updated

egs/commonvoice/ASR/local/compute_fbank_commonvoice_splits.py

@@ -17,7 +17,6 @@
 
 import argparse
 import logging
-from datetime import datetime
 from pathlib import Path
 
 import torch
@@ -30,6 +29,8 @@ from lhotse import (
     set_caching_enabled,
 )
 
+from icefall.utils import str2bool
+
 # Torch's multithreaded behavior needs to be disabled or
 # it wastes a lot of CPU and slow things down.
 # Do this outside of main() in case it needs to take effect
@@ -83,6 +84,13 @@ def get_args():
         help="Stop processing pieces until this number (exclusive).",
     )
 
+    parser.add_argument(
+        "--perturb-speed",
+        type=str2bool,
+        default=False,
+        help="""Perturb speed with factor 0.9 and 1.1 on train subset.""",
+    )
+
     return parser.parse_args()
 
 
@@ -130,6 +138,10 @@ def compute_fbank_commonvoice_splits(args):
             keep_overlapping=False, min_duration=None
         )
 
+        if args.perturb_speed:
+            logging.info(f"Doing speed perturb")
+            cut_set = cut_set + cut_set.perturb_speed(0.9) + cut_set.perturb_speed(1.1)
+
         logging.info("Computing features")
         cut_set = cut_set.compute_and_store_features_batch(
             extractor=extractor,

egs/commonvoice/ASR/prepare.sh

@@ -38,6 +38,7 @@ num_splits=1000
 dl_dir=$PWD/download
 release=cv-corpus-12.0-2022-12-07
 lang=fr
+perturb_speed=false
 
 . shared/parse_options.sh || exit 1
 
@@ -149,7 +150,8 @@ if [ $stage -le 6 ] && [ $stop_stage -ge 6 ]; then
       --batch-duration 200 \
       --start 0 \
       --num-splits $num_splits \
-      --language $lang
+      --language $lang \
+      --perturb-speed $perturb_speed
     touch data/${lang}/fbank/.cv-${lang}_train.done
   fi
 fi