYTHONDONTWRITEBYTECODE=1
-# Use UTF-8 in Python to avoid UnicodeDecodeError when LC_ALL=C
-export PYTHONIOENCODING=UTF-8
-export PYTHONPATH=${MAIN_ROOT}:${PYTHONPATH}
-
-MODEL=ernie_sat
-export BIN_DIR=${MAIN_ROOT}/paddlespeech/t2s/exps/${MODEL}

examples/aishell3_vctk/ernie_sat/path.sh

@@ -0,0 +1 @@
+../../aishell3/ernie_sat/path.sh

examples/csmsc/voc1/local/preprocess.sh

@@ -42,14 +42,18 @@ if [ ${stage} -le 3 ] && [ ${stop_stage} -ge 3 ]; then
     python3 ${BIN_DIR}/../normalize.py \
         --metadata=dump/train/raw/metadata.jsonl \
         --dumpdir=dump/train/norm \
-        --stats=dump/train/feats_stats.npy
+        --stats=dump/train/feats_stats.npy \
+        --skip-wav-copy
+        
     python3 ${BIN_DIR}/../normalize.py \
         --metadata=dump/dev/raw/metadata.jsonl \
         --dumpdir=dump/dev/norm \
-        --stats=dump/train/feats_stats.npy
+        --stats=dump/train/feats_stats.npy \
+        --skip-wav-copy
     
     python3 ${BIN_DIR}/../normalize.py \
         --metadata=dump/test/raw/metadata.jsonl \
         --dumpdir=dump/test/norm \
-        --stats=dump/train/feats_stats.npy
+        --stats=dump/train/feats_stats.npy \
+        --skip-wav-copy
 fi

examples/csmsc/voc3/finetune.sh

@@ -1,64 +0,0 @@
-#!/bin/bash
-
-source path.sh
-
-gpus=0
-stage=0
-stop_stage=100
-
-source ${MAIN_ROOT}/utils/parse_options.sh || exit 1
-
-if [ ${stage} -le 0 ] && [ ${stop_stage} -ge 0 ]; then
-    python3 ${MAIN_ROOT}/paddlespeech/t2s/exps/fastspeech2/gen_gta_mel.py \
-        --fastspeech2-config=fastspeech2_nosil_baker_ckpt_0.4/default.yaml \
-        --fastspeech2-checkpoint=fastspeech2_nosil_baker_ckpt_0.4/snapshot_iter_76000.pdz \
-        --fastspeech2-stat=fastspeech2_nosil_baker_ckpt_0.4/speech_stats.npy \
-        --dur-file=durations.txt \
-        --output-dir=dump_finetune \
-        --phones-dict=fastspeech2_nosil_baker_ckpt_0.4/phone_id_map.txt \
-        --dataset=baker \
-        --rootdir=~/datasets/BZNSYP/
-fi
-
-if [ ${stage} -le 1 ] && [ ${stop_stage} -ge 1 ]; then
-    python3 ${MAIN_ROOT}/utils/link_wav.py \
-        --old-dump-dir=dump \
-        --dump-dir=dump_finetune
-fi
-
-if [ ${stage} -le 2 ] && [ ${stop_stage} -ge 2 ]; then
-    # get features' stats(mean and std)
-    echo "Get features' stats ..."
-    cp dump/train/feats_stats.npy dump_finetune/train/
-fi
-
-if [ ${stage} -le 3 ] && [ ${stop_stage} -ge 3 ]; then
-    # normalize, dev and test should use train's stats
-    echo "Normalize ..."
-   
-    python3 ${BIN_DIR}/../normalize.py \
-        --metadata=dump_finetune/train/raw/metadata.jsonl \
-        --dumpdir=dump_finetune/train/norm \
-        --stats=dump_finetune/train/feats_stats.npy
-    python3 ${BIN_DIR}/../normalize.py \
-        --metadata=dump_finetune/dev/raw/metadata.jsonl \
-        --dumpdir=dump_finetune/dev/norm \
-        --stats=dump_finetune/train/feats_stats.npy
-    
-    python3 ${BIN_DIR}/../normalize.py \
-        --metadata=dump_finetune/test/raw/metadata.jsonl \
-        --dumpdir=dump_finetune/test/norm \
-        --stats=dump_finetune/train/feats_stats.npy
-fi
-
-if [ ${stage} -le 4 ] && [ ${stop_stage} -ge 4 ]; then
-    CUDA_VISIBLE_DEVICES=${gpus} \
-    FLAGS_cudnn_exhaustive_search=true \
-    FLAGS_conv_workspace_size_limit=4000 \
-    python ${BIN_DIR}/train.py \
-        --train-metadata=dump_finetune/train/norm/metadata.jsonl \
-        --dev-metadata=dump_finetune/dev/norm/metadata.jsonl \
-        --config=conf/finetune.yaml \
-        --output-dir=exp/finetune \
-        --ngpu=1
-fi 

examples/csmsc/voc3/finetune.sh

@@ -0,0 +1 @@
+../voc5/finetune.sh

examples/csmsc/voc3/local/preprocess.sh

@@ -1,55 +0,0 @@
-#!/bin/bash
-
-stage=0
-stop_stage=100
-
-config_path=$1
-
-if [ ${stage} -le 0 ] && [ ${stop_stage} -ge 0 ]; then
-    # get durations from MFA's result
-    echo "Generate durations.txt from MFA results ..."
-    python3 ${MAIN_ROOT}/utils/gen_duration_from_textgrid.py \
-        --inputdir=./baker_alignment_tone \
-        --output=durations.txt \
-        --config=${config_path}
-fi
-
-if [ ${stage} -le 1 ] && [ ${stop_stage} -ge 1 ]; then
-    # extract features
-    echo "Extract features ..."
-    python3 ${BIN_DIR}/../preprocess.py \
-        --rootdir=~/datasets/BZNSYP/ \
-        --dataset=baker \
-        --dumpdir=dump \
-        --dur-file=durations.txt \
-        --config=${config_path} \
-        --cut-sil=True \
-        --num-cpu=20
-fi
-
-if [ ${stage} -le 2 ] && [ ${stop_stage} -ge 2 ]; then
-    # get features' stats(mean and std)
-    echo "Get features' stats ..."
-    python3 ${MAIN_ROOT}/utils/compute_statistics.py \
-        --metadata=dump/train/raw/metadata.jsonl \
-        --field-name="feats"
-fi
-
-if [ ${stage} -le 3 ] && [ ${stop_stage} -ge 3 ]; then
-    # normalize, dev and test should use train's stats
-    echo "Normalize ..."
-   
-    python3 ${BIN_DIR}/../normalize.py \
-        --metadata=dump/train/raw/metadata.jsonl \
-        --dumpdir=dump/train/norm \
-        --stats=dump/train/feats_stats.npy
-    python3 ${BIN_DIR}/../normalize.py \
-        --metadata=dump/dev/raw/metadata.jsonl \
-        --dumpdir=dump/dev/norm \
-        --stats=dump/train/feats_stats.npy
-    
-    python3 ${BIN_DIR}/../normalize.py \
-        --metadata=dump/test/raw/metadata.jsonl \
-        --dumpdir=dump/test/norm \
-        --stats=dump/train/feats_stats.npy
-fi

examples/csmsc/voc3/local/preprocess.sh

@@ -0,0 +1 @@
+../../voc1/local/preprocess.sh

examples/csmsc/voc3/local/train.sh

@@ -1,13 +0,0 @@
-#!/bin/bash
-
-config_path=$1
-train_output_path=$2
-
-FLAGS_cudnn_exhaustive_search=true \
-FLAGS_conv_workspace_size_limit=4000 \
-python ${BIN_DIR}/train.py \
-    --train-metadata=dump/train/norm/metadata.jsonl \
-    --dev-metadata=dump/dev/norm/metadata.jsonl \
-    --config=${config_path} \
-    --output-dir=${train_output_path} \
-    --ngpu=1

examples/csmsc/voc3/local/train.sh

@@ -0,0 +1 @@
+../../voc1/local/train.sh

examples/csmsc/voc4/local/preprocess.sh

@@ -1,55 +0,0 @@
-#!/bin/bash
-
-stage=0
-stop_stage=100
-
-config_path=$1
-
-if [ ${stage} -le 0 ] && [ ${stop_stage} -ge 0 ]; then
-    # get durations from MFA's result
-    echo "Generate durations.txt from MFA results ..."
-    python3 ${MAIN_ROOT}/utils/gen_duration_from_textgrid.py \
-        --inputdir=./baker_alignment_tone \
-        --output=durations.txt \
-        --config=${config_path}
-fi
-
-if [ ${stage} -le 1 ] && [ ${stop_stage} -ge 1 ]; then
-    # extract features
-    echo "Extract features ..."
-    python3 ${BIN_DIR}/../preprocess.py \
-        --rootdir=~/datasets/BZNSYP/ \
-        --dataset=baker \
-        --dumpdir=dump \
-        --dur-file=durations.txt \
-        --config=${config_path} \
-        --cut-sil=True \
-        --num-cpu=20
-fi
-
-if [ ${stage} -le 2 ] && [ ${stop_stage} -ge 2 ]; then
-    # get features' stats(mean and std)
-    echo "Get features' stats ..."
-    python3 ${MAIN_ROOT}/utils/compute_statistics.py \
-        --metadata=dump/train/raw/metadata.jsonl \
-        --field-name="feats"
-fi
-
-if [ ${stage} -le 3 ] && [ ${stop_stage} -ge 3 ]; then
-    # normalize, dev and test should use train's stats
-    echo "Normalize ..."
-   
-    python3 ${BIN_DIR}/../normalize.py \
-        --metadata=dump/train/raw/metadata.jsonl \
-        --dumpdir=dump/train/norm \
-        --stats=dump/train/feats_stats.npy
-    python3 ${BIN_DIR}/../normalize.py \
-        --metadata=dump/dev/raw/metadata.jsonl \
-        --dumpdir=dump/dev/norm \
-        --stats=dump/train/feats_stats.npy
-    
-    python3 ${BIN_DIR}/../normalize.py \
-        --metadata=dump/test/raw/metadata.jsonl \
-        --dumpdir=dump/test/norm \
-        --stats=dump/train/feats_stats.npy
-fi

examples/csmsc/voc4/local/preprocess.sh

@@ -0,0 +1 @@
+../../voc1/local/preprocess.sh

examples/csmsc/voc4/local/train.sh

@@ -1,13 +0,0 @@
-#!/bin/bash
-
-config_path=$1
-train_output_path=$2
-
-FLAGS_cudnn_exhaustive_search=true \
-FLAGS_conv_workspace_size_limit=4000 \
-python ${BIN_DIR}/train.py \
-    --train-metadata=dump/train/norm/metadata.jsonl \
-    --dev-metadata=dump/dev/norm/metadata.jsonl \
-    --config=${config_path} \
-    --output-dir=${train_output_path} \
-    --ngpu=1

examples/csmsc/voc4/local/train.sh

@@ -0,0 +1 @@
+../../voc1/local/train.sh

examples/csmsc/voc5/finetune.sh

@@ -39,16 +39,19 @@ if [ ${stage} -le 3 ] && [ ${stop_stage} -ge 3 ]; then
     python3 ${BIN_DIR}/../normalize.py \
         --metadata=dump_finetune/train/raw/metadata.jsonl \
         --dumpdir=dump_finetune/train/norm \
-        --stats=dump_finetune/train/feats_stats.npy
+        --stats=dump_finetune/train/feats_stats.npy \
+        --skip-wav-copy
     python3 ${BIN_DIR}/../normalize.py \
         --metadata=dump_finetune/dev/raw/metadata.jsonl \
         --dumpdir=dump_finetune/dev/norm \
-        --stats=dump_finetune/train/feats_stats.npy
+        --stats=dump_finetune/train/feats_stats.npy \
+        --skip-wav-copy
     
     python3 ${BIN_DIR}/../normalize.py \
         --metadata=dump_finetune/test/raw/metadata.jsonl \
         --dumpdir=dump_finetune/test/norm \
-        --stats=dump_finetune/train/feats_stats.npy
+        --stats=dump_finetune/train/feats_stats.npy \
+        --skip-wav-copy
 fi
 
 if [ ${stage} -le 4 ] && [ ${stop_stage} -ge 4 ]; then

examples/csmsc/voc5/local/preprocess.sh

@@ -1,55 +0,0 @@
-#!/bin/bash
-
-stage=0
-stop_stage=100
-
-config_path=$1
-
-if [ ${stage} -le 0 ] && [ ${stop_stage} -ge 0 ]; then
-    # get durations from MFA's result
-    echo "Generate durations.txt from MFA results ..."
-    python3 ${MAIN_ROOT}/utils/gen_duration_from_textgrid.py \
-        --inputdir=./baker_alignment_tone \
-        --output=durations.txt \
-        --config=${config_path}
-fi
-
-if [ ${stage} -le 1 ] && [ ${stop_stage} -ge 1 ]; then
-    # extract features
-    echo "Extract features ..."
-    python3 ${BIN_DIR}/../preprocess.py \
-        --rootdir=~/datasets/BZNSYP/ \
-        --dataset=baker \
-        --dumpdir=dump \
-        --dur-file=durations.txt \
-        --config=${config_path} \
-        --cut-sil=True \
-        --num-cpu=20
-fi
-
-if [ ${stage} -le 2 ] && [ ${stop_stage} -ge 2 ]; then
-    # get features' stats(mean and std)
-    echo "Get features' stats ..."
-    python3 ${MAIN_ROOT}/utils/compute_statistics.py \
-        --metadata=dump/train/raw/metadata.jsonl \
-        --field-name="feats"
-fi
-
-if [ ${stage} -le 3 ] && [ ${stop_stage} -ge 3 ]; then
-    # normalize, dev and test should use train's stats
-    echo "Normalize ..."
-   
-    python3 ${BIN_DIR}/../normalize.py \
-        --metadata=dump/train/raw/metadata.jsonl \
-        --dumpdir=dump/train/norm \
-        --stats=dump/train/feats_stats.npy
-    python3 ${BIN_DIR}/../normalize.py \
-        --metadata=dump/dev/raw/metadata.jsonl \
-        --dumpdir=dump/dev/norm \
-        --stats=dump/train/feats_stats.npy
-    
-    python3 ${BIN_DIR}/../normalize.py \
-        --metadata=dump/test/raw/metadata.jsonl \
-        --dumpdir=dump/test/norm \
-        --stats=dump/train/feats_stats.npy
-fi

examples/csmsc/voc5/local/preprocess.sh

@@ -0,0 +1 @@
+../../voc1/local/preprocess.sh

examples/csmsc/voc5/local/train.sh

@@ -1,13 +0,0 @@
-#!/bin/bash
-
-config_path=$1
-train_output_path=$2
-
-FLAGS_cudnn_exhaustive_search=true \
-FLAGS_conv_workspace_size_limit=4000 \
-python ${BIN_DIR}/train.py \
-    --train-metadata=dump/train/norm/metadata.jsonl \
-    --dev-metadata=dump/dev/norm/metadata.jsonl \
-    --config=${config_path} \
-    --output-dir=${train_output_path} \
-    --ngpu=1

examples/csmsc/voc5/local/train.sh

@@ -0,0 +1 @@
+../../voc1/local/train.sh

examples/csmsc/voc6/local/preprocess.sh

@@ -42,14 +42,18 @@ if [ ${stage} -le 3 ] && [ ${stop_stage} -ge 3 ]; then
     python3 ${BIN_DIR}/../gan_vocoder/normalize.py \
         --metadata=dump/train/raw/metadata.jsonl \
         --dumpdir=dump/train/norm \
-        --stats=dump/train/feats_stats.npy
+        --stats=dump/train/feats_stats.npy \
+        --skip-wav-copy
+
     python3 ${BIN_DIR}/../gan_vocoder/normalize.py \
         --metadata=dump/dev/raw/metadata.jsonl \
         --dumpdir=dump/dev/norm \
-        --stats=dump/train/feats_stats.npy
+        --stats=dump/train/feats_stats.npy \
+        --skip-wav-copy
     
     python3 ${BIN_DIR}/../gan_vocoder/normalize.py \
         --metadata=dump/test/raw/metadata.jsonl \
         --dumpdir=dump/test/norm \
-        --stats=dump/train/feats_stats.npy
+        --stats=dump/train/feats_stats.npy \
+        --skip-wav-copy
 fi

examples/csmsc/voc6/local/train.sh

@@ -1,13 +0,0 @@
-#!/bin/bash
-
-config_path=$1
-train_output_path=$2
-
-FLAGS_cudnn_exhaustive_search=true \
-FLAGS_conv_workspace_size_limit=4000 \
-python ${BIN_DIR}/train.py \
-    --train-metadata=dump/train/norm/metadata.jsonl \
-    --dev-metadata=dump/dev/norm/metadata.jsonl \
-    --config=${config_path} \
-    --output-dir=${train_output_path} \
-    --ngpu=1

examples/csmsc/voc6/local/train.sh

@@ -0,0 +1 @@
+../../voc1/local/train.sh

examples/ljspeech/tts0/local/train.sh

@@ -1,12 +0,0 @@
-#!/bin/bash
-
-config_path=$1
-train_output_path=$2
-
-python3 ${BIN_DIR}/train.py \
-    --train-metadata=dump/train/norm/metadata.jsonl \
-    --dev-metadata=dump/dev/norm/metadata.jsonl \
-    --config=${config_path} \
-    --output-dir=${train_output_path} \
-    --ngpu=1 \
-    --phones-dict=dump/phone_id_map.txt

examples/ljspeech/tts0/local/train.sh

@@ -0,0 +1 @@
+../../../csmsc/tts0/local/train.sh

examples/ljspeech/tts0/path.sh

@@ -1,13 +0,0 @@
-#!/bin/bash
-export MAIN_ROOT=`realpath ${PWD}/../../../`
-
-export PATH=${MAIN_ROOT}:${MAIN_ROOT}/utils:${PATH}
-export LC_ALL=C
-
-export PYTHONDONTWRITEBYTECODE=1
-# Use UTF-8 in Python to avoid UnicodeDecodeError when LC_ALL=C
-export PYTHONIOENCODING=UTF-8
-export PYTHONPATH=${MAIN_ROOT}:${PYTHONPATH}
-
-MODEL=tacotron2
-export BIN_DIR=${MAIN_ROOT}/paddlespeech/t2s/exps/${MODEL}

examples/ljspeech/tts0/path.sh

@@ -0,0 +1 @@
+../../csmsc/tts0/path.sh

examples/ljspeech/tts3/local/train.sh

@@ -1,12 +0,0 @@
-#!/bin/bash
-
-config_path=$1
-train_output_path=$2
-
-python3 ${BIN_DIR}/train.py \
-    --train-metadata=dump/train/norm/metadata.jsonl \
-    --dev-metadata=dump/dev/norm/metadata.jsonl \
-    --config=${config_path} \
-    --output-dir=${train_output_path} \
-    --ngpu=1 \
-    --phones-dict=dump/phone_id_map.txt

examples/ljspeech/tts3/local/train.sh

@@ -0,0 +1 @@
+../../../csmsc/tts3/local/train.sh

examples/ljspeech/tts3/path.sh

@@ -1,13 +0,0 @@
-#!/bin/bash
-export MAIN_ROOT=`realpath ${PWD}/../../../`
-
-export PATH=${MAIN_ROOT}:${MAIN_ROOT}/utils:${PATH}
-export LC_ALL=C
-
-export PYTHONDONTWRITEBYTECODE=1
-# Use UTF-8 in Python to avoid UnicodeDecodeError when LC_ALL=C
-export PYTHONIOENCODING=UTF-8
-export PYTHONPATH=${MAIN_ROOT}:${PYTHONPATH}
-
-MODEL=fastspeech2
-export BIN_DIR=${MAIN_ROOT}/paddlespeech/t2s/exps/${MODEL}

examples/ljspeech/tts3/path.sh

@@ -0,0 +1 @@
+../../csmsc/tts3/path.sh

examples/ljspeech/voc1/local/preprocess.sh

@@ -42,14 +42,18 @@ if [ ${stage} -le 3 ] && [ ${stop_stage} -ge 3 ]; then
     python3 ${BIN_DIR}/../normalize.py \
         --metadata=dump/train/raw/metadata.jsonl \
         --dumpdir=dump/train/norm \
-        --stats=dump/train/feats_stats.npy
+        --stats=dump/train/feats_stats.npy \
+        --skip-wav-copy
+
     python3 ${BIN_DIR}/../normalize.py \
         --metadata=dump/dev/raw/metadata.jsonl \
         --dumpdir=dump/dev/norm \
-        --stats=dump/train/feats_stats.npy
+        --stats=dump/train/feats_stats.npy \
+        --skip-wav-copy
     
     python3 ${BIN_DIR}/../normalize.py \
         --metadata=dump/test/raw/metadata.jsonl \
         --dumpdir=dump/test/norm \
-        --stats=dump/train/feats_stats.npy
+        --stats=dump/train/feats_stats.npy \
+        --skip-wav-copy
 fi

examples/ljspeech/voc1/local/synthesize.sh

@@ -1,14 +0,0 @@
-#!/bin/bash
-
-config_path=$1
-train_output_path=$2
-ckpt_name=$3
-
-FLAGS_allocator_strategy=naive_best_fit \
-FLAGS_fraction_of_gpu_memory_to_use=0.01 \
-python3 ${BIN_DIR}/../synthesize.py \
-    --config=${config_path} \
-    --checkpoint=${train_output_path}/checkpoints/${ckpt_name} \
-    --test-metadata=dump/test/norm/metadata.jsonl \
-    --output-dir=${train_output_path}/test \
-    --generator-type=pwgan

examples/ljspeech/voc1/local/synthesize.sh

@@ -0,0 +1 @@
+../../../csmsc/voc1/local/synthesize.sh

examples/ljspeech/voc1/local/train.sh

@@ -1,13 +0,0 @@
-#!/bin/bash
-
-config_path=$1
-train_output_path=$2
-
-FLAGS_cudnn_exhaustive_search=true \
-FLAGS_conv_workspace_size_limit=4000 \
-python ${BIN_DIR}/train.py \
-    --train-metadata=dump/train/norm/metadata.jsonl \
-    --dev-metadata=dump/dev/norm/metadata.jsonl \
-    --config=${config_path} \
-    --output-dir=${train_output_path} \
-    --ngpu=1

examples/ljspeech/voc1/local/train.sh

@@ -0,0 +1 @@
+../../../csmsc/voc1/local/train.sh

examples/ljspeech/voc1/path.sh

@@ -1,13 +0,0 @@
-#!/bin/bash
-export MAIN_ROOT=`realpath ${PWD}/../../../`
-
-export PATH=${MAIN_ROOT}:${MAIN_ROOT}/utils:${PATH}
-export LC_ALL=C
-
-export PYTHONDONTWRITEBYTECODE=1
-# Use UTF-8 in Python to avoid UnicodeDecodeError when LC_ALL=C
-export PYTHONIOENCODING=UTF-8
-export PYTHONPATH=${MAIN_ROOT}:${PYTHONPATH}
-
-MODEL=parallelwave_gan
-export BIN_DIR=${MAIN_ROOT}/paddlespeech/t2s/exps/gan_vocoder/${MODEL}

examples/ljspeech/voc1/path.sh

@@ -0,0 +1 @@
+../../csmsc/voc1/path.sh

examples/ljspeech/voc5/local/preprocess.sh

@@ -1,55 +0,0 @@
-#!/bin/bash
-
-stage=0
-stop_stage=100
-
-config_path=$1
-
-if [ ${stage} -le 0 ] && [ ${stop_stage} -ge 0 ]; then
-    # get durations from MFA's result
-    echo "Generate durations.txt from MFA results ..."
-    python3 ${MAIN_ROOT}/utils/gen_duration_from_textgrid.py \
-        --inputdir=./ljspeech_alignment \
-        --output=durations.txt \
-        --config=${config_path}
-fi
-
-if [ ${stage} -le 1 ] && [ ${stop_stage} -ge 1 ]; then
-    # extract features
-    echo "Extract features ..."
-    python3 ${BIN_DIR}/../preprocess.py \
-        --rootdir=~/datasets/LJSpeech-1.1/ \
-        --dataset=ljspeech \
-        --dumpdir=dump \
-        --dur-file=durations.txt \
-        --config=${config_path} \
-        --cut-sil=True \
-        --num-cpu=20
-fi
-
-if [ ${stage} -le 2 ] && [ ${stop_stage} -ge 2 ]; then
-    # get features' stats(mean and std)
-    echo "Get features' stats ..."
-    python3 ${MAIN_ROOT}/utils/compute_statistics.py \
-        --metadata=dump/train/raw/metadata.jsonl \
-        --field-name="feats"
-fi
-
-if [ ${stage} -le 3 ] && [ ${stop_stage} -ge 3 ]; then
-    # normalize, dev and test should use train's stats
-    echo "Normalize ..."
-   
-    python3 ${BIN_DIR}/../normalize.py \
-        --metadata=dump/train/raw/metadata.jsonl \
-        --dumpdir=dump/train/norm \
-        --stats=dump/train/feats_stats.npy
-    python3 ${BIN_DIR}/../normalize.py \
-        --metadata=dump/dev/raw/metadata.jsonl \
-        --dumpdir=dump/dev/norm \
-        --stats=dump/train/feats_stats.npy
-    
-    python3 ${BIN_DIR}/../normalize.py \
-        --metadata=dump/test/raw/metadata.jsonl \
-        --dumpdir=dump/test/norm \
-        --stats=dump/train/feats_stats.npy
-fi

examples/ljspeech/voc5/local/preprocess.sh

@@ -0,0 +1 @@
+../../voc1/local/preprocess.sh

examples/ljspeech/voc5/local/synthesize.sh

@@ -1,14 +0,0 @@
-#!/bin/bash
-
-config_path=$1
-train_output_path=$2
-ckpt_name=$3
-
-FLAGS_allocator_strategy=naive_best_fit \
-FLAGS_fraction_of_gpu_memory_to_use=0.01 \
-python3 ${BIN_DIR}/../synthesize.py \
-    --config=${config_path} \
-    --checkpoint=${train_output_path}/checkpoints/${ckpt_name} \
-    --test-metadata=dump/test/norm/metadata.jsonl \
-    --output-dir=${train_output_path}/test \
-    --generator-type=hifigan

examples/ljspeech/voc5/local/synthesize.sh

@@ -0,0 +1 @@
+../../../csmsc/voc5/local/synthesize.sh

examples/ljspeech/voc5/local/train.sh

@@ -1,13 +0,0 @@
-#!/bin/bash
-
-config_path=$1
-train_output_path=$2
-
-FLAGS_cudnn_exhaustive_search=true \
-FLAGS_conv_workspace_size_limit=4000 \
-python ${BIN_DIR}/train.py \
-    --train-metadata=dump/train/norm/metadata.jsonl \
-    --dev-metadata=dump/dev/norm/metadata.jsonl \
-    --config=${config_path} \
-    --output-dir=${train_output_path} \
-    --ngpu=1

examples/ljspeech/voc5/local/train.sh

@@ -0,0 +1 @@
+../../../csmsc/voc1/local/train.sh

examples/ljspeech/voc5/path.sh

@@ -1,13 +0,0 @@
-#!/bin/bash
-export MAIN_ROOT=`realpath ${PWD}/../../../`
-
-export PATH=${MAIN_ROOT}:${MAIN_ROOT}/utils:${PATH}
-export LC_ALL=C
-
-export PYTHONDONTWRITEBYTECODE=1
-# Use UTF-8 in Python to avoid UnicodeDecodeError when LC_ALL=C
-export PYTHONIOENCODING=UTF-8
-export PYTHONPATH=${MAIN_ROOT}:${PYTHONPATH}
-
-MODEL=hifigan
-export BIN_DIR=${MAIN_ROOT}/paddlespeech/t2s/exps/gan_vocoder/${MODEL}

examples/ljspeech/voc5/path.sh

@@ -0,0 +1 @@
+../../csmsc/voc5/path.sh

examples/vctk/ernie_sat/local/train.sh

@@ -1,12 +0,0 @@
-#!/bin/bash
-
-config_path=$1
-train_output_path=$2
-
-python3 ${BIN_DIR}/train.py \
-    --train-metadata=dump/train/norm/metadata.jsonl \
-    --dev-metadata=dump/dev/norm/metadata.jsonl \
-    --config=${config_path} \
-    --output-dir=${train_output_path} \
-    --ngpu=8 \
-    --phones-dict=dump/phone_id_map.txt

examples/vctk/ernie_sat/local/train.sh

@@ -0,0 +1 @@
+../../../aishell3/ernie_sat/local/train.sh

examples/vctk/ernie_sat/path.sh

@@ -1,13 +0,0 @@
-#!/bin/bash
-export MAIN_ROOT=`realpath ${PWD}/../../../`
-
-export PATH=${MAIN_ROOT}:${MAIN_ROOT}/utils:${PATH}
-export LC_ALL=C
-
-export PYTHONDONTWRITEBYTECODE=1
-# Use UTF-8 in Python to avoid UnicodeDecodeError when LC_ALL=C
-export PYTHONIOENCODING=UTF-8
-export PYTHONPATH=${MAIN_ROOT}:${PYTHONPATH}
-
-MODEL=ernie_sat
-export BIN_DIR=${MAIN_ROOT}/paddlespeech/t2s/exps/${MODEL}

examples/vctk/ernie_sat/path.sh

@@ -0,0 +1 @@
+../../aishell3/ernie_sat/path.sh

examples/vctk/tts3/local/train.sh

@@ -1,13 +0,0 @@
-#!/bin/bash
-
-config_path=$1
-train_output_path=$2
-
-python3 ${BIN_DIR}/train.py \
-    --train-metadata=dump/train/norm/metadata.jsonl \
-    --dev-metadata=dump/dev/norm/metadata.jsonl \
-    --config=${config_path} \
-    --output-dir=${train_output_path} \
-    --ngpu=1 \
-    --phones-dict=dump/phone_id_map.txt \
-    --speaker-dict=dump/speaker_id_map.txt

examples/vctk/tts3/local/train.sh

@@ -0,0 +1 @@
+../../../aishell3/tts3/local/train.sh

examples/vctk/tts3/path.sh

@@ -1,13 +0,0 @@
-#!/bin/bash
-export MAIN_ROOT=`realpath ${PWD}/../../../`
-
-export PATH=${MAIN_ROOT}:${MAIN_ROOT}/utils:${PATH}
-export LC_ALL=C
-
-export PYTHONDONTWRITEBYTECODE=1
-# Use UTF-8 in Python to avoid UnicodeDecodeError when LC_ALL=C
-export PYTHONIOENCODING=UTF-8
-export PYTHONPATH=${MAIN_ROOT}:${PYTHONPATH}
-
-MODEL=fastspeech2
-export BIN_DIR=${MAIN_ROOT}/paddlespeech/t2s/exps/${MODEL}

examples/vctk/tts3/path.sh

@@ -0,0 +1 @@
+../../csmsc/tts3/path.sh

examples/vctk/voc1/local/preprocess.sh

@@ -42,14 +42,18 @@ if [ ${stage} -le 3 ] && [ ${stop_stage} -ge 3 ]; then
     python3 ${BIN_DIR}/../normalize.py \
         --metadata=dump/train/raw/metadata.jsonl \
         --dumpdir=dump/train/norm \
-        --stats=dump/train/feats_stats.npy
+        --stats=dump/train/feats_stats.npy \
+        --skip-wav-copy
+
     python3 ${BIN_DIR}/../normalize.py \
         --metadata=dump/dev/raw/metadata.jsonl \
         --dumpdir=dump/dev/norm \
-        --stats=dump/train/feats_stats.npy
+        --stats=dump/train/feats_stats.npy \
+        --skip-wav-copy
     
     python3 ${BIN_DIR}/../normalize.py \
         --metadata=dump/test/raw/metadata.jsonl \
         --dumpdir=dump/test/norm \
-        --stats=dump/train/feats_stats.npy
+        --stats=dump/train/feats_stats.npy \
+        --skip-wav-copy
 fi

examples/vctk/voc1/local/synthesize.sh

@@ -1,14 +0,0 @@
-#!/bin/bash
-
-config_path=$1
-train_output_path=$2
-ckpt_name=$3
-
-FLAGS_allocator_strategy=naive_best_fit \
-FLAGS_fraction_of_gpu_memory_to_use=0.01 \
-python3 ${BIN_DIR}/../synthesize.py \
-    --config=${config_path} \
-    --checkpoint=${train_output_path}/checkpoints/${ckpt_name} \
-    --test-metadata=dump/test/norm/metadata.jsonl \
-    --output-dir=${train_output_path}/test \
-    --generator-type=pwgan

examples/vctk/voc1/local/synthesize.sh

@@ -0,0 +1 @@
+../../../csmsc/voc1/local/synthesize.sh

examples/vctk/voc1/local/train.sh

@@ -1,13 +0,0 @@
-#!/bin/bash
-
-config_path=$1
-train_output_path=$2
-
-FLAGS_cudnn_exhaustive_search=true \
-FLAGS_conv_workspace_size_limit=4000 \
-python ${BIN_DIR}/train.py \
-    --train-metadata=dump/train/norm/metadata.jsonl \
-    --dev-metadata=dump/dev/norm/metadata.jsonl \
-    --config=${config_path} \
-    --output-dir=${train_output_path} \
-    --ngpu=1

examples/vctk/voc1/local/train.sh

@@ -0,0 +1 @@
+../../../csmsc/voc1/local/train.sh

examples/vctk/voc1/path.sh

@@ -1,13 +0,0 @@
-#!/bin/bash
-export MAIN_ROOT=`realpath ${PWD}/../../../`
-
-export PATH=${MAIN_ROOT}:${MAIN_ROOT}/utils:${PATH}
-export LC_ALL=C
-
-export PYTHONDONTWRITEBYTECODE=1
-# Use UTF-8 in Python to avoid UnicodeDecodeError when LC_ALL=C
-export PYTHONIOENCODING=UTF-8
-export PYTHONPATH=${MAIN_ROOT}:${PYTHONPATH}
-
-MODEL=parallelwave_gan
-export BIN_DIR=${MAIN_ROOT}/paddlespeech/t2s/exps/gan_vocoder/${MODEL}

examples/vctk/voc1/path.sh

@@ -0,0 +1 @@
+../../csmsc/voc1/path.sh

examples/vctk/voc5/local/preprocess.sh

@@ -1,55 +0,0 @@
-#!/bin/bash
-
-stage=0
-stop_stage=100
-
-config_path=$1
-
-if [ ${stage} -le 0 ] && [ ${stop_stage} -ge 0 ]; then
-    # get durations from MFA's result
-    echo "Generate durations.txt from MFA results ..."
-    python3 ${MAIN_ROOT}/utils/gen_duration_from_textgrid.py \
-        --inputdir=./vctk_alignment \
-        --output=durations.txt \
-        --config=${config_path}
-fi
-
-if [ ${stage} -le 1 ] && [ ${stop_stage} -ge 1 ]; then
-    # extract features
-    echo "Extract features ..."
-    python3 ${BIN_DIR}/../preprocess.py \
-        --rootdir=~/datasets/VCTK-Corpus-0.92/ \
-        --dataset=vctk \
-        --dumpdir=dump \
-        --dur-file=durations.txt \
-        --config=${config_path} \
-        --cut-sil=True \
-        --num-cpu=20
-fi
-
-if [ ${stage} -le 2 ] && [ ${stop_stage} -ge 2 ]; then
-    # get features' stats(mean and std)
-    echo "Get features' stats ..."
-    python3 ${MAIN_ROOT}/utils/compute_statistics.py \
-        --metadata=dump/train/raw/metadata.jsonl \
-        --field-name="feats"
-fi
-
-if [ ${stage} -le 3 ] && [ ${stop_stage} -ge 3 ]; then
-    # normalize, dev and test should use train's stats
-    echo "Normalize ..."
-   
-    python3 ${BIN_DIR}/../normalize.py \
-        --metadata=dump/train/raw/metadata.jsonl \
-        --dumpdir=dump/train/norm \
-        --stats=dump/train/feats_stats.npy
-    python3 ${BIN_DIR}/../normalize.py \
-        --metadata=dump/dev/raw/metadata.jsonl \
-        --dumpdir=dump/dev/norm \
-        --stats=dump/train/feats_stats.npy
-    
-    python3 ${BIN_DIR}/../normalize.py \
-        --metadata=dump/test/raw/metadata.jsonl \
-        --dumpdir=dump/test/norm \
-        --stats=dump/train/feats_stats.npy
-fi

examples/vctk/voc5/local/preprocess.sh

@@ -0,0 +1 @@
+../../voc1/local/preprocess.sh

examples/vctk/voc5/local/synthesize.sh

@@ -1,14 +0,0 @@
-#!/bin/bash
-
-config_path=$1
-train_output_path=$2
-ckpt_name=$3
-
-FLAGS_allocator_strategy=naive_best_fit \
-FLAGS_fraction_of_gpu_memory_to_use=0.01 \
-python3 ${BIN_DIR}/../synthesize.py \
-    --config=${config_path} \
-    --checkpoint=${train_output_path}/checkpoints/${ckpt_name} \
-    --test-metadata=dump/test/norm/metadata.jsonl \
-    --output-dir=${train_output_path}/test \
-    --generator-type=hifigan

examples/vctk/voc5/local/synthesize.sh

@@ -0,0 +1 @@
+../../../csmsc/voc5/local/synthesize.sh

examples/vctk/voc5/local/train.sh

@@ -1,13 +0,0 @@
-#!/bin/bash
-
-config_path=$1
-train_output_path=$2
-
-FLAGS_cudnn_exhaustive_search=true \
-FLAGS_conv_workspace_size_limit=4000 \
-python ${BIN_DIR}/train.py \
-    --train-metadata=dump/train/norm/metadata.jsonl \
-    --dev-metadata=dump/dev/norm/metadata.jsonl \
-    --config=${config_path} \
-    --output-dir=${train_output_path} \
-    --ngpu=1

examples/vctk/voc5/local/train.sh

@@ -0,0 +1 @@
+../../../csmsc/voc1/local/train.sh

examples/vctk/voc5/path.sh

@@ -1,13 +0,0 @@
-#!/bin/bash
-export MAIN_ROOT=`realpath ${PWD}/../../../`
-
-export PATH=${MAIN_ROOT}:${MAIN_ROOT}/utils:${PATH}
-export LC_ALL=C
-
-export PYTHONDONTWRITEBYTECODE=1
-# Use UTF-8 in Python to avoid UnicodeDecodeError when LC_ALL=C
-export PYTHONIOENCODING=UTF-8
-export PYTHONPATH=${MAIN_ROOT}:${PYTHONPATH}
-
-MODEL=hifigan
-export BIN_DIR=${MAIN_ROOT}/paddlespeech/t2s/exps/gan_vocoder/${MODEL}

examples/vctk/voc5/path.sh

@@ -0,0 +1 @@
+../../csmsc/voc5/path.sh