Build A Talking AI with LLAMA 3 (Python tutorial)

🔑 Get your AssemblyAI API key here: www.assemblyai.com/?...
Code Repo: github.com/smithakolan/Assemb...
Learn to build a talking AI! This tutorial covers real-time transcription with AssemblyAI, using LLAMA 3 as the language model with Ollama, and ElevenLabs for text-to-speech.
Timestamps:
00:00 - Demo
00:17 - How we will build a talking AI with LLAMA 3
00:34 - Streaming Speech-to-text with AssemblyAI
02:07 - Step 1: Installing python libraries
05:35 - Step 2: Streaming Speech-to-text with AssemblyAI
09:10 - Step 3: Generating an AI response with Ollama & LLAMA 3 and Text-to-speech using ElevenLabs
▬▬▬▬▬▬▬▬▬▬▬▬ CONNECT ▬▬▬▬▬▬▬▬▬▬▬▬
🖥️ Website: www.assemblyai.com
🐦 Twitter: / assemblyai
🦾 Discord: / discord
▶️ Subscribe: ua-cam.com/users/AssemblyAI?...
🔥 We're hiring! Check our open roles: www.assemblyai.com/careers
▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬
#MachineLearning #DeepLearning

[@AssemblyAI]

🚀MEGA UPDATE 🚀
We've launched Universal-1, our most powerful and accurate multilingual speech-to-text model to date-trained on 12.5M hours of multilingual audio data.
www.assemblyai.com/blog/announcing-universal-1-speech-recognition-model/?

[@sanjaykrish8719]

is it free 😜😜

[@Threecommaaclub]

When running the code i receive the following message in the terminal "`ALSA lib confmic.c.:160,
i think this may be a warning message but i want to surpress them, how would i go about doing that?

[@cheybrown2076]

Amazing video 🎉.
Was portaudio the library used to listen and capture the audio?

[@naveenalla3000]

wonderful project

[@gewinnde]

Can live speech transcription be used in German or any other language than English ??

[@MrOktony]

Don’t forget you can also use text to speech , found in settings accessibility - it’s not real time, but at least it’s private. It would be awesome if you could get terminal speaking in real time, without using web based services..

[@AceOnBase1]

But then assemblyai wouldnt make money :/

[@fiorellademedina8419]

hello, how you can know the version of llama 3 with ollama? Is it 8B or 70B.?

[@AssemblyAI]

The default 'llama3' model is 8B with ollama. If you want to call the 70B model you need to specifiy 'llama3:70b'. Check our their naming conventions here: github.com/ollama/ollama/blob/main/docs/api.md#conventions

[@marketfinds]

Sorry, newbie, what is the 'source' command you use at the start?
You also jumped from step 1 to 6.
ie when I run: The term 'ollama' is not recognized as the name of a cmdlet
Thanks

[@AssemblyAI]

The 'source' command is used when activating the python virtual environment.
As for ollama you might not have installed it. So run 'pip install ollama' before using it. And additionally, you need to download it here: ollama.com

[@wesleydunn169]

Genius!

[@rabidrabbitmeow]

Any idea how to deal with this?
OSError: [Errno -9996] Invalid input device (no default output device)

[@tanb13]

How can I integrate with a bot created using chainlit ?

[@Roopkishore726]

some problems to make android application.
Please make a video to make application for convert audio file to text in android studio when button click and text show in textview.

[@cheybrown2076]

Source is used when activating virtual environment on Linux system.
Ollama not being recognised is probably because you haven’t installed it or you have but aren’t in the environment where you’ve installed it

[@cheybrown2076]

Ah I replied to the wrong comment lol. My bad. This was meant for a different comment

[@ojasvisingh786]

👏👏

[@gileneusz]

I'm not sure if talking to animals is a good idea....

[@Pixelatedcharts]

Can we use this api free for Life time ?

[@TheNewVision1003]

is the AI limited to what it can say?

[@unkillablespirits8582]

i refuse to pay for speech to text , i want this idea but free

[@GamersHobby]

Just show us the github and copy past, seeing you typing is so frustrated