I Made An AI That Says All The Crazy Sh*t You Can’t Say At Work

Full stream here: https://www.youtube.com/watch?v=Scs-8BOrjYo

I streaming this for 7 hours the other day building a pipeline to gather podcast transcripts, convert them to a Q/A dataset (uses diarization will work with any podcast) and pass them through to a fine-tuning process (llama 2 + QLoRa) Enjoy!

💰Support the stream!
CashApp: https://cash.app/$jawerty210
Venmo: https://venmo.com/jawerty210
Buy Me a Coffee: https://www.buymeacoffee.com/bjGHFVW355

Join The Discord: https://discord.gg/dv4TSzsk27

My Socials:
Github: https://github.com/jawerty
Twitter: https://twitter.com/jawerty
LinkedIn: https://www.linkedin.com/in/jawerty/
Twitch: https://www.twitch.tv/jaredthecoder10x
Rumble: https://rumble.com/c/c-3572412

00:00 Intro
00:30 Explanation / How it'll work
03:00 The architecture + what is diarization?
04:20 Setting up the colab
05:46 Testing the audio segmenting
06:45 Fetching a Tate video
7:59 Plugging my Podcast
08:27 Organizing the code
08:52 Audio Segments to Dataset script
12:29 I start training the model (finally)
13:25 The first test
14:00 Fixing the model
16:05 Model's getting better...
17:00 Prompt tuning
18:10 Now it's sounding BASED
19:05 Saying Goodbye