updated Implementation

egs/librispeech/ASR/chech_version.py

@@ -1,11 +0,0 @@
-import torch
-import torchaudio
-
-print("PyTorch version:", torch.__version__)
-# print("Torchvision version:", torchvision.__version__)
-print("Torchaudio version:", torchaudio.__version__)
-print("CUDA available:", torch.cuda.is_available())
-if torch.cuda.is_available():
-        print("CUDA version:", torch.version.cuda)
-        print("cuDNN version:", torch.backends.cudnn.version())
-        print("GPU name:", torch.cuda.get_device_name(0))

egs/librispeech/ASR/local/compute_fbank_librispeech.py

@@ -149,18 +149,18 @@ def compute_fbank_librispeech(
                         + cut_set.perturb_speed(0.9)
                         + cut_set.perturb_speed(1.1)
                     )
-            # 修改這部分代碼
+
             if ex is None:
-                # 為 None 的情況（本地執行）創建自定義進程池上下文
+                # Create a custom process pool context for None (local execution)
                 import multiprocessing as mp
                 from concurrent.futures import ProcessPoolExecutor
 
-                # 計算工作數
+                # Calculate the number of jobs
                 actual_jobs = (
                     min(num_jobs * 2, 20) if "train" in partition else num_jobs
                 )
 
-                # 使用 forkserver 方法
+                # Use the forkserver method
                 ctx = mp.get_context("forkserver")
                 with ProcessPoolExecutor(
                     max_workers=actual_jobs, mp_context=ctx
@@ -172,7 +172,7 @@ def compute_fbank_librispeech(
                         storage_type=LilcomChunkyWriter,
                     )
             else:
-                # 分佈式環境，使用提供的執行器
+                # Distributed environment, use the provided executor
                 cut_set = cut_set.compute_and_store_features(
                     extractor=extractor,
                     storage_path=f"{output_dir}/{prefix}_feats_{partition}",

egs/librispeech/ASR/prepare.sh

@@ -99,9 +99,9 @@ if [ $stage -le 0 ] && [ $stop_stage -ge 0 ]; then
   #
   #   ln -sfv /path/to/LibriSpeech $dl_dir/LibriSpeech
   #
-  #if [ ! -d $dl_dir/LibriSpeech/train-other-500 ]; then
-  #  lhotse download librispeech --full $dl_dir
-  #fi
+  if [ ! -d $dl_dir/LibriSpeech/train-other-500 ]; then
+    lhotse download librispeech --full $dl_dir
+  fi
 
   # If you have pre-downloaded it to /path/to/musan,
   # you can create a symlink
@@ -139,7 +139,7 @@ if [ $stage -le 3 ] && [ $stop_stage -ge 3 ]; then
   log "Stage 3: Compute fbank for librispeech"
   mkdir -p data/fbank
   if [ ! -e data/fbank/.librispeech.done ]; then
-    ./local/compute_fbank_librispeech.py --num-workers 4
+    ./local/compute_fbank_librispeech.py --num-workers $nj
     touch data/fbank/.librispeech.done
   fi