update

examples/mpnet_aishell/run.sh

@@ -109,7 +109,7 @@ if [ ${stage} -le 2 ] && [ ${stop_stage} -ge 2 ]; then
   $verbose && echo "stage 2: train model"
   cd "${work_dir}" || exit 1
   python3 step_2_train_model.py \
-  --train_dataset "${valid_dataset}" \
+  --train_dataset "${train_dataset}" \
   --valid_dataset "${valid_dataset}" \
   --serialization_dir "${file_dir}" \
   --config_file "${config_file}" \

examples/mpnet_aishell/step_1_prepare_data.py

@@ -42,6 +42,8 @@ def get_args():
 
     parser.add_argument("--target_sample_rate", default=8000, type=int)
 
+    parser.add_argument("--scale", default=0.01, type=float)
+
     args = parser.parse_args()
     return args
 
@@ -99,6 +101,9 @@ def get_dataset(args):
     count = 0
     process_bar = tqdm(desc="build dataset excel")
     for noise, speech in zip(noise_generator, speech_generator):
+        flag = random.random()
+        if flag > args.scale:
+            continue
 
         noise_filename = noise["filename"]
         noise_raw_duration = noise["raw_duration"]