How to Fine Tune ChatGPT on Your Gmail

Рет қаралды 1,486

Күн бұрын

Master your Gmail with ChatGPT fine-tuning 🤖📧: Email like a pro without lifting a finger! ✨
Notebook: colab.research.google.com/dri...
🚀Custom AI Solutions: www.vrsen.ai/
🤝Need help with your next AI project? calendly.com/vrsen/ai-project...
🐦Twitter: / __vrsen__
Intro - 00:00
Guide: Fine Tuning ChatGPT on Gmail - 00:55
Conclusion - 05:18
About: Discover the power of fine-tuning ChatGPT for your Gmail! Unveil the magic behind AI-driven email automation, effortlessly replicate your email style, and free up valuable time for what truly matters. Dive into this step-by-step tutorial and transform the way you handle emails. 🚀💌

Пікірлер: 15

@Mr6499 7 ай бұрын

Amazing work mate 💪 Thank you so much, this is real content. A Channel very different !

@severian42 9 ай бұрын

Brilliant!

@nanukaskhilaze7885 9 ай бұрын

Great tutorial🚀

@vrsen 9 ай бұрын

Thank you!

@robboerman9378 4 ай бұрын

Hm, now I am starting to wonder if it is you replying to these comments or ChatGPT :D Great content, again, thanks for sharing.

@100cipher 7 ай бұрын

Love the content. You mentioned in the video to check in with you and ask "how do you code like that with GPT 4?" if we are interested. I'm curious - how do you code so well and is it duplicatable?

@merkury28 9 ай бұрын

I'm sure if it's not a good idea to chat gpt have access to all your emails just for convenience. Just imagine all the secrets that your email box contains, and also imagine that all those data is leaked..... Well for now I will pass.

@vrsen 9 ай бұрын

They claim not to train their models on data that you pass through the API. The files for fine tuning are stored on their side, however, so I guess there is a risk

@Evox402 9 ай бұрын

Hey nice video :) Good to know that OpenAI does now support fine tuning. I'd like to see how your coding workflow with ChatGPT looks like 👍 Have you ever used embedchain? I recently heard of it as a langchain alternative but haven't had the time to get a closer look.

@vrsen 9 ай бұрын

Thank you for the feedback! I’ve never tried embedchain, but the issue remains the same as with LangChain. They abstract things that can easily be implemented from scratch. I always prefer to create my own agent for better control over its behavior.

@nadiaprivalikhina4101 9 ай бұрын

Great content on your channel. Keep it up! May I ask what was the total cost for you to fine-tune the model? How much do you think fine-tuning is better for content creation compared to just providing instructions and referencing past content pieces?

@vrsen 9 ай бұрын

The cost for gpt-3.5 fine-tuning is not high; it should be less than a dollar for 100 emails, as in this tutorial. It will be significantly more effective for content generation if you have more than 100 pieces for training.

@gokulraja1580 9 ай бұрын

Hey arsen , put a video regarding llm with web search or Google search . Explain and implement this.

@vrsen 9 ай бұрын

I already have a similar video, but it is a bit outdated. You essentially want to recreate a web browsing plugin?

@Evox402 9 ай бұрын

@@vrsen maybe you can combine this and show your coding workflow. As you not use the common tools like langchain or Llama-Index I would be interested to see how you would code your own Agent, handle conversation memory, parsing etc. Maybe a small series where you build ChatGPT or a chat bot with conversational memory, the capability to use knowledge bases, parsing of the output and search the web (via an agent or so) and maybe some additional typical agent tasks. Don't know if this would get to technical though but I would love such a series/course :)