Unable to use fairseq modules?

[mahdi155000]

Hello everyone,

I wanted to use Fairseq for the Persian language, but I'm encountering some issues. When I try to send a custom model in the web interface, I get the following error:

Error: cannot unpack non-iterable async_generator object

And when I try to run it in headless mode, I get this error:

./ebook2audiobook.sh --headless --ebook tts/a.txt --language fas --custom_model tts/fas/G_100000.pth --custom_config tts/fas/config.json --custom_vocab tts/fas/vocab.txt

Running in native mode
Error: Unrecognized option "--custom_config"

Could someone please help me? I believe I might be making a mistake somewhere.

Here is the link to the model I used for this task:

wget https://dl.fbaipublicfiles.com/mms/tts/fas.tar.gz

I then converted it to a zip file. I also tested it with other models, and I get the same result.

---

I have no issues with regular usage; I can use English, Arabic, and other XTTS models correctly, but I am having trouble with the Fairseq models.

[ROBERT-MCDOWELL]

you are using the old syntax which is not available anymore (see README)
to upload a custom model in headless mode:

./ebook2audiobook.sh --headless --ebook tts/a.txt --language fas --custom_model '/path/to/a/model.zip'

model.zip must contains only what the TTS engine requires so:
config.json, model.pth, and vocab.json for XTTSv2

for FAIRSEQ, the custom model upload is not available for now, I'm in discussion with coqui-tts guys to know if it's possible or their are improvements on the subject. I know some dev started but don't know the status of it.
however if you are really concerned to upload FAIRSEQ custom model so you can make some tests if som implementation done.
are you ok fo this?

[ROBERT-MCDOWELL]

btw the link https://dl.fbaipublicfiles.com/mms/tts/fas.tar.gz
is the standard fairseq fas model so you don't need to upload it as a custom model,
just choose persian language and fairseq tts engine!

[mahdi155000]

you are using the old syntax which is not available anymore (see README) to upload a custom model in headless mode:

./ebook2audiobook.sh --headless --ebook tts/a.txt --language fas --custom_model '/path/to/a/model.zip'

model.zip must contains only what the TTS engine requires so: config.json, model.pth, and vocab.json for XTTSv2

for FAIRSEQ, the custom model upload is not available for now, I'm in discussion with coqui-tts guys to know if it's possible or their are improvements on the subject. I know some dev started but don't know the status of it. however if you are really concerned to upload FAIRSEQ custom model so you can make some tests if som implementation done. are you ok fo this?

I am in no hurry, but if needed, I can help with the testing as long as my knowledge and the system allow it.
Also, since when I run it without uploading, the report tells me that the model was not found, I tried to upload it:

./ebook2audiobook.sh --headless --ebook tts/a.txt --language fas

File "path/ebook2audiobook-main/python_env/lib/python3.12/site-packages/TTS/utils/manage.py", line 468, in _find_files
    raise ValueError(" [!] Model file not found in the output path")
ValueError:  [!] Model file not found in the output path
Caught DependencyError:  [!] Model file not found in the output path
convert_chapters_to_audio 0.00%: 0/2 

[ROBERT-MCDOWELL]

please provide the full terminal log

[ROBERT-MCDOWELL]

try to use --tts_engin fairseq and see if it's better

[mahdi155000]

I apologize for the delay.
The option --tts_engin fairseq is giving me "unrecognized option"
I also tried with different formats, and the result was the same. Here is the output:

./ebook2audiobook.sh --headless --ebook tts/a.txt --language fas
Running in native mode
Available Processor Unit: cpu
path/ebook2audiobook/ebook2audiobook-main/python_env/lib/python3.12/site-packages/ebooklib/epub.py:1395: UserWarning: In the future version we will turn default option ignore_ncx to True.
  warnings.warn('In the future version we will turn default option ignore_ncx to True.')
1% Converting input to HTML...
InputFormatPlugin: TXT Input running
on path/ebook2audiobook/ebook2audiobook-main/tmp/ebook-5234e09d-65b0-4dc8-af2d-d5b7ae09b608/a.txt
Language not specified
Creator not specified
Building file list...
Normalizing filename cases
Rewriting HTML links
flow is too short, not running heuristics
Forcing index.html into XHTML namespace                                              
34% Running transforms on e-book...                                                  
Merging user specified metadata...
Detecting structure...
Auto generated TOC with 0 entries.
Flattening CSS and remapping font sizes...
Source base font size is 12.00000pt
Removing fake margins...
Cleaning up manifest...
Trimming unused files from manifest...
Creating EPUB Output...
67% Running EPUB Output plugin
Splitting markup on page breaks and flow limits, if any...
        Looking for large trees in index.html...
        No large trees found
Generating default cover
This EPUB file has no Table of Contents. Creating a default TOC
EPUB output written topath/ebook2audiobook/ebook2audiobook-main/tmp/ebook-5234e09d-65b0-4dc8-af2d-d5b7ae09b608/__a.epub
Output saved to   path/ebook2audiobook/ebook2audiobook-main/tmp/ebook-5234e09d-65b0-4dc8-af2d-d5b7ae09b608/__a.epub
/home/mahdi/github/ebook2audiobook/ebook2audiobook-main/python_env/lib/python3.12/site-packages/ebooklib/epub.py:1423: FutureWarning: This search incorrectly ignores the root element, and will be fixed in a future version.  If you rely on the current behaviour, change it to './/xmlns:rootfile[@media-type]'
  for root_file in tree.findall('//xmlns:rootfile[@media-type]', namespaces={'xmlns': NAMESPACES['CONTAINERNS']}):
Loading TTS tts_models/fas/fairseq/vits model from tts_models/fas/fairseq/vits...
convert_chapters_to_audio 0.00%: 0/2 
Chapter 1 containing 2 sentences...
unknown, a;
Character 'u' not found in the vocabulary. Discarding it.
unknown, a;
Character 'n' not found in the vocabulary. Discarding it.
unknown, a;
Character 'k' not found in the vocabulary. Discarding it.
unknown, a;
Character 'o' not found in the vocabulary. Discarding it.
unknown, a;
Character 'w' not found in the vocabulary. Discarding it.
unknown, a;
Character ',' not found in the vocabulary. Discarding it.
unknown, a;
Character 'a' not found in the vocabulary. Discarding it.
unknown, a;
Character ';' not found in the vocabulary. Discarding it.
[W129 07:06:15.822121283 NNPACK.cpp:61] Could not initialize NNPACK! Reason: Unsupported hardware.
Traceback (most recent call last):
  File "path/ebook2audiobook/ebook2audiobook-main/lib/functions.py", line 646, in convert_sentence_to_audio
    params['tts'].tts_with_vc_to_file(
  File "path/ebook2audiobook/ebook2audiobook-main/python_env/lib/python3.12/site-packages/TTS/api.py", line 508, in tts_with_vc_to_file
    wav = self.tts_with_vc(
          ^^^^^^^^^^^^^^^^^
  File "path/ebook2audiobook/ebook2audiobook-main/python_env/lib/python3.12/site-packages/TTS/api.py", line 468, in tts_with_vc
    self.load_vc_model_by_name("voice_conversion_models/multilingual/vctk/freevc24")
  File "path/ebook2audiobook/ebook2audiobook-main/python_env/lib/python3.12/site-packages/TTS/api.py", line 196, in load_vc_model_by_name
    model_path, config_path, vocoder_path, vocoder_config_path, model_dir = self.download_model_by_name(
                                                                            ^^^^^^^^^^^^^^^^^^^^^^^^^^^^
  File "path/ebook2audiobook/ebook2audiobook-main/python_env/lib/python3.12/site-packages/TTS/api.py", line 161, in download_model_by_name
    model_path, config_path, model_item = self.manager.download_model(model_name)
                                          ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
  File "path/ebook2audiobook/ebook2audiobook-main/python_env/lib/python3.12/site-packages/TTS/utils/manage.py", line 440, in download_model
    output_model_path, output_config_path = self._find_files(output_path)
                                            ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
  File "path/ebook2audiobook/ebook2audiobook-main/python_env/lib/python3.12/site-packages/TTS/utils/manage.py", line 468, in _find_files
    raise ValueError(" [!] Model file not found in the output path")
ValueError:  [!] Model file not found in the output path
Caught DependencyError:  [!] Model file not found in the output path
convert_chapters_to_audio 0.00%: 0/2 

[ROBERT-MCDOWELL]

try my fork and do the original command you made.

[mahdi155000]

try my fork and do the original command you made.
At the beginning it download the module but after that I get error "file_types".

/home/mahdi/github/robert-ebook2audiobook/ebook2audiobook/python_env/lib/python3.12/site-packages/gradio/components/audio.py:195: UserWarning: The `show_controls` parameter is deprecated and will be removed in a future release. Use `show_recording_waveform` instead.
  warnings.warn(
/home/mahdi/github/robert-ebook2audiobook/ebook2audiobook/python_env/lib/python3.12/site-packages/gradio/components/file.py:113: UserWarning: The `file_types` parameter is ignored when `file_count` is 'directory'.
  warnings.warn(
Processing eBook file: a.txt
Available Processor Unit: cuda
Running command: /usr/bin/ebook-convert /media/windows/driveD/linux/github/robert-ebook2audiobook/ebook2audiobook/tmp/ebook-92cde1b0-014a-4d10-9201-0302cf52b2ff/d368799032ca707be2e4d54fb864a49b/a.txt /media/windows/driveD/linux/github/robert-ebook2audiobook/ebook2audiobook/tmp/ebook-92cde1b0-014a-4d10-9201-0302cf52b2ff/d368799032ca707be2e4d54fb864a49b/__a.epub --input-encoding=utf-8 --output-profile=generic_eink --verbose
Conversion options changed from defaults:
  output_profile: 'generic_eink'
  verbose: 1
  input_encoding: 'utf-8'
1% Converting input to HTML...
InputFormatPlugin: TXT Input running
on /media/windows/driveD/linux/github/robert-ebook2audiobook/ebook2audiobook/tmp/ebook-92cde1b0-014a-4d10-9201-0302cf52b2ff/d368799032ca707be2e4d54fb864a49b/a.txt
Reading text from file...
Using user specified input encoding of utf-8
Auto detected paragraph type as block
Auto detected formatting as heuristic
Running text through basic conversion...
Language not specified
Creator not specified
Building file list...
        Found files...
                 HTMLFile:0:a:'/media/windows/driveD/linux/github/robert-ebook2audiobook/ebook2audiobook/tmp/calibre_7.24.0_tmp_ikma5opw/v2eo96w9_plumber/index.html'
Normalizing filename cases
Rewriting HTML links
Parsing index.html ...
*********  Heuristic processing HTML  *********
flow is too short, not running heuristics
Forcing index.html into XHTML namespace
34% Running transforms on e-book...
Merging user specified metadata...
Detecting structure...
Auto generated TOC with 0 entries.
Flattening CSS and remapping font sizes...
Source base font size is 12.00000pt
Removing fake margins...
Found 1 items of level: p_1
Ignoring level p_1
Cleaning up manifest...
Trimming unused files from manifest...
Creating EPUB Output...
67% Running EPUB Output plugin
Splitting markup on page breaks and flow limits, if any...
        Looking for large trees in index.html...
        No large trees found
Generating default cover
This EPUB file has no Table of Contents. Creating a default TOC
EPUB output written to /media/windows/driveD/linux/github/robert-ebook2audiobook/ebook2audiobook/tmp/ebook-92cde1b0-014a-4d10-9201-0302cf52b2ff/d368799032ca707be2e4d54fb864a49b/__a.epub
Output saved to   /media/windows/driveD/linux/github/robert-ebook2audiobook/ebook2audiobook/tmp/ebook-92cde1b0-014a-4d10-9201-0302cf52b2ff/d368799032ca707be2e4d54fb864a49b/__a.epub

/home/mahdi/github/robert-ebook2audiobook/ebook2audiobook/python_env/lib/python3.12/site-packages/ebooklib/epub.py:1423: FutureWarning: This search incorrectly ignores the root element, and will be fixed in a future version.  If you rely on the current behaviour, change it to './/xmlns:rootfile[@media-type]'
  for root_file in tree.findall('//xmlns:rootfile[@media-type]', namespaces={'xmlns': NAMESPACES['CONTAINERNS']}):
******* NOTE: YOU CAN SAFELY IGNORE "Character xx not found in the vocabulary." *******
Loading TTS tts_models/fas/fairseq/vits model from tts_models/fas/fairseq/vits...
convert_chapters_to_audio 0.00%: 0/1 
Chapter 1 containing 1 sentences...
Traceback (most recent call last):
  File "/media/windows/driveD/linux/github/robert-ebook2audiobook/ebook2audiobook/lib/functions.py", line 873, in convert_sentence_to_audio
    audio_data = params['tts'].tts_with_vc(
                 ^^^^^^^^^^^^^^^^^^^^^^^^^^
  File "/home/mahdi/github/robert-ebook2audiobook/ebook2audiobook/python_env/lib/python3.12/site-packages/TTS/api.py", line 468, in tts_with_vc
    self.load_vc_model_by_name("voice_conversion_models/multilingual/vctk/freevc24")
  File "/home/mahdi/github/robert-ebook2audiobook/ebook2audiobook/python_env/lib/python3.12/site-packages/TTS/api.py", line 199, in load_vc_model_by_name
    self.voice_converter = Synthesizer(
                           ^^^^^^^^^^^^
  File "/home/mahdi/github/robert-ebook2audiobook/ebook2audiobook/python_env/lib/python3.12/site-packages/TTS/utils/synthesizer.py", line 102, in __init__
    self._load_vc(self.vc_checkpoint, self.vc_config, use_cuda)
  File "/home/mahdi/github/robert-ebook2audiobook/ebook2audiobook/python_env/lib/python3.12/site-packages/TTS/utils/synthesizer.py", line 145, in _load_vc
    self.vc_model = setup_vc_model(config=self.vc_config)
                    ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
  File "/home/mahdi/github/robert-ebook2audiobook/ebook2audiobook/python_env/lib/python3.12/site-packages/TTS/vc/models/__init__.py", line 22, in setup_model
    return MyModel.init_from_config(config)
           ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
  File "/home/mahdi/github/robert-ebook2audiobook/ebook2audiobook/python_env/lib/python3.12/site-packages/TTS/vc/models/freevc.py", line 480, in init_from_config
    model = FreeVC(config)
            ^^^^^^^^^^^^^^
  File "/home/mahdi/github/robert-ebook2audiobook/ebook2audiobook/python_env/lib/python3.12/site-packages/TTS/vc/models/freevc.py", line 307, in __init__
    self.wavlm = get_wavlm()
                 ^^^^^^^^^^^
  File "/home/mahdi/github/robert-ebook2audiobook/ebook2audiobook/python_env/lib/python3.12/site-packages/TTS/vc/layers/freevc/wavlm/__init__.py", line 30, in get_wavlm
    checkpoint = torch.load(output_path, map_location=torch.device(device), weights_only=is_pytorch_at_least_2_4())
                 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
  File "/home/mahdi/github/robert-ebook2audiobook/ebook2audiobook/python_env/lib/python3.12/site-packages/torch/serialization.py", line 1432, in load
    with _open_zipfile_reader(opened_file) as opened_zipfile:
         ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
  File "/home/mahdi/github/robert-ebook2audiobook/ebook2audiobook/python_env/lib/python3.12/site-packages/torch/serialization.py", line 763, in __init__
    super().__init__(torch._C.PyTorchFileReader(name_or_buffer))
                     ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
RuntimeError: PytorchStreamReader failed reading zip archive: failed finding central directory
Caught DependencyError: PytorchStreamReader failed reading zip archive: failed finding central directory
convert_chapters_to_audio 0.00%: 0/1 
Traceback (most recent call last):
  File "/media/windows/driveD/linux/github/robert-ebook2audiobook/ebook2audiobook/lib/functions.py", line 873, in convert_sentence_to_audio
    audio_data = params['tts'].tts_with_vc(
                 ^^^^^^^^^^^^^^^^^^^^^^^^^^
  File "/home/mahdi/github/robert-ebook2audiobook/ebook2audiobook/python_env/lib/python3.12/site-packages/TTS/api.py", line 468, in tts_with_vc
    self.load_vc_model_by_name("voice_conversion_models/multilingual/vctk/freevc24")
  File "/home/mahdi/github/robert-ebook2audiobook/ebook2audiobook/python_env/lib/python3.12/site-packages/TTS/api.py", line 199, in load_vc_model_by_name
    self.voice_converter = Synthesizer(
                           ^^^^^^^^^^^^
  File "/home/mahdi/github/robert-ebook2audiobook/ebook2audiobook/python_env/lib/python3.12/site-packages/TTS/utils/synthesizer.py", line 102, in __init__
    self._load_vc(self.vc_checkpoint, self.vc_config, use_cuda)
  File "/home/mahdi/github/robert-ebook2audiobook/ebook2audiobook/python_env/lib/python3.12/site-packages/TTS/utils/synthesizer.py", line 145, in _load_vc
    self.vc_model = setup_vc_model(config=self.vc_config)
                    ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
  File "/home/mahdi/github/robert-ebook2audiobook/ebook2audiobook/python_env/lib/python3.12/site-packages/TTS/vc/models/__init__.py", line 22, in setup_model
    return MyModel.init_from_config(config)
           ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
  File "/home/mahdi/github/robert-ebook2audiobook/ebook2audiobook/python_env/lib/python3.12/site-packages/TTS/vc/models/freevc.py", line 480, in init_from_config
    model = FreeVC(config)
            ^^^^^^^^^^^^^^
  File "/home/mahdi/github/robert-ebook2audiobook/ebook2audiobook/python_env/lib/python3.12/site-packages/TTS/vc/models/freevc.py", line 307, in __init__
    self.wavlm = get_wavlm()
                 ^^^^^^^^^^^
  File "/home/mahdi/github/robert-ebook2audiobook/ebook2audiobook/python_env/lib/python3.12/site-packages/TTS/vc/layers/freevc/wavlm/__init__.py", line 30, in get_wavlm
    checkpoint = torch.load(output_path, map_location=torch.device(device), weights_only=is_pytorch_at_least_2_4())
                 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
  File "/home/mahdi/github/robert-ebook2audiobook/ebook2audiobook/python_env/lib/python3.12/site-packages/torch/serialization.py", line 1432, in load
    with _open_zipfile_reader(opened_file) as opened_zipfile:
         ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
  File "/home/mahdi/github/robert-ebook2audiobook/ebook2audiobook/python_env/lib/python3.12/site-packages/torch/serialization.py", line 763, in __init__
    super().__init__(torch._C.PyTorchFileReader(name_or_buffer))
                     ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
RuntimeError: PytorchStreamReader failed reading zip archive: failed finding central directory

During handling of the above exception, another exception occurred:

Traceback (most recent call last):
  File "/media/windows/driveD/linux/github/robert-ebook2audiobook/ebook2audiobook/lib/functions.py", line 808, in convert_chapters_to_audio
    if convert_sentence_to_audio(params, session):
    ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
  File "/media/windows/driveD/linux/github/robert-ebook2audiobook/ebook2audiobook/lib/functions.py", line 891, in convert_sentence_to_audio
    raise DependencyError(e)
lib.functions.DependencyError: PytorchStreamReader failed reading zip archive: failed finding central directory
Caught DependencyError: PytorchStreamReader failed reading zip archive: failed finding central directory
convert_ebook() Exception: PytorchStreamReader failed reading zip archive: failed finding central directory
Traceback (most recent call last):
  File "/home/mahdi/github/robert-ebook2audiobook/ebook2audiobook/python_env/lib/python3.12/site-packages/gradio/queueing.py", line 625, in process_events
    response = await route_utils.call_process_api(
               ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
  File "/home/mahdi/github/robert-ebook2audiobook/ebook2audiobook/python_env/lib/python3.12/site-packages/gradio/route_utils.py", line 322, in call_process_api
    output = await app.get_blocks().process_api(
             ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
  File "/home/mahdi/github/robert-ebook2audiobook/ebook2audiobook/python_env/lib/python3.12/site-packages/gradio/blocks.py", line 2044, in process_api
    result = await self.call_function(
             ^^^^^^^^^^^^^^^^^^^^^^^^^
  File "/home/mahdi/github/robert-ebook2audiobook/ebook2audiobook/python_env/lib/python3.12/site-packages/gradio/blocks.py", line 1591, in call_function
    prediction = await anyio.to_thread.run_sync(  # type: ignore
                 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
  File "/home/mahdi/github/robert-ebook2audiobook/ebook2audiobook/python_env/lib/python3.12/site-packages/anyio/to_thread.py", line 56, in run_sync
    return await get_async_backend().run_sync_in_worker_thread(
           ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
  File "/home/mahdi/github/robert-ebook2audiobook/ebook2audiobook/python_env/lib/python3.12/site-packages/anyio/_backends/_asyncio.py", line 2461, in run_sync_in_worker_thread
    return await future
           ^^^^^^^^^^^^
  File "/home/mahdi/github/robert-ebook2audiobook/ebook2audiobook/python_env/lib/python3.12/site-packages/anyio/_backends/_asyncio.py", line 962, in run
    result = context.run(func, *args)
             ^^^^^^^^^^^^^^^^^^^^^^^^
  File "/home/mahdi/github/robert-ebook2audiobook/ebook2audiobook/python_env/lib/python3.12/site-packages/gradio/utils.py", line 883, in wrapper
    response = f(*args, **kwargs)
               ^^^^^^^^^^^^^^^^^^
  File "/media/windows/driveD/linux/github/robert-ebook2audiobook/ebook2audiobook/lib/functions.py", line 1851, in refresh_interface
    audiobook_options, session['audiobook'] = update_audiobook_list(session['audiobooks_dir'])
                                              ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
  File "/media/windows/driveD/linux/github/robert-ebook2audiobook/ebook2audiobook/lib/functions.py", line 2169, in update_audiobook_list
    return list, default_value
                 ^^^^^^^^^^^^^
UnboundLocalError: cannot access local variable 'default_value' where it is not associated with a value
```
```

[ROBERT-MCDOWELL]

you are not in headless mode. retry
./ebook2audiobook.sh --headless --ebook tts/a.txt --language fas

[mahdi155000]

I'm sorry, I thought it was okay to use the web interface.

./ebook2audiobook.sh --headless --ebook tts/a.txt --language fas
Running in native mode
Traceback (most recent call last):
  File "/media/windows/driveD/linux/github/robert-ebook2audiobook/ebook2audiobook/app.py", line 262, in <module>
    main()
  File "/media/windows/driveD/linux/github/robert-ebook2audiobook/ebook2audiobook/app.py", line 224, in main
    if not os.path.exists(args['ebooks_dir']):
           ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
  File "<frozen genericpath>", line 19, in exists
TypeError: stat: path should be string, bytes, os.PathLike or integer, not NoneType

[ROBERT-MCDOWELL]

ok I think it's better you wait the coming release update....

[DrewThomasson]

This one's fixed now right?

[ROBERT-MCDOWELL]

yes, it's fixed since long time. but in fact he mixed with custom model and internal, so I guess he wanted to add the feature custom model for fairseq. reason why I added new feature.