Open30

uv環境でESPnetの推論スクリプトを実行

Python
Windows
ESPnet
uv
豚&紙箱豚&紙箱

tts.py

import soundfile
from espnet2.bin.tts_inference import Text2Speech
text2speech = Text2Speech.from_pretrained("kan-bayashi/ljspeech_vits")
text = "Hello, this is a text-to-speech test using ESPnet. Does my speech sound good?"
speech = text2speech(text)["wav"]
soundfile.write("output.wav", speech.numpy(), text2speech.fs, "PCM_16")
豚&紙箱豚&紙箱
uv run --with espnet --with soundfile tts.py




  x Failed to build `ctc-segmentation==1.7.4`
  |-> The build backend returned an error
  `-> Call to `setuptools.build_meta:__legacy__.build_wheel` failed (exit code: 1)

      [stdout]
      running bdist_wheel
      running build
      running build_py
      creating build\lib.win-amd64-cpython-313\ctc_segmentation
      copying ctc_segmentation\ctc_segmentation.py -> build\lib.win-amd64-cpython-313\ctc_segmentation
      copying ctc_segmentation\partitioning.py -> build\lib.win-amd64-cpython-313\ctc_segmentation
      copying ctc_segmentation\__init__.py -> build\lib.win-amd64-cpython-313\ctc_segmentation
      running build_ext
      building 'ctc_segmentation.ctc_segmentation_dyn' extension

      [stderr]
      C:\Users\zzxia\AppData\Local\uv\cache\builds-v0\.tmpodXSzV\Lib\site-packages\setuptools\_distutils\dist.py:289: UserWarning: Unknown
      distribution option: 'tests_require'
        warnings.warn(msg)
      error: Microsoft Visual C++ 14.0 or greater is required. Get it with "Microsoft C++ Build Tools":
      https://visualstudio.microsoft.com/visual-cpp-build-tools/

      hint: This usually indicates a problem with the package or the build environment.
  help: `ctc-segmentation` (v1.7.4) was included because `espnet` (v202412) depends on `ctc-segmentation`
豚&紙箱豚&紙箱

パッケージをインストールする前に、先にPythonをインストールする必要です。

豚&紙箱豚&紙箱

ESPnetとsoundfileをインストールします。

uv pip install espnet soundfile

同じエラーが出ました。

豚&紙箱豚&紙箱

Perplexity.aiに聞いてみました。

https://www.perplexity.ai/search/the-following-error-occurred-w-udqoI6HqTISOS_6KkZDW0A#0

Visual C++ 14.0以上が入っていなくてctc-segmentationがビルドできなかったようです。

error: Microsoft Visual C++ 14.0 or greater is required. Get it with "Microsoft C++ Build Tools": https://visualstudio.microsoft.com/visual-cpp-build-tools/

上記のリンクを踏んでインストールすればよさそうです。

豚&紙箱豚&紙箱

Visual Studio Build Tools 2022をインストールしました。

豚&紙箱豚&紙箱

もう一度espnetとsoundfileをインストールしてみます。
さっき失敗したところから再開しました。

豚&紙箱豚&紙箱

.venvが生成されました。

なんと1.5GBが取られました。

豚&紙箱豚&紙箱

uv pip installの時に以下の警告メッセージが出ました。

warning: Failed to hardlink files; falling back to full copy. This may lead to degraded performance.
         If the cache and target directories are on different filesystems, hardlinking may not be supported.
         If this is intentional, set `export UV_LINK_MODE=copy` or use `--link-mode=copy` to suppress this warning.

キャッシュフォルダと仮想環境フォルダが別々のファイルシステム（ディスク）にあるのは原因のようです。

https://github.com/astral-sh/uv/issues/7285#issuecomment-2343651950

uv cache dir

それぞれ同じコピーのパッケージがインストールされています。スペースがもったいないです。

uvのキャッシュについての公式資料は以下を参照
https://docs.astral.sh/uv/concepts/cache/#cache-directory

豚&紙箱豚&紙箱

espnet_model_zooも必要です。

flash_attnもインストールしたほうがよさそうです。

豚&紙箱豚&紙箱

flash_attnをインストールする時にtorchが見つからなくて怒られました。
でもespnetをインストールするときにtorchはインストールされたはずです。

豚&紙箱豚&紙箱

torchのバージョンが高すぎるかもしれません。

豚&紙箱豚&紙箱

flash_attnをなしに実行してみましたが、import Text2Speech2のところにエラーが出ました。

OSError: [WinError 127] The specified procedure could not be found

豚&紙箱豚&紙箱

別のエラーが出ました。nltk_dataが見つからなかったです。

PS D:\espnet-lab> uv run .\tts.py
Failed to import Flash Attention, using ESPnet default: No module named 'flash_attn'
https://zenodo.org/record/5443814/files/tts_train_vits_raw_phn_tacotron_g2p_en_no_space_train.total_count.ave.zip?download=1: 100%|█| 356M/356M [02:5
D:\espnet-lab\.venv\lib\site-packages\espnet_model_zoo\downloader.py:364: UserWarning: Not validating checksum
  warnings.warn("Not validating checksum")
D:\espnet-lab\.venv\lib\site-packages\torch\nn\utils\weight_norm.py:143: FutureWarning: `torch.nn.utils.weight_norm` is deprecated in favor of `torch.nn.utils.parametrizations.weight_norm`.
  WeightNorm.apply(module, name, dim)
D:\espnet-lab\.venv\lib\site-packages\espnet2\gan_tts\vits\monotonic_align\__init__.py:19: UserWarning: Cython version is not available. Fallback to 'EXPERIMETAL' numba version. If you want to use the cython version, please build it as follows: `cd espnet2/gan_tts/vits/monotonic_align; python setup.py build_ext --inplace`
  warnings.warn(
WARNING:root:It seems weight norm is not applied in the pretrained model but the current model uses it. To keep the compatibility, we remove the norm from the current model. This may cause unexpected behavior due to the parameter mismatch in finetuning. To avoid this issue, please change the following parameters in config to false:
 - discriminator_params.follow_official_norm
 - discriminator_params.scale_discriminator_params.use_weight_norm
 - discriminator_params.scale_discriminator_params.use_spectral_norm

See also:
 - https://github.com/espnet/espnet/pull/5240
 - https://github.com/espnet/espnet/pull/5249
Traceback (most recent call last):
  File "D:\espnet-lab\tts.py", line 5, in <module>
    speech = text2speech(text)["wav"]
  File "D:\espnet-lab\.venv\lib\site-packages\torch\utils\_contextlib.py", line 116, in decorate_context
    return func(*args, **kwargs)
  File "D:\espnet-lab\.venv\lib\site-packages\espnet2\bin\tts_inference.py", line 173, in __call__
    text = self.preprocess_fn("<dummy>", dict(text=text))["text"]
  File "D:\espnet-lab\.venv\lib\site-packages\espnet2\train\preprocessor.py", line 548, in __call__
    data = self._text_process(data)
  File "D:\espnet-lab\.venv\lib\site-packages\espnet2\train\preprocessor.py", line 483, in _text_process
    tokens = self.tokenizer.text2tokens(text)
  File "D:\espnet-lab\.venv\lib\site-packages\espnet2\text\phoneme_tokenizer.py", line 623, in text2tokens
    tokens = self.g2p(line)
  File "D:\espnet-lab\.venv\lib\site-packages\espnet2\text\phoneme_tokenizer.py", line 260, in __call__
    phones = self.g2p(text)
  File "D:\espnet-lab\.venv\lib\site-packages\g2p_en\g2p.py", line 162, in __call__
    tokens = pos_tag(words)  # tuples of (word, tag)
  File "D:\espnet-lab\.venv\lib\site-packages\nltk\tag\__init__.py", line 168, in pos_tag
    tagger = _get_tagger(lang)
  File "D:\espnet-lab\.venv\lib\site-packages\nltk\tag\__init__.py", line 110, in _get_tagger
    tagger = PerceptronTagger()
  File "D:\espnet-lab\.venv\lib\site-packages\nltk\tag\perceptron.py", line 183, in __init__
    self.load_from_json(lang)
  File "D:\espnet-lab\.venv\lib\site-packages\nltk\tag\perceptron.py", line 273, in load_from_json
    loc = find(f"taggers/averaged_perceptron_tagger_{lang}/")
  File "D:\espnet-lab\.venv\lib\site-packages\nltk\data.py", line 579, in find
    raise LookupError(resource_not_found)
LookupError:
**********************************************************************
  Resource averaged_perceptron_tagger_eng not found.
  Please use the NLTK Downloader to obtain the resource:

  >>> import nltk
  >>> nltk.download('averaged_perceptron_tagger_eng')

  For more information see: https://www.nltk.org/data.html

  Attempted to load taggers/averaged_perceptron_tagger_eng/

  Searched in:
    - 'C:\\Users\\zzxia/nltk_data'
    - 'D:\\espnet-lab\\.venv\\nltk_data'
    - 'D:\\espnet-lab\\.venv\\share\\nltk_data'
    - 'D:\\espnet-lab\\.venv\\lib\\nltk_data'
    - 'C:\\Users\\zzxia\\AppData\\Roaming\\nltk_data'
    - 'C:\\nltk_data'
    - 'D:\\nltk_data'
    - 'E:\\nltk_data'
**********************************************************************

PS D:\espnet-lab>
豚&紙箱豚&紙箱

D:\espnet-labフォルダに、以下のPythonスクリプトを作成し実行します。

import nltk
nltk.download('averaged_perceptron_tagger_eng')

またCドライブにダウンロードされました。

PS D:\espnet-lab> & D:/espnet-lab/.venv/Scripts/python.exe d:/espnet-lab/download_nltk_data.py
[nltk_data] Downloading package averaged_perceptron_tagger_eng to
[nltk_data]     C:\Users\zzxia\AppData\Roaming\nltk_data...
[nltk_data]   Unzipping taggers\averaged_perceptron_tagger_eng.zip.
PS D:\espnet-lab>

Dドライブの仮想環境に移動します。

mv C:\Users\zzxia\AppData\Roaming\nltk_data  .\.venv\

確認

PS D:\espnet-lab> ls .\.venv\ | select-string nltk

nltk_data


PS D:\espnet-lab>
豚&紙箱豚&紙箱

やっとtts.pyを実行できました。

PS D:\espnet-lab> uv run .\tts.py
Failed to import Flash Attention, using ESPnet default: No module named 'flash_attn'
D:\espnet-lab\.venv\lib\site-packages\torch\nn\utils\weight_norm.py:143: FutureWarning: `torch.nn.utils.weight_norm` is deprecated in favor of `torch.nn.utils.parametrizations.weight_norm`.
  WeightNorm.apply(module, name, dim)
D:\espnet-lab\.venv\lib\site-packages\espnet2\gan_tts\vits\monotonic_align\__init__.py:19: UserWarning: Cython version is not available. Fallback to 'EXPERIMETAL' numba version. If you want to use the cython version, please build it as follows: `cd espnet2/gan_tts/vits/monotonic_align; python setup.py build_ext --inplace`
  warnings.warn(
WARNING:root:It seems weight norm is not applied in the pretrained model but the current model uses it. To keep the compatibility, we remove the norm from the current model. This may cause unexpected behavior due to the parameter mismatch in finetuning. To avoid this issue, please change the following parameters in config to false:
 - discriminator_params.follow_official_norm
 - discriminator_params.scale_discriminator_params.use_weight_norm
 - discriminator_params.scale_discriminator_params.use_spectral_norm

See also:
 - https://github.com/espnet/espnet/pull/5240
 - https://github.com/espnet/espnet/pull/5249
PS D:\espnet-lab> ls


    Directory: D:\espnet-lab


Mode                 LastWriteTime         Length Name
----                 -------------         ------ ----
d-----         6/29/2025   4:06 PM                .venv
-a----         6/29/2025   4:03 PM             60 download_nltk_data.py
-a----         6/29/2025   4:08 PM         255532 output.wav
-a----         6/28/2025   7:36 PM            332 tts.py


PS D:\espnet-lab>
豚&紙箱豚&紙箱

uv pip installよりuv addがいいかもしれません。