parent
ef5f0436f2
commit
b648f0c2d1
@ -1,37 +1,35 @@
|
|||||||
DATA_PATH=/pfs/dlnel/public/dataset/speech/libri
|
DATA_PATH=$1
|
||||||
|
MODEL_PATH=$2
|
||||||
#setted by user
|
#setted by user
|
||||||
TRAIN_MANI=${DATA_PATH}/manifest_pcloud.train
|
TRAIN_MANI=${DATA_PATH}/cloud.train.manifest
|
||||||
#setted by user
|
#setted by user
|
||||||
DEV_MANI=${DATA_PATH}/manifest_pcloud.dev
|
DEV_MANI=${DATA_PATH}/cloud.test.manifest
|
||||||
#setted by user
|
#setted by user
|
||||||
TRAIN_TAR=${DATA_PATH}/data.train.tar
|
TRAIN_TAR=${DATA_PATH}/cloud.train.tar
|
||||||
#setted by user
|
#setted by user
|
||||||
DEV_TAR=${DATA_PATH}/data.dev.tar
|
DEV_TAR=${DATA_PATH}/cloud.test.tar
|
||||||
#setted by user
|
#setted by user
|
||||||
VOCAB_PATH=${DATA_PATH}/eng_vocab.txt
|
VOCAB_PATH=${DATA_PATH}/eng_vocab.txt
|
||||||
#setted by user
|
#setted by user
|
||||||
MEAN_STD_FILE=${DATA_PATH}/mean_std.npz
|
MEAN_STD_FILE=${DATA_PATH}/mean_std.npz
|
||||||
|
|
||||||
tar -xzf deepspeech.tar.gz
|
|
||||||
rm -rf ./cloud/data/*
|
|
||||||
|
|
||||||
# split train data for each pcloud node
|
# split train data for each pcloud node
|
||||||
python ./cloud/pcloud_split_data.py \
|
python ./cloud/split_data.py \
|
||||||
--in_manifest_path=$TRAIN_MANI \
|
--in_manifest_path=$TRAIN_MANI \
|
||||||
--data_tar_path=$TRAIN_TAR \
|
--data_tar_path=$TRAIN_TAR \
|
||||||
--out_manifest_path='./cloud/data/train.mani'
|
--out_manifest_path='./local.train.manifest'
|
||||||
|
|
||||||
# split dev data for each pcloud node
|
# split dev data for each pcloud node
|
||||||
python pcloud_split_data.py \
|
python ./cloud/split_data.py \
|
||||||
--in_manifest_path=$DEV_MANI \
|
--in_manifest_path=$DEV_MANI \
|
||||||
--data_tar_path=$DEV_TAR \
|
--data_tar_path=$DEV_TAR \
|
||||||
--out_manifest_path='./cloud/data/dev.mani'
|
--out_manifest_path='./local.test.manifest'
|
||||||
|
|
||||||
python train.py \
|
python train.py \
|
||||||
--use_gpu=1 \
|
--use_gpu=1 \
|
||||||
--trainer_count=4 \
|
--trainer_count=4 \
|
||||||
--batch_size=256 \
|
--batch_size=256 \
|
||||||
--mean_std_filepath=$MEAN_STD_FILE \
|
--mean_std_filepath=$MEAN_STD_FILE \
|
||||||
--train_manifest_path='./cloud/data/train.mani' \
|
--train_manifest_path='./local.train.manifest' \
|
||||||
--dev_manifest_path='./cloud/data/dev.mani' \
|
--dev_manifest_path='./local.test.manifest' \
|
||||||
--vocab_filepath=$VOCAB_PATH \
|
--vocab_filepath=$VOCAB_PATH \
|
||||||
|
@ -1,37 +1,35 @@
|
|||||||
DATA_PATH=/pfs/dlnel/public/dataset/speech/libri
|
DATA_PATH=$1
|
||||||
|
MODEL_PATH=$2
|
||||||
#setted by user
|
#setted by user
|
||||||
TRAIN_MANI=${DATA_PATH}/manifest_pcloud.train
|
TRAIN_MANI=${DATA_PATH}/cloud.train.manifest
|
||||||
#setted by user
|
#setted by user
|
||||||
DEV_MANI=${DATA_PATH}/manifest_pcloud.dev
|
DEV_MANI=${DATA_PATH}/cloud.test.manifest
|
||||||
#setted by user
|
#setted by user
|
||||||
TRAIN_TAR=${DATA_PATH}/data.train.tar
|
TRAIN_TAR=${DATA_PATH}/cloud.train.tar
|
||||||
#setted by user
|
#setted by user
|
||||||
DEV_TAR=${DATA_PATH}/data.dev.tar
|
DEV_TAR=${DATA_PATH}/cloud.test.tar
|
||||||
#setted by user
|
#setted by user
|
||||||
VOCAB_PATH=${DATA_PATH}/eng_vocab.txt
|
VOCAB_PATH=${DATA_PATH}/eng_vocab.txt
|
||||||
#setted by user
|
#setted by user
|
||||||
MEAN_STD_FILE=${DATA_PATH}/mean_std.npz
|
MEAN_STD_FILE=${DATA_PATH}/mean_std.npz
|
||||||
|
|
||||||
tar -xzvf deepspeech.tar.gz
|
|
||||||
rm -rf ./cloud/data/*
|
|
||||||
|
|
||||||
# split train data for each pcloud node
|
# split train data for each pcloud node
|
||||||
python ./cloud/pcloud_split_data.py \
|
python ./cloud/split_data.py \
|
||||||
--in_manifest_path=$TRAIN_MANI \
|
--in_manifest_path=$TRAIN_MANI \
|
||||||
--data_tar_path=$TRAIN_TAR \
|
--data_tar_path=$TRAIN_TAR \
|
||||||
--out_manifest_path='./cloud/data/train.mani'
|
--out_manifest_path='./local.train.manifest'
|
||||||
|
|
||||||
# split dev data for each pcloud node
|
# split dev data for each pcloud node
|
||||||
python pcloud_split_data.py \
|
python ./cloud/split_data.py \
|
||||||
--in_manifest_path=$DEV_MANI \
|
--in_manifest_path=$DEV_MANI \
|
||||||
--data_tar_path=$DEV_TAR \
|
--data_tar_path=$DEV_TAR \
|
||||||
--out_manifest_path='./cloud/data/dev.mani'
|
--out_manifest_path='./local.test.manifest'
|
||||||
|
|
||||||
python train.py \
|
python train.py \
|
||||||
--use_gpu=1 \
|
--use_gpu=1 \
|
||||||
--trainer_count=4 \
|
--trainer_count=4 \
|
||||||
--batch_size=256 \
|
--batch_size=256 \
|
||||||
--mean_std_filepath=$MEAN_STD_FILE \
|
--mean_std_filepath=$MEAN_STD_FILE \
|
||||||
--train_manifest_path='./cloud/data/train.mani' \
|
--train_manifest_path='./local.train.manifest' \
|
||||||
--dev_manifest_path='./cloud/data/dev.mani' \
|
--dev_manifest_path='./local.test.manifest' \
|
||||||
--vocab_filepath=$VOCAB_PATH \
|
--vocab_filepath=$VOCAB_PATH \
|
||||||
|
Loading…
Reference in new issue