INSANELY FAST Talking AI: Powered by Groq & Deepgram
HTML-код
- Опубликовано: 2 июн 2024
- Fastest Voice Chat Inference with Groq and DeepGram
In this video, I show how to achieve the fastest voice chat inference using Groq and DeepGram APIs. I compare their speeds to OpenAI’s Whisper and demonstrate how to set up and code the process. Learn about handling rate limits, buffering issues, and how to get started with these services. Stay tuned for future videos on local model implementations.
#groq #voicechat #whisper
🦾 Discord: / discord
☕ Buy me a Coffee: ko-fi.com/promptengineering
|🔴 Patreon: / promptengineering
💼Consulting: calendly.com/engineerprompt/c...
📧 Business Contact: engineerprompt@gmail.com
Become Member: tinyurl.com/y5h28s6h
💻 Pre-configured localGPT VM: bit.ly/localGPT (use Code: PromptEngineering for 50% off).
Signup for Advanced RAG:
tally.so/r/3y9bb0
LINKS:
Updated code now released
Project Verbi: github.com/PromtEngineer/Verbi
Whisper on Groq vs OpenAI: tinyurl.com/5ea42yn4
00:00 Introduction to Advanced Voice Chat Inference
00:10 Meet Ada: Your AI Assistant
01:17 Exploring OpenAI's Implementation
01:45 Switching to Groq and Deepgram for Speed
02:43 Deep Dive into the New Implementation
06:22 Setting Up Your Environment
10:28 Understanding Rate Limits and Service Credits
11:19 Looking Ahead: Local Models and Community Engagement
All Interesting Videos:
Everything LangChain: • LangChain
Everything LLM: • Large Language Models
Everything Midjourney: • MidJourney Tutorials
AI Image Generation: • AI Image Generation Tu... 
Наука
Releasing the updated code with "project Verbi": Same functionality shown in the video but better! github.com/PromtEngineer/Verbi
Completely local or runpod-even if slow, would be wonderful. No external services beyond your own.
That version is coming soon
The vocie spund pissed 😂😂😂
yeah came here to say this also 😂
So glad to have stumbled across this video, I've been looking for something similar to this all week. Thanks for putting this together, I can tell it will be very useful for me! I'm excited to see the streaming functionality on the top of the roadmap as well.
Glad it was helpful! The repo already has some updates. More to come soon.
Bro 4s response times are god fearingly slow😂
Just checked your code. Set stream to true for your TTS,SST and LLM. Then switch your TTS to PlayHT.
Looking forward to local version
Exciting!
You're a legend
Thank you 🙏🙏🙏
Very cool
I'm in Fresno CA and it's 4am😮
Please upload the same using local model. Thanks.
Incredible content! Could you post a version of this implementation that is locally hosted without subscriptions or fees? Thanks!!!!
Thank you, yes, will be updating it in future
I guess i should expect performance issues if I extend this to RAG use cases? Any recommendations for a performance optimized vector database?
I don't see performance issues but you will need to manage when the audio generation starts
How to use it with Flowise?
please do video for local opensource setup of this
Just pushed open source models for to TTS and STT. Will create a video on the updated package after some optimization
@@engineerprompt Great Thanks looking forward to that.
I keep getting "Server disconnected without sending a response."
Which model are you using?
Why do you have sleep(1) in the code?
I was running into issues of the llm generating the response but the audio would be just last few words without giving the system time to finishing playing the audio. Need a better solution :)
I am doing Bachelor in AI.I want more to learn Gen AI
Unfortunately, with Groq it doesn't sound as if it would be suitable for rebuilding free of charge :(
You could do this almost with only Deepgram as well.
I hope groq API have option to let dev continue their API for free, but in sense of incubator fee...
I hope so too
We cannot sacrifice quality of voice anyways.
Probably should have picked a better voice from available voices 😀
@@engineerprompt May be. but this will get outdated soon, what do you say? AI have taken a rapid pace. one after another advancements.
Would implement instantly the day groq releases api payments bc rn rates limits are too strict since it's free