Skip to content
/ Infernos Public

Real-time multi-modal inference server for interacting with humans and other intelligences around

License

Notifications You must be signed in to change notification settings

sippy/Infernos

Repository files navigation

I.N.F.E.R.N.O.S.

Interactive Neural Framework for Efficient Realtime Network Operations on Streams

🔥 Welcome to Infernos, where data comes to life in real-time! 🔥

Overview

Harness the power of I.N.F.E.R.N.O.S. to transform audio, video, and text streams with state-of-the-art inference in an instant. Embrace a blazing-fast future, free from lag.

News

Work is underway to scale TTS performance 50x. Stay tuned.

I am going to present at FOSDEM 2024 (RTC track). See you all there!

Features

  • Interactive: Infernos isn't just another tool; it's an experience. Speak in one voice and marvel as it's automatically translated into a completely different tone or even language, and then seamlessly transmitted in real-time during phone or web meetings.

  • Neural Power: With deep learning at its core, Infernos is optimized for top-notch performance.

  • Multimodal Support: Whether it's audio, video, or text, Infernos handles them with elegance.

  • Efficiency: Designed for low-latency, high-throughput operations.

  • Realtime: Don't wait. Experience the magic as it unfolds.

Quick Start

  1. Clone the repository:

    git clone https://github.com/sippy/Infernos.git
  2. Navigate to the project directory and install dependencies:

    cd Infernos && pip install -r requirements.txt
  3. Create a configuration file. In the following example we would listen and accept SIP calls from MY_IP and pass them into Live Translator application. Then use SIP account to send outbound call legs to DEST_NUM@MY_SIP_SRV:

    MY_IP="A.B.C.D"
    MY_SIP_SRV="E.F.G.H"
    DEST_NUM="12345"
    DEST_USER="foo"
    DEST_PWD="bar"
    cat > config.yaml <<EOF
    sip:
      settings:
        bind: ${MY_IP}:5060
      profiles:
        me:
          sip_server: ${MY_IP}:*
          sink: apps/live_translator/configuration1
        bar:
          sip_server: ${MY_SIP_SRV}:5060
          username: '${DEST_NUM}'
          password: '${DEST_PWD}'
          register: True
    apps:
      live_translator:
        profiles:
          configuration1:
            stt_langs: ['en', 'pt']
            tts_langs: ['pt', 'en']
            outbound: sip/bar;cld=${DEST_NUM}
    EOF
  4. Light the fire:

    python Infernos.py -f -L ~/Infernos.log
  5. Use SIP device or software such as Linphone to place a SIP call to sip:anything@localhost:5060. Replace localhost with a local IP of machine running Infernos if testing over LAN.

Ready to experience real-time inferencing?

Contribute

Feeling the warmth? 🔥 Eager to stoke the flames of Infernos? Delve into our contribution guidelines and join the firestorm!

License & Acknowledgements

Powered by the 2-clause BSD license. A heartfelt shoutout to the community for their priceless insights and tireless contributions.

Media

Join US


Stay on the lookout for more sizzling updates, and always remember: Infernos makes the future sizzle!

About

Real-time multi-modal inference server for interacting with humans and other intelligences around

Resources

License

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published