Тёмный

Advanced Function Calling with Mistral-7B - Multi function and Nested Tool Usage 

Prompt Engineering
Подписаться 161 тыс.
Просмотров 6 тыс.
50% 1

Testing Multi and Nested Function Calls with Mistral 7b
In this video, I explore the advanced function calling capabilities of the Mistral 7b v3 model, including multi-function and nested function calls. Using a Google Colab notebook by Uncle Code, I demonstrate how to set up, install the Mistral inference package, and log into Hugging Face hub. Practical examples of the model handling multiple and sequential function calls are provided. Check the description for related links and detailed how-tos.
#llm #mistral #functioncalling
🦾 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/3y9bb0a
LINK:
Github: tinyurl.com/284bw4ka
X-post: x.com/unclecode/status/179402...
/ @unclecode788
TIMESTAMP:
00:00 Introduction to Advanced Function Calling in Mistral 7b
00:44 Understanding the Basics of Function Calling
02:54 Exploring Multi and Nested Function Calls
04:14 Deep Dive into Uncle Code's Notebook and Setup
06:13 Single Function Call Example and Analysis
08:16 Testing Multi Function Calls
09:28 Unpacking Nested Function Calls with Examples
12:30 Experimenting with Mixed Function Calls and Conclusion
All Interesting Videos:
Everything LangChain: • LangChain
Everything LLM: • Large Language Models
Everything Midjourney: • MidJourney Tutorials
AI Image Generation: • AI Image Generation Tu...

Наука

Опубликовано:

 

29 июн 2024

Поделиться:

Ссылка:

Скачать:

Готовим ссылку...

Добавить в:

Мой плейлист
Посмотреть позже
Комментарии : 20   
@unclecode
@unclecode Месяц назад
Really appreciate you creating this content and giving me credit. I have to admit that your term "nested tool" is much better than my "function dependency." 😅 I was trying to find a good term, and yours is definitely better. Again, thank you so much for reviewing this. Hopefully, it can help your audience. I'm a fan of your content and channels, and I wish you and everyone contributing to the open-source community the best. May the force be with you.
@engineerprompt
@engineerprompt Месяц назад
thank you for putting this together. glad you approve "Nested Tool" :) Looking forward to showcase some of your other work as well. Keep up the great work. I also love open-source community because of the collaborative nature and what it brings to the table :)
@ernestuz
@ernestuz Месяц назад
I have been building agents with the old Mixtral 8x7B and Mistral 0.2, and I have to say Mixtral many times outperforms models that have been trained for function calling, eager to see how this new Mistral performs..
@engineerprompt
@engineerprompt Месяц назад
Seems like its much smarter than what we have been able to achieve so far. I think specialized models is the way to go.
@MustafaMagdy1
@MustafaMagdy1 Месяц назад
Looking forward for a langgraph implementation of this
@engineerprompt
@engineerprompt Месяц назад
that would be actually good. Let me see what I can put together.
@aa-xn5hc
@aa-xn5hc Месяц назад
Great video
@engineerprompt
@engineerprompt Месяц назад
Thanks 😊
@Boudybuilder
@Boudybuilder Месяц назад
Maybe this question not about this video topic. But i want to know how i can create on my php web site smart search based on some local storages llm.
@engineerprompt
@engineerprompt Месяц назад
I have a video on what with your website, that might be a good start. Here (ru-vid.com/video/%D0%B2%D0%B8%D0%B4%D0%B5%D0%BE-RBnuhhmD21U.html) Its a very old one though
@user-rp2vv5oq6e
@user-rp2vv5oq6e 14 дней назад
Can this model have a conversation with the user and call the appropriate functions when necessary and use those function results to answer the question asked by the user in a conversational manner?
@engineerprompt
@engineerprompt 14 дней назад
Yes, that is the goal. If you ask questions that do not require a function call, it will act like normal LLM and when the input requires function call, it will use that to get info and then reply to the user with relevant info.
@user-rp2vv5oq6e
@user-rp2vv5oq6e 14 дней назад
@@engineerprompt Thanks! Would you be able to do a video on how to do that exactly?
@luigitech3169
@luigitech3169 Месяц назад
cool does it also work with GGUF model ?
@engineerprompt
@engineerprompt Месяц назад
Yes, it should
@mohsenghafari7652
@mohsenghafari7652 Месяц назад
gpu out of memory ... please help
@engineerprompt
@engineerprompt Месяц назад
It's not going to run on free colab. You need A100 for this one.
@mohsenghafari7652
@mohsenghafari7652 Месяц назад
@@engineerprompt thanks 🥲🥲
@kdjsgkhjhmrzazhgu
@kdjsgkhjhmrzazhgu Месяц назад
chainlink
Далее
С Анджилишей на тусе💃
00:15
Просмотров 164 тыс.
КУДА БОЛЬНЕЕ УПАСТЬ С ВЫСОТЫ?
25:11
НЕ ПОКУПАЙТЕ ЭТО!
00:29
Просмотров 26 тыс.
LLaMA3 & TinyLlama Fine-Tuned for Function Calls
11:00
Просмотров 2,1 тыс.
Have You Picked the Wrong AI Agent Framework?
13:10
Просмотров 45 тыс.
EXCLUSIVE: Torture Testing GPT-4o w/ SHOCKING Results!
22:00
iPhone 16 - КРУТЕЙШИЕ ИННОВАЦИИ
4:50