From d273f7e12693e5b366cbf2ff7d01dde0264b01d9 Mon Sep 17 00:00:00 2001
From: 嘉渊 <wangjiaming.wjm@alibaba-inc.com>
Date: 星期五, 21 七月 2023 14:26:15 +0800
Subject: [PATCH] update

---
 egs/callhome/eend_ola/run.sh |   66 +++++++++++++++++++++++----------
 1 files changed, 46 insertions(+), 20 deletions(-)

diff --git a/egs/callhome/eend_ola/run.sh b/egs/callhome/eend_ola/run.sh
index fb030c5..11e1771 100644
--- a/egs/callhome/eend_ola/run.sh
+++ b/egs/callhome/eend_ola/run.sh
@@ -8,6 +8,11 @@
 count=1
 
 # general configuration
+dump_cmd=utils/run.pl
+nj=64
+
+# feature configuration
+data_dir="./data"
 simu_feats_dir="/nfs/wangjiaming.wjm/EEND_ARK_DATA/dump/simu_data/data"
 simu_feats_dir_chunk2000="/nfs/wangjiaming.wjm/EEND_ARK_DATA/dump/simu_data_chunk2000/data"
 callhome_feats_dir_chunk2000="/nfs/wangjiaming.wjm/EEND_ARK_DATA/dump/callhome_chunk2000/data"
@@ -31,7 +36,7 @@
 stop_stage=-1
 
 # exp tag
-tag="exp_fix"
+tag="exp1"
 
 . local/parse_options.sh || exit 1;
 
@@ -52,23 +57,40 @@
 
 # simulate mixture data for training and inference
 if [ ${stage} -le -1 ] && [ ${stop_stage} -ge -1 ]; then
-    echo "stage 0: Simulate mixture data for training and inference"
+    echo "stage -1: Simulate mixture data for training and inference"
     echo "The detail can be found in https://github.com/hitachi-speech/EEND"
     echo "Before running this step, you should download and compile kaldi and set KALDI_ROOT in this script and path.sh"
     echo "This stage may take a long time, please waiting..."
     KALDI_ROOT=
     ln -s $KALDI_ROOT/egs/wsj/s5/steps steps
     ln -s $KALDI_ROOT/egs/wsj/s5/utils utils
-    . local/run_prepare_shared_eda.sh
+    local/run_prepare_shared_eda.sh
 fi
 
-## Prepare data for training and inference
-#if [ ${stage} -le 0 ] && [ ${stop_stage} -ge 0 ]; then
-#    echo "stage 0: Prepare data for training and inference"
-#    echo "The detail can be found in https://github.com/hitachi-speech/EEND"
-#    . ./local/
-#fi
-#
+# Prepare data for training and inference
+if [ ${stage} -le 0 ] && [ ${stop_stage} -ge 0 ]; then
+    echo "stage 0: Prepare data for training and inference"
+    simu_opts_num_speaker_array=(1 2 3 4)
+    simu_opts_sil_scale_array=(2 2 5 9)
+    simu_opts_num_train=100000
+
+    # for simulated data of chunk500
+    for dset in swb_sre_tr swb_sre_cv; do
+        if [ "$dset" == "swb_sre_tr" ]; then
+            n_mixtures=${simu_opts_num_train}
+        else
+            n_mixtures=500
+        fi
+        simu_data_dir=${dset}_ns"$(IFS="n"; echo "${simu_opts_num_speaker_array[*]}")"_beta"$(IFS="n"; echo "${simu_opts_sil_scale_array[*]}")"_${n_mixtures}
+        mkdir -p ${data_dir}/simu/data/${simu_data_dir}/.work
+        split_scps=
+        for n in $(seq $nj); do
+            split_scps="$split_scps ${data_dir}/simu/data/${simu_data_dir}/.work/wav.$n.scp"
+        done
+        utils/split_scp.pl "${data_dir}/simu/data/${simu_data_dir}/wav.scp" $split_scps || exit 1
+        python local/split.py ${data_dir}/simu/data/${simu_data_dir}
+    done
+fi
 
 # Training on simulated two-speaker data
 world_size=$gpu_num
@@ -245,13 +267,17 @@
     python local/model_averaging.py ${exp_dir}/exp/${callhome_model_dir}/$callhome_ave_id.pb $models
 fi
 
-## inference
-#if [ ${stage} -le 5 ] && [ ${stop_stage} -ge 5 ]; then
-#    echo "Inference"
-#    mkdir -p ${exp_dir}/exp/${callhome_model_dir}/inference/log
-#    CUDA_VISIBLE_DEVICES=$CUDA_VISIBLE_DEVICES python local/infer.py \
-#        --config_file ${exp_dir}/exp/${callhome_model_dir}/config.yaml \
-#        --model_file ${exp_dir}/exp/${callhome_model_dir}/$callhome_ave_id.pb \
-#        --output_rttm_file ${exp_dir}/exp/${callhome_model_dir}/inference/rttm \
-#        --wav_scp_file ${callhome_feats_dir_chunk2000}/${callhome_valid_dataset}/${callhome2_wav_scp_file} 1> ${exp_dir}/exp/${callhome_model_dir}/inference/log/infer.log 2>&1
-#fi
\ No newline at end of file
+# inference and compute DER
+if [ ${stage} -le 5 ] && [ ${stop_stage} -ge 5 ]; then
+    echo "Inference"
+    mkdir -p ${exp_dir}/exp/${callhome_model_dir}/inference/log
+    CUDA_VISIBLE_DEVICES=$CUDA_VISIBLE_DEVICES python local/infer.py \
+        --config_file ${exp_dir}/exp/${callhome_model_dir}/config.yaml \
+        --model_file ${exp_dir}/exp/${callhome_model_dir}/$callhome_ave_id.pb \
+        --output_rttm_file ${exp_dir}/exp/${callhome_model_dir}/inference/rttm \
+        --wav_scp_file ${callhome_feats_dir_chunk2000}/${callhome_valid_dataset}/${callhome2_wav_scp_file} \
+        1> ${exp_dir}/exp/${callhome_model_dir}/inference/log/infer.log 2>&1
+    md-eval.pl -c 0.25 \
+          -r ${callhome_feats_dir_chunk2000}/${callhome_valid_dataset}/rttm \
+          -s ${exp_dir}/exp/${callhome_model_dir}/inference/rttm > ${exp_dir}/exp/${callhome_model_dir}/inference/result_med11_collar0.25 2>/dev/null || exit
+fi
\ No newline at end of file

--
Gitblit v1.9.1