a web interface with gradio?

[mikhail2013ru]

Can you make a web interface with gradio that allows us to adjust parameters such as speech speed, duration, warmth, etc.?

[wrightmikea]

I have forked this repo and added several scripts including some that provide a gradio UI (and some that provide a FastAPI REST server). fork. I've tested some of the scripts and will push more changes to the fork if I need to make changes after more testing.

[Jandown]

I have forked this repo and added several scripts including some that provide a gradio UI (and some that provide a FastAPI REST server). fork. I've tested some of the scripts and will push more changes to the fork if I need to make changes after more testing.

Why use your version to execute Python scripts/local-gradio.py
Afterwards, web page access http://127.0.0.1:7860 There is no content and no error messages

[mikhail2013ru]

I have forked this repo and added several scripts including some that provide a gradio UI (and some that provide a FastAPI REST server). fork. I've tested some of the scripts and will push more changes to the fork if I need to make changes after more testing.

Can you make a notebook to run your fork through colab with gradio? (Not only local use)

[mikhail2013ru]

I have forked this repo and added several scripts including some that provide a gradio UI (and some that provide a FastAPI REST server). fork. I've tested some of the scripts and will push more changes to the fork if I need to make changes after more testing.

Why use your version to execute Python scripts/local-gradio.py Afterwards, web page access http://127.0.0.1:7860 There is no content and no error messages

The approach is wrong, but the idea is good.

[Jandown]

I have forked this repo and added several scripts including some that provide a gradio UI (and some that provide a FastAPI REST server). fork. I've tested some of the scripts and will push more changes to the fork if I need to make changes after more testing.

Can you make a notebook to run your fork through colab with gradio? (Not only local use)

The problem has been resolved. Gradio version is too high and does not support Gradio5

[mikhail2013ru]

I have forked this repo and added several scripts including some that provide a gradio UI (and some that provide a FastAPI REST server). fork. I've tested some of the scripts and will push more changes to the fork if I need to make changes after more testing.

Can you make a notebook to run your fork through colab with gradio? (Not only local use)

The problem has been resolved. Gradio version is too high and does not support Gradio5

I don't understand if there is a way to run dmospeech2 in gradio on colab?

[wrightmikea]

I have forked this repo and added several scripts including some that provide a gradio UI (and some that provide a FastAPI REST server). fork. I've tested some of the scripts and will push more changes to the fork if I need to make changes after more testing.

Can you make a notebook to run your fork through colab with gradio? (Not only local use)

The problem has been resolved. Gradio version is too high and does not support Gradio5

I don't understand if there is a way to run dmospeech2 in gradio on colab?

I wil look info porting this to (i.e. testing this on) colab.

*** see comments below for resolution ***

[wrightmikea]

I attempted to port this to colab and after many iterations I got the gradio UI to show and work for the reference audio; however the generation step fails with "Error" and I won't be able to look into that until later today or tomorrow. In the meantime, someone more familiar with colab might make progress. gradio-GPU-colab-test notebook

the thing I've been trying to fix:
Caught ModuleNotFoundError in LMxv-ilhEFbU: No module named 'torchdiffeq' Sys.path at the time of error in LMxv-ilhEFbU: ['./src', './src', './src', './src', './src', './src', './src', '/content/DMOSpeech2/src', '/content', '/env/python', '/usr/lib/python311.zip', '/usr/lib/python3.11', '/usr/lib/python3.11/lib-dynload', '', '/usr/local/lib/python3.11/dist-packages', '/usr/lib/python3/dist-packages', '/usr/local/lib/python3.11/dist-packages/IPython/extensions', '/root/.ipython'] Working directory at the time of error: /content/DMOSpeech2 Skipping tts object creation due to import error. Colab notebook detected. To show errors in colab notebook, set debug=True in launch()

*** see my comment below for the final tested/working resolution***

[wrightmikea]

I've got a tested/working Google collab DMOSpeech2 (fork) gradio notebook working with a T4 GPU. latest shared notebook and I've updated my forked repo that has the added notebook. I had to run the script explicitly, as a workaround, because I could not get the last cell to work with inline python without using the wrong version. Maybe someone else can improve it.

[mikhail2013ru]

wrightmikea, good work. I launched dmospeech2 on colab an hour ago from your notepad, but I couldn't initialize the voice.

Traceback (most recent call last):
  File "/content/dmo_env/lib/python3.10/site-packages/gradio/queueing.py", line 626, in process_events
    response = await route_utils.call_process_api(
  File "/content/dmo_env/lib/python3.10/site-packages/gradio/route_utils.py", line 350, in call_process_api
    output = await app.get_blocks().process_api(
  File "/content/dmo_env/lib/python3.10/site-packages/gradio/blocks.py", line 2235, in process_api
    result = await self.call_function(
  File "/content/dmo_env/lib/python3.10/site-packages/gradio/blocks.py", line 1746, in call_function
    prediction = await anyio.to_thread.run_sync(  # type: ignore
  File "/content/dmo_env/lib/python3.10/site-packages/anyio/to_thread.py", line 56, in run_sync
    return await get_async_backend().run_sync_in_worker_thread(
  File "/content/dmo_env/lib/python3.10/site-packages/anyio/_backends/_asyncio.py", line 2470, in run_sync_in_worker_thread
    return await future
  File "/content/dmo_env/lib/python3.10/site-packages/anyio/_backends/_asyncio.py", line 967, in run
    result = context.run(func, *args)
  File "/content/dmo_env/lib/python3.10/site-packages/gradio/utils.py", line 917, in wrapper
    response = f(*args, **kwargs)
  File "/content/DMOSpeech2/scripts/colab-gradio.py", line 32, in prepare_audio_for_download
    torchaudio.save(audio_path, audio_tensor, sr)
  File "/content/dmo_env/lib/python3.10/site-packages/torchaudio/_backend/utils.py", line 313, in save
    return backend.save(
  File "/content/dmo_env/lib/python3.10/site-packages/torchaudio/_backend/ffmpeg.py", line 316, in save
    save_audio(
  File "/content/dmo_env/lib/python3.10/site-packages/torchaudio/_backend/ffmpeg.py", line 248, in save_audio
    s.add_audio_stream(
  File "/content/dmo_env/lib/python3.10/site-packages/torio/io/_streaming_media_encoder.py", line 278, in add_audio_stream
    self._s.add_audio_stream(
RuntimeError: Failed to create input filter: "time_base=1/44100:sample_rate=44100:sample_fmt=s16:channel_layout=0x0" (Invalid argument)
Exception raised from add_src at /__w/audio/audio/pytorch/audio/src/libtorio/ffmpeg/filter_graph.cpp:91 (most recent call first):
frame #0: c10::Error::Error(c10::SourceLocation, std::string) + 0x57 (0x79ecafd81d87 in /content/dmo_env/lib/python3.10/site-packages/torch/lib/libc10.so)
frame #1: c10::detail::torchCheckFail(char const*, char const*, unsigned int, std::string const&) + 0x64 (0x79ecafd3275f in /content/dmo_env/lib/python3.10/site-packages/torch/lib/libc10.so)
frame #2: torio::io::FilterGraph::add_src(AVFilter const*, std::string const&) + 0x37f (0x79ec3f2a771f in /content/dmo_env/lib/python3.10/site-packages/torio/lib/libtorio_ffmpeg4.so)
frame #3: torio::io::FilterGraph::add_audio_src(AVSampleFormat, AVRational, int, unsigned long) + 0x31 (0x79ec3f2a7821 in /content/dmo_env/lib/python3.10/site-packages/torio/lib/libtorio_ffmpeg4.so)
frame #4: torio::io::get_audio_encode_process(AVFormatContext*, int, int, std::string const&, std::optional<std::string> const&, std::optional<std::map<std::string, std::string, std::less<std::string>, std::allocator<std::pair<std::string const, std::string> > > > const&, std::optional<std::string> const&, std::optional<int> const&, std::optional<int> const&, std::optional<torio::io::CodecConfig> const&, std::optional<std::string> const&, bool) + 0x7b7 (0x79ec3f2d6207 in /content/dmo_env/lib/python3.10/site-packages/torio/lib/libtorio_ffmpeg4.so)
frame #5: torio::io::StreamingMediaEncoder::add_audio_stream(int, int, std::string const&, std::optional<std::string> const&, std::optional<std::map<std::string, std::string, std::less<std::string>, std::allocator<std::pair<std::string const, std::string> > > > const&, std::optional<std::string> const&, std::optional<int> const&, std::optional<int> const&, std::optional<torio::io::CodecConfig> const&, std::optional<std::string> const&) + 0x90 (0x79ec3f2dd5d0 in /content/dmo_env/lib/python3.10/site-packages/torio/lib/libtorio_ffmpeg4.so)
frame #6: <unknown function> + 0x3b233 (0x79ebe059a233 in /content/dmo_env/lib/python3.10/site-packages/torio/lib/_torio_ffmpeg4.so)
frame #7: <unknown function> + 0x32147 (0x79ebe0591147 in /content/dmo_env/lib/python3.10/site-packages/torio/lib/_torio_ffmpeg4.so)
<omitting python frames>
frame #34: <unknown function> + 0x94ac3 (0x79ecb0b56ac3 in /lib/x86_64-linux-gnu/libc.so.6)
frame #35: <unknown function> + 0x126850 (0x79ecb0be8850 in /lib/x86_64-linux-gnu/libc.so.6)

Traceback (most recent call last):
  File "/content/dmo_env/lib/python3.10/site-packages/gradio/queueing.py", line 626, in process_events
    response = await route_utils.call_process_api(
  File "/content/dmo_env/lib/python3.10/site-packages/gradio/route_utils.py", line 350, in call_process_api
    output = await app.get_blocks().process_api(
  File "/content/dmo_env/lib/python3.10/site-packages/gradio/blocks.py", line 2235, in process_api
    result = await self.call_function(
  File "/content/dmo_env/lib/python3.10/site-packages/gradio/blocks.py", line 1746, in call_function
    prediction = await anyio.to_thread.run_sync(  # type: ignore
  File "/content/dmo_env/lib/python3.10/site-packages/anyio/to_thread.py", line 56, in run_sync
    return await get_async_backend().run_sync_in_worker_thread(
  File "/content/dmo_env/lib/python3.10/site-packages/anyio/_backends/_asyncio.py", line 2470, in run_sync_in_worker_thread
    return await future
  File "/content/dmo_env/lib/python3.10/site-packages/anyio/_backends/_asyncio.py", line 967, in run
    result = context.run(func, *args)
  File "/content/dmo_env/lib/python3.10/site-packages/gradio/utils.py", line 917, in wrapper
    response = f(*args, **kwargs)
  File "/content/DMOSpeech2/scripts/colab-gradio.py", line 48, in initialize_voice
    torchaudio.save(audio_path, audio_tensor, sr)
  File "/content/dmo_env/lib/python3.10/site-packages/torchaudio/_backend/utils.py", line 313, in save
    return backend.save(
  File "/content/dmo_env/lib/python3.10/site-packages/torchaudio/_backend/ffmpeg.py", line 316, in save
    save_audio(
  File "/content/dmo_env/lib/python3.10/site-packages/torchaudio/_backend/ffmpeg.py", line 248, in save_audio
    s.add_audio_stream(
  File "/content/dmo_env/lib/python3.10/site-packages/torio/io/_streaming_media_encoder.py", line 278, in add_audio_stream
    self._s.add_audio_stream(
RuntimeError: Failed to create input filter: "time_base=1/44100:sample_rate=44100:sample_fmt=s16:channel_layout=0x0" (Invalid argument)
Exception raised from add_src at /__w/audio/audio/pytorch/audio/src/libtorio/ffmpeg/filter_graph.cpp:91 (most recent call first):
frame #0: c10::Error::Error(c10::SourceLocation, std::string) + 0x57 (0x79ecafd81d87 in /content/dmo_env/lib/python3.10/site-packages/torch/lib/libc10.so)
frame #1: c10::detail::torchCheckFail(char const*, char const*, unsigned int, std::string const&) + 0x64 (0x79ecafd3275f in /content/dmo_env/lib/python3.10/site-packages/torch/lib/libc10.so)
frame #2: torio::io::FilterGraph::add_src(AVFilter const*, std::string const&) + 0x37f (0x79ec3f2a771f in /content/dmo_env/lib/python3.10/site-packages/torio/lib/libtorio_ffmpeg4.so)
frame #3: torio::io::FilterGraph::add_audio_src(AVSampleFormat, AVRational, int, unsigned long) + 0x31 (0x79ec3f2a7821 in /content/dmo_env/lib/python3.10/site-packages/torio/lib/libtorio_ffmpeg4.so)
frame #4: torio::io::get_audio_encode_process(AVFormatContext*, int, int, std::string const&, std::optional<std::string> const&, std::optional<std::map<std::string, std::string, std::less<std::string>, std::allocator<std::pair<std::string const, std::string> > > > const&, std::optional<std::string> const&, std::optional<int> const&, std::optional<int> const&, std::optional<torio::io::CodecConfig> const&, std::optional<std::string> const&, bool) + 0x7b7 (0x79ec3f2d6207 in /content/dmo_env/lib/python3.10/site-packages/torio/lib/libtorio_ffmpeg4.so)
frame #5: torio::io::StreamingMediaEncoder::add_audio_stream(int, int, std::string const&, std::optional<std::string> const&, std::optional<std::map<std::string, std::string, std::less<std::string>, std::allocator<std::pair<std::string const, std::string> > > > const&, std::optional<std::string> const&, std::optional<int> const&, std::optional<int> const&, std::optional<torio::io::CodecConfig> const&, std::optional<std::string> const&) + 0x90 (0x79ec3f2dd5d0 in /content/dmo_env/lib/python3.10/site-packages/torio/lib/libtorio_ffmpeg4.so)
frame #6: <unknown function> + 0x3b233 (0x79ebe059a233 in /content/dmo_env/lib/python3.10/site-packages/torio/lib/_torio_ffmpeg4.so)
frame #7: <unknown function> + 0x32147 (0x79ebe0591147 in /content/dmo_env/lib/python3.10/site-packages/torio/lib/_torio_ffmpeg4.so)
<omitting python frames>
frame #34: <unknown function> + 0x94ac3 (0x79ecb0b56ac3 in /lib/x86_64-linux-gnu/libc.so.6)
frame #35: <unknown function> + 0x126850 (0x79ecb0be8850 in /lib/x86_64-linux-gnu/libc.so.6)

Traceback (most recent call last):
  File "/content/dmo_env/lib/python3.10/site-packages/gradio/queueing.py", line 626, in process_events
    response = await route_utils.call_process_api(
  File "/content/dmo_env/lib/python3.10/site-packages/gradio/route_utils.py", line 350, in call_process_api
    output = await app.get_blocks().process_api(
  File "/content/dmo_env/lib/python3.10/site-packages/gradio/blocks.py", line 2235, in process_api
    result = await self.call_function(
  File "/content/dmo_env/lib/python3.10/site-packages/gradio/blocks.py", line 1746, in call_function
    prediction = await anyio.to_thread.run_sync(  # type: ignore
  File "/content/dmo_env/lib/python3.10/site-packages/anyio/to_thread.py", line 56, in run_sync
    return await get_async_backend().run_sync_in_worker_thread(
  File "/content/dmo_env/lib/python3.10/site-packages/anyio/_backends/_asyncio.py", line 2470, in run_sync_in_worker_thread
    return await future
  File "/content/dmo_env/lib/python3.10/site-packages/anyio/_backends/_asyncio.py", line 967, in run
    result = context.run(func, *args)
  File "/content/dmo_env/lib/python3.10/site-packages/gradio/utils.py", line 917, in wrapper
    response = f(*args, **kwargs)
  File "/content/DMOSpeech2/scripts/colab-gradio.py", line 58, in generate_audio
    raise gr.Error("Initialize voice first!")
gradio.exceptions.Error: 'Initialize voice first!'
Traceback (most recent call last):
  File "/content/dmo_env/lib/python3.10/site-packages/gradio/queueing.py", line 626, in process_events
    response = await route_utils.call_process_api(
  File "/content/dmo_env/lib/python3.10/site-packages/gradio/route_utils.py", line 350, in call_process_api
    output = await app.get_blocks().process_api(
  File "/content/dmo_env/lib/python3.10/site-packages/gradio/blocks.py", line 2235, in process_api
    result = await self.call_function(
  File "/content/dmo_env/lib/python3.10/site-packages/gradio/blocks.py", line 1746, in call_function
    prediction = await anyio.to_thread.run_sync(  # type: ignore
  File "/content/dmo_env/lib/python3.10/site-packages/anyio/to_thread.py", line 56, in run_sync
    return await get_async_backend().run_sync_in_worker_thread(
  File "/content/dmo_env/lib/python3.10/site-packages/anyio/_backends/_asyncio.py", line 2470, in run_sync_in_worker_thread
    return await future
  File "/content/dmo_env/lib/python3.10/site-packages/anyio/_backends/_asyncio.py", line 967, in run
    result = context.run(func, *args)
  File "/content/dmo_env/lib/python3.10/site-packages/gradio/utils.py", line 917, in wrapper
    response = f(*args, **kwargs)
  File "/content/DMOSpeech2/scripts/colab-gradio.py", line 48, in initialize_voice
    torchaudio.save(audio_path, audio_tensor, sr)
  File "/content/dmo_env/lib/python3.10/site-packages/torchaudio/_backend/utils.py", line 313, in save
    return backend.save(
  File "/content/dmo_env/lib/python3.10/site-packages/torchaudio/_backend/ffmpeg.py", line 316, in save
    save_audio(
  File "/content/dmo_env/lib/python3.10/site-packages/torchaudio/_backend/ffmpeg.py", line 248, in save_audio
    s.add_audio_stream(
  File "/content/dmo_env/lib/python3.10/site-packages/torio/io/_streaming_media_encoder.py", line 278, in add_audio_stream
    self._s.add_audio_stream(
RuntimeError: Failed to create input filter: "time_base=1/44100:sample_rate=44100:sample_fmt=s16:channel_layout=0x0" (Invalid argument)
Exception raised from add_src at /__w/audio/audio/pytorch/audio/src/libtorio/ffmpeg/filter_graph.cpp:91 (most recent call first):
frame #0: c10::Error::Error(c10::SourceLocation, std::string) + 0x57 (0x79ecafd81d87 in /content/dmo_env/lib/python3.10/site-packages/torch/lib/libc10.so)
frame #1: c10::detail::torchCheckFail(char const*, char const*, unsigned int, std::string const&) + 0x64 (0x79ecafd3275f in /content/dmo_env/lib/python3.10/site-packages/torch/lib/libc10.so)
frame #2: torio::io::FilterGraph::add_src(AVFilter const*, std::string const&) + 0x37f (0x79ec3f2a771f in /content/dmo_env/lib/python3.10/site-packages/torio/lib/libtorio_ffmpeg4.so)
frame #3: torio::io::FilterGraph::add_audio_src(AVSampleFormat, AVRational, int, unsigned long) + 0x31 (0x79ec3f2a7821 in /content/dmo_env/lib/python3.10/site-packages/torio/lib/libtorio_ffmpeg4.so)
frame #4: torio::io::get_audio_encode_process(AVFormatContext*, int, int, std::string const&, std::optional<std::string> const&, std::optional<std::map<std::string, std::string, std::less<std::string>, std::allocator<std::pair<std::string const, std::string> > > > const&, std::optional<std::string> const&, std::optional<int> const&, std::optional<int> const&, std::optional<torio::io::CodecConfig> const&, std::optional<std::string> const&, bool) + 0x7b7 (0x79ec3f2d6207 in /content/dmo_env/lib/python3.10/site-packages/torio/lib/libtorio_ffmpeg4.so)
frame #5: torio::io::StreamingMediaEncoder::add_audio_stream(int, int, std::string const&, std::optional<std::string> const&, std::optional<std::map<std::string, std::string, std::less<std::string>, std::allocator<std::pair<std::string const, std::string> > > > const&, std::optional<std::string> const&, std::optional<int> const&, std::optional<int> const&, std::optional<torio::io::CodecConfig> const&, std::optional<std::string> const&) + 0x90 (0x79ec3f2dd5d0 in /content/dmo_env/lib/python3.10/site-packages/torio/lib/libtorio_ffmpeg4.so)
frame #6: <unknown function> + 0x3b233 (0x79ebe059a233 in /content/dmo_env/lib/python3.10/site-packages/torio/lib/_torio_ffmpeg4.so)
frame #7: <unknown function> + 0x32147 (0x79ebe0591147 in /content/dmo_env/lib/python3.10/site-packages/torio/lib/_torio_ffmpeg4.so)
<omitting python frames>
frame #34: <unknown function> + 0x94ac3 (0x79ecb0b56ac3 in /lib/x86_64-linux-gnu/libc.so.6)
frame #35: <unknown function> + 0x126850 (0x79ecb0be8850 in /lib/x86_64-linux-gnu/libc.so.6)

[wrightmikea]

wrightmikea, good work. I launched dmospeech2 on colab an hour ago from your notepad, but I couldn't initialize the voice.

What steps did you take? Did you use the microphone or upload a file, and if so, what kind of file? Are you using the T4 GPU? I only tested with that in colab. ---> @mikhail2013ru did you get it to work? <--- [edit 2025 July 26]

There is a working reference audio in the repo: src/f5_tts/infer/examples/basic/basic_ref_en.wav which I just tested with:

A) select tab Initialize Voice;

1. upload ref file 
   

2. type in the matching text 
   

3. click Prepare Reference Audio for download (download this for a quality check--earlier attempts distorted this audio) 
   

4. click Initialize Voice
   

[EDIT]
From the stacktrace File "/content/DMOSpeech2/scripts/colab-gradio.py", line 58, in generate_audio raise gr.Error("Initialize voice first!") gradio.exceptions.Error: 'Initialize voice first!' it seems you may have not clicked this button?

B) select tab Generate Speech

1. type in text to synthesize
   

2. click Generate Audio.  
   

these steps just worked for me with no output errors in the notebook (the ref text shows up in the last cell's output). I didn't spend any time improving the UI and took the first one that Claude Code generated.

Maybe someone else can help with the stack trace (I don't use Python a lot). Good luck

(I may be updating my repo next week with a FastAPI example notebook and sample curl commands or other API clients). I generally do not use colab; I have a homelab with some old servers (old GPUs) running Jupyter Lab and standalone Python scripts (Gradio, Fast API) which I call from an MCP server I wrote in Rust to generate dialogs between multiple speakers.