Feedback for NatureLM-audio UI: Hugging Face Spaces Demo

NatureLM-audio
1

We just released our first demo of NatureLM-audio on :hugs: Hugging Face Spaces!

This is the main thread for sharing feedback, use cases, and asking questions for the Hugging Face UI demo.

We’d love to hear from you! Please share:

  • Background context: Is this for a research project? Side hobby project? Just exploring for fun?
  • What you tested: What type of animal sounds did you upload? (whale songs, bird calls, etc.)
  • How it went: Share your back and forth conversation with NatureLM-audio. Did the model answer correctly? Any surprising results (good or bad)?
  • Audio files: If you’re comfortable sharing, links to the audio files are super helpful for our team to understand better what happened
  • Future ideas or questions: Feature requests, use cases we hadn’t considered, technical questions

The more details you can share, the better we can improve NatureLM-audio together!

3

So I tried successfully the no-code UI and now I would like to access it from my application, so do you have an API that I could use to get the species from short dolphin whistles that my app picks up live from dolphins at sea? My app works fine on a tablet and I could add some http requests to it to access your api.

I tried this curl without success:
curl -X POST \

-H “Authorization: Bearer my-token” \

-H “Content-Type: application/json” \

-d ‘{“data”: [“@5801400J.txt”, “What is the species of this call?”]}’ \

NatureLM-audio

My app would use equivalent http requests to this curl.

Thank you,
Serge

4

Hi serge,

We created this Huggingface space to understand how the first (beta) version of NatureLM-audio UI is used, it is not ready for an API release. This may come in the future.

thanks for trying it out

best,

Gagan

ML engineer at ESP