update the installing script for openfst

benchmark/README.md

@@ -1,14 +0,0 @@
-### Prepare the environment
-Please follow the instructions shown in [here](https://github.com/PaddlePaddle/DeepSpeech/blob/develop/docs/src/install.md) to install the Deepspeech first.
-
-### Prepare the benchmark environment
-bash prepare.sh
-
-### Start benchmarking
-bash run.sh
-
-### Analyse the sp
-bash run_analysis_sp.sh
-
-### Analyse the mp
-bash run_analysis_mp.sh

benchmark/run_benchmark.sh

@@ -1,65 +0,0 @@
-#!/usr/bin/env bash
-set -xe
-# 运行示例：CUDA_VISIBLE_DEVICES=0 bash run_benchmark.sh ${run_mode} ${bs_item} ${fp_item} 500 ${model_mode}
-# 参数说明
-function _set_params(){
-
-    run_mode=${1:-"sp"}          # 单卡sp|多卡mp
-    config_path=${2:-"conf/conformer.yaml"}
-    output=${3:-"exp/conformer"}
-    seed=${4:-"0"}
-    ngpu=${5:-"1"}
-    profiler_options=${6:-"None"}
-    batch_size=${7:-"32"}
-    fp_item=${8:-"fp32"}
-    TRAIN_LOG_DIR=${9:-$(pwd)}
-
-    benchmark_max_step=0
-
-    run_log_path=${TRAIN_LOG_DIR:-$(pwd)}  # TRAIN_LOG_DIR 后续QA设置该参数
-
-#   以下不用修改
-    device=${CUDA_VISIBLE_DEVICES//,/ }
-    arr=(${device})
-    num_gpu_devices=${#arr[*]}
-    log_file=${run_log_path}/recoder_${run_mode}_bs${batch_size}_${fp_item}_ngpu${ngpu}.txt
-}
-
-function _train(){
-    echo "Train on ${num_gpu_devices} GPUs"
-    echo "current CUDA_VISIBLE_DEVICES=$CUDA_VISIBLE_DEVICES, gpus=$num_gpu_devices, batch_size=$batch_size"
-    train_cmd="--config=${config_path}
-               --output=${output}
-               --seed=${seed}
-               --nproc=${ngpu}
-               --profiler-options "${profiler_options}"
-               --benchmark-batch-size ${batch_size}
-               --benchmark-max-step ${benchmark_max_step} "
-
-    echo "run_mode "${run_mode}
-
-    case ${run_mode} in
-    sp) train_cmd="python3 -u ${BIN_DIR}/train.py "${train_cmd} ;;
-    mp) train_cmd="python3 -u ${BIN_DIR}/train.py "${train_cmd} ;;
-    *) echo "choose run_mode(sp or mp)"; exit 1;
-    esac
-    echo ${train_cmd}
-# 以下不用修改
-    timeout 15m ${train_cmd} > ${log_file} 2>&1
-    if [ $? -ne 0 ];then
-        echo -e "${model_name}, FAIL"
-        export job_fail_flag=1
-    else
-        echo -e "${model_name}, SUCCESS"
-        export job_fail_flag=0
-    fi
-    kill -9 `ps -ef|grep 'python'|awk '{print $2}'`
-
-    if [ $run_mode = "mp" -a -d mylog ]; then
-        rm ${log_file}
-        cp mylog/workerlog.0 ${log_file}
-    fi
-}
-
-_set_params $@
-_train

deepspeech/decoders/swig/setup.sh

@@ -10,7 +10,7 @@ fi
 
 if [ ! -d openfst-1.6.3 ]; then
     echo "Download and extract openfst ..."
-    wget http://www.openfst.org/twiki/pub/FST/FstDownload/openfst-1.6.3.tar.gz
+    wget http://www.openfst.org/twiki/pub/FST/FstDownload/openfst-1.6.3.tar.gz --no-check-certificate
     tar -xzvf openfst-1.6.3.tar.gz
     echo -e "\n"
 fi

tests/benchmark/.gitignore

@@ -1,2 +0,0 @@
-old-pd_env.txt
-pd_env.txt

tests/benchmark/README.md

@@ -1,11 +1,22 @@
-# Benchmark Test
+### Prepare the environment
+Please follow the instructions shown in [here](https://github.com/PaddlePaddle/DeepSpeech/blob/develop/docs/src/install.md) to install the Deepspeech first.
 
-## Data
+### Prepare the benchmark environment
+```
+bash prepare.sh
+```
 
-* Aishell
+### Start benchmarking
+```
+bash run.sh
+```
 
-## Docker
+### Analyse the sp
+```
+bash run_analysis_sp.sh
+```
 
+### Analyse the mp
 ```
-registry.baidubce.com/paddlepaddle/paddle   2.1.1-gpu-cuda10.2-cudnn7   59d5ec1de486  
+bash run_analysis_mp.sh
 ```

tests/benchmark/prepare.sh

@@ -1,7 +1,7 @@
 source ../tools/venv/bin/activate
 
-#进入执行目录
+#Enter the example dir
 pushd ../examples/aishell/s1
 
-#准备数据
+#Prepare the data
 bash run.sh --stage 0 --stop_stage 0

tests/benchmark/run_all.sh

@@ -1,49 +0,0 @@
-#!/bin/bash
-
-CUR_DIR=${PWD}
-ROOT_DIR=../../
-
-# 提供可稳定复现性能的脚本，默认在标准docker环境内py37执行：
-# collect env info
-bash ${ROOT_DIR}/utils/pd_env_collect.sh
-#cat pd_env.txt
-
-
-# 1 安装该模型需要的依赖 (如需开启优化策略请注明)
-#pushd ${ROOT_DIR}/tools; make; popd
-#source ${ROOT_DIR}/tools/venv/bin/activate
-#pushd ${ROOT_DIR}; bash setup.sh; popd
-
-
-# 2 拷贝该模型需要数据、预训练模型
-
-# 执行目录：需说明
-#pushd ${ROOT_DIR}/examples/aishell/s1
-pushd ${ROOT_DIR}/examples/tiny/s1
-
-mkdir -p exp/log
-. path.sh
-#bash local/data.sh &> exp/log/data.log
-
-# 3 批量运行（如不方便批量，1，2需放到单个模型中）
-
-model_mode_list=(conformer transformer)
-fp_item_list=(fp32)
-bs_item_list=(32 64 96)
-for model_mode in ${model_mode_list[@]}; do
-      for fp_item in ${fp_item_list[@]}; do
-          for bs_item in ${bs_item_list[@]}
-            do
-            echo "index is speed, 1gpus, begin, ${model_name}"
-            run_mode=sp
-            CUDA_VISIBLE_DEVICES=0 bash ${CUR_DIR}/run_benchmark.sh ${run_mode} ${bs_item} ${fp_item} 500 ${model_mode}     #  (5min)
-            sleep 60
-            echo "index is speed, 8gpus, run_mode is multi_process, begin, ${model_name}"
-            run_mode=mp
-            CUDA_VISIBLE_DEVICES=0,1,2,3,4,5,6,7 bash ${CUR_DIR}/run_benchmark.sh ${run_mode} ${bs_item} ${fp_item} 500 ${model_mode}
-            sleep 60
-            done
-      done
-done
-
-popd # aishell/s1

tests/benchmark/run_benchmark.sh

@@ -1,39 +1,49 @@
-#!/bin/bash
+#!/usr/bin/env bash
-
 set -xe
-
 # 运行示例：CUDA_VISIBLE_DEVICES=0 bash run_benchmark.sh ${run_mode} ${bs_item} ${fp_item} 500 ${model_mode}
 # 参数说明
 function _set_params(){
+
     run_mode=${1:-"sp"}          # 单卡sp|多卡mp
-    batch_size=${2:-"64"}
+    config_path=${2:-"conf/conformer.yaml"}
-    fp_item=${3:-"fp32"}        # fp32|fp16
+    output=${3:-"exp/conformer"}
-    max_iter=${4:-"500"}       # 可选，如果需要修改代码提前中断
+    seed=${4:-"0"}
-    model_name=${5:-"model_name"}
+    ngpu=${5:-"1"}
+    profiler_options=${6:-"None"}
+    batch_size=${7:-"32"}
+    fp_item=${8:-"fp32"}
+    TRAIN_LOG_DIR=${9:-$(pwd)}
+
+    benchmark_max_step=0
+
     run_log_path=${TRAIN_LOG_DIR:-$(pwd)}  # TRAIN_LOG_DIR 后续QA设置该参数
 
 #   以下不用修改
     device=${CUDA_VISIBLE_DEVICES//,/ }
     arr=(${device})
     num_gpu_devices=${#arr[*]}
-    log_file=${run_log_path}/${model_name}_${run_mode}_bs${batch_size}_${fp_item}_${num_gpu_devices}
+    log_file=${run_log_path}/recoder_${run_mode}_bs${batch_size}_${fp_item}_ngpu${ngpu}.txt
 }
 
 function _train(){
     echo "Train on ${num_gpu_devices} GPUs"
     echo "current CUDA_VISIBLE_DEVICES=$CUDA_VISIBLE_DEVICES, gpus=$num_gpu_devices, batch_size=$batch_size"
+    train_cmd="--config=${config_path}
+               --output=${output}
+               --seed=${seed}
+               --nproc=${ngpu}
+               --profiler-options "${profiler_options}"
+               --benchmark-batch-size ${batch_size}
+               --benchmark-max-step ${benchmark_max_step} "
 
-    train_cmd="--benchmark-batch-size ${batch_size}
+    echo "run_mode "${run_mode}
-               --benchmark-max-step ${max_iter}
-               conf/${model_name}.yaml ${model_name}"
 
     case ${run_mode} in
-    sp) train_cmd="bash local/train.sh "${train_cmd}"" ;;
+    sp) train_cmd="python3 -u ${BIN_DIR}/train.py "${train_cmd} ;;
-    mp)
+    mp) train_cmd="python3 -u ${BIN_DIR}/train.py "${train_cmd} ;;
-        train_cmd="bash local/train.sh "${train_cmd}"" ;;
     *) echo "choose run_mode(sp or mp)"; exit 1;
     esac
-
+    echo ${train_cmd}
 # 以下不用修改
     timeout 15m ${train_cmd} > ${log_file} 2>&1
     if [ $? -ne 0 ];then
@@ -43,8 +53,7 @@ function _train(){
         echo -e "${model_name}, SUCCESS"
         export job_fail_flag=0
     fi
-
+    kill -9 `ps -ef|grep 'python'|awk '{print $2}'`
-    trap 'for pid in $(jobs -pr); do kill -KILL $pid; done' INT QUIT TERM
 
     if [ $run_mode = "mp" -a -d mylog ]; then
         rm ${log_file}
@@ -54,4 +63,3 @@ function _train(){
 
 _set_params $@
 _train
-