fix style

.flake8

@@ -15,7 +15,7 @@ per-file-ignores =
     egs/librispeech/ASR/zipformer_mmi/*.py: E501, E203
     egs/librispeech/ASR/zipformer/*.py: E501, E203
     egs/librispeech/ASR/RESULTS.md: E999,
-
+    egs/ljspeech/TTS/vits/*.py: E501, E203
     # invalid escape sequence (cause by tex formular), W605
     icefall/utils.py: E501, W605
 

docs/source/recipes/TTS/ljspeech/vits.rst

@@ -103,4 +103,10 @@ You can test the exported ONNX model with:
       --model-filename vits/exp/vits-epoch-1000.onnx \
       --tokens data/tokens.txt
 
+Download pretrained models
+--------------------------
 
+If you don't want to train from scratch, you can download the pretrained models
+by visiting the following link:
+
+  - `<https://huggingface.co/Zengwei/icefall-tts-ljspeech-vits-2023-11-29>`_

egs/ljspeech/TTS/local/compute_spectrogram_ljspeech.py

@@ -29,7 +29,13 @@ import os
 from pathlib import Path
 
 import torch
-from lhotse import CutSet, Spectrogram, SpectrogramConfig, LilcomChunkyWriter, load_manifest
+from lhotse import (
+    CutSet,
+    LilcomChunkyWriter,
+    Spectrogram,
+    SpectrogramConfig,
+    load_manifest,
+)
 from lhotse.audio import RecordingSet
 from lhotse.supervision import SupervisionSet
 

egs/ljspeech/TTS/local/prepare_token_file.py

@@ -73,7 +73,7 @@ def get_token2id(manifest_file: Path) -> Dict[str, int]:
     extra_tokens = [
         "<blk>",  # 0 for blank
         "<sos/eos>",  # 1 for sos and eos symbols.
-        "<unk>"  # 2 for OOV
+        "<unk>",  # 2 for OOV
     ]
     all_tokens = set()
 

egs/ljspeech/TTS/local/prepare_tokens_ljspeech.py

@@ -34,9 +34,7 @@ def prepare_tokens_ljspeech():
     suffix = "jsonl.gz"
     partition = "all"
 
-    cut_set = load_manifest(
-        output_dir / f"{prefix}_cuts_{partition}.{suffix}"
-    )
+    cut_set = load_manifest(output_dir / f"{prefix}_cuts_{partition}.{suffix}")
     g2p = g2p_en.G2p()
 
     new_cuts = []
@@ -51,9 +49,7 @@ def prepare_tokens_ljspeech():
         new_cuts.append(cut)
 
     new_cut_set = CutSet.from_cuts(new_cuts)
-    new_cut_set.to_file(
-        output_dir / f"{prefix}_cuts_with_tokens_{partition}.{suffix}"
-    )
+    new_cut_set.to_file(output_dir / f"{prefix}_cuts_with_tokens_{partition}.{suffix}")
 
 
 if __name__ == "__main__":

egs/ljspeech/TTS/vits/README.md

@@ -1 +1,3 @@
 See https://k2-fsa.github.io/icefall/recipes/TTS/ljspeech/vits.html for detailed tutorials.
+
+Training logs, Tensorboard logs, and checkpoints are uploaded to https://huggingface.co/Zengwei/icefall-tts-ljspeech-vits-2023-11-29.