no output

[felixbur]

hi, and thanks for doing this!

all seems fine, but i get the output

orpheus-tts-local$ python gguf_orpheus.py --text "This is mostly a test sentence, I'd like to know how it sounds. Haha, that's awesome!" --voice tara --output "output.wav"
Generating speech for: <|audio|>tara: This is mostly a test sentence, I'd like to know how it sounds. Haha, that's awesome!<|eot_id|>
Token generation complete
Generated 0 audio segments
Generated 0.00 seconds of audio
Speech generation completed in 14.19 seconds
Audio saved to output.wav

i don't see an error, also not on LM studio log:

2025-03-20 16:46:29  [INFO] 
[LM STUDIO SERVER] Processing...
2025-03-20 16:46:30  [INFO] 
[LM STUDIO SERVER] First token generated. Continuing to stream response..
2025-03-20 16:46:43  [INFO] 
[LM STUDIO SERVER] Finished streaming response
2025-03-20 16:46:43  [INFO] 
[LM STUDIO SERVER] Client disconnected. Stopping generation... (If the model is busy processing the prompt, it will finish first.)

any hint appreciated
(on ubuntu 20.04)

[m-atthart]

This might not be your specific issue, but I still think there's a chance this might help some people who run into this.
It was my first time using the LM Studio local server, and I got the same thing until I read another comment and realized how their local server works.
Make sure after you start the server, you click "Select a model to load" at the top and load in the Orpheus model.

[felixbur]

thanks, that may even be it, i don't have experience with LM Studio and found the UI not super intuitive.
actually, thanks to this colab i could run orpheus models on a gpu.

[sh4iny]

I got same issue too. I really want to run that but i can't.
orpheus-tts-local> python gguf_orpheus.py --text "Hello, this is a test" --voice tara No output file specified. Saving to outputs/tara_20250406_052821.wav Generating speech for: <|audio|>tara: Hello, this is a test<|eot_id|> Error: API request failed with status code 502 Error details: Generated 0 audio segments Generated 0.00 seconds of audio Speech generation completed in 0.13 seconds Audio saved to outputs/tara_20250406_052821.wav