- Видео 65
- Просмотров 9 635
AI Tech Explained Right
Канада
Добавлен 22 июл 2023
Hey, I'm Right, and I've been working on AI for many years ... and I'm one of the most fans of machine learning challenges 😜.
I like to transfer knowledge about AI, machine learning and keep people up to date with the last trends in the Tech industry.
If you're interested to learn about what's shaping our present and our future, consider subscribing !!
see you !!
I like to transfer knowledge about AI, machine learning and keep people up to date with the last trends in the Tech industry.
If you're interested to learn about what's shaping our present and our future, consider subscribing !!
see you !!
Retrieval remotely using LangServe (LangChain Server) and FastAPI in few minutes!
Hi everyone I hope you're doing well and welcome back to the AI Tech explained Right channel, a place for AI professinals and entouthiasts to stay updated with AI advances
In this video I will present a simple example where we use a LangChain server to expose a retriever.
Please don't forget to subsctibe to the channel to get notified about new ai tech videos like this, put a thumb up on the video and to leave a comment to give a feedback :).
☕ Buy me a Coffee: ko-fi.com/aitechexplainedright
References:
github.com/langchain-ai/langserve
notebook (client part) and server file:
drive.google.com/drive/folders/1JLQJ1EctT0NPfibSILsKXYfdkHEjHg2G?usp=drive_link
In this video I will present a simple example where we use a LangChain server to expose a retriever.
Please don't forget to subsctibe to the channel to get notified about new ai tech videos like this, put a thumb up on the video and to leave a comment to give a feedback :).
☕ Buy me a Coffee: ko-fi.com/aitechexplainedright
References:
github.com/langchain-ai/langserve
notebook (client part) and server file:
drive.google.com/drive/folders/1JLQJ1EctT0NPfibSILsKXYfdkHEjHg2G?usp=drive_link
Просмотров: 22
Видео
Build an AGENT powered CHATBOT using LANGCHAIN in few minutes !
Просмотров 127Месяц назад
Hi everyone I hope you're doing well and welcome back to the AI Tech explained Right channel, a place for AI professinals and entouthiasts to stay updated with AI advances In this video I will show how to build and AGENT powered chatbot using Langchain in few minutes Please don't forget to subsctibe to the channel to get notified about new ai tech videos like this, put a thumb up on the video a...
UNCENSORED NEW MISTRAL-7B-INSTRUCT: more powerful with Function Calling feature!
Просмотров 66Месяц назад
Hi everyone I hope you're doing well and welcome back to the AI Tech explained Right channel, a place for AI professinals and entouthiasts to stay updated with AI advances In this video I will talk about the new Mistral-7B-Instruct-v0.3 Please don't forget to subsctibe to the channel to get notified about new ai tech videos like this, put a thumb up on the video and to leave a comment to give a...
A NEW and MORE CAPABLE ChatGPT with GPT-4o Released by OpenAI - all you need to know in few minutes!
Просмотров 43Месяц назад
Hi everyone I hope you're doing well and welcome back to the AI Tech explained Right channel, a place for AI professinals and entouthiasts to stay updated with AI advances In This video I talk about the Release of GPT-4o by OpenAI Please don't forget to subsctibe to the channel to get notified about new ai tech videos like this, put a thumb up on the video and to leave a comment to give a feedb...
7 GOLDEN RULES for prompt engineering by Anthropic team !
Просмотров 732 месяца назад
Hi everyone I hope you're doing well, and welcome back to the AI Tech explained Right channel, the place for AI professinals and entouthiasts to stay updated with AI advances. In This video I will talk about 7 golden techniques for prompt engeering based on a presentation made by Anthropic team. ☕ Buy me a Coffee: ko-fi.com/aitechexplainedright References: docs.anthropic.com/claude/docs/intro-t...
Prompt engineering and llms evaluations explained right !
Просмотров 672 месяца назад
hi everyone I hope you're doing well, and welcome back to the AI Tech explained Right channel, the place for AI professinals and entouthiasts to stay updated with AI advances In This video I will talk about prompt engeering and evals tips and tricks based on a presentation made by Anthropic team ☕ Buy me a Coffee: ko-fi.com/aitechexplainedright references: docs.anthropic.com/claude/docs/intro-t...
6000 $ PC is home !! All characteristics prensented !!
Просмотров 1482 месяца назад
hi everyone I hope you're doing well and welcome back to the AI Tech explained Right channel, a place for AI professinals and entouthiasts to stay updated with AI advances In this video I will present my 6000 CA$ GPU that I received. Please don't forget to subsctibe to the channel to get notified about new ai tech videos like this, put a thumb up on the video and to leave a comment to give a fe...
Microsoft Phi-3 Open Models FIT ON YOUR PHONE !!
Просмотров 602 месяца назад
hi everyone I hope you're doing well and welcome back to the AI Tech explained Right channel, a place for AI professinals and entouthiasts to stay updated with AI advances In This video I talk about the Release of Phi-3 models by Microsoft Please don't forget to subsctibe to the channel to get notified about new ai tech videos like this, put a thumb up on the video and to leave a comment to giv...
Meta Llama 3 Release: The most capable OPENLY available LLM - ALL YOU NEED TO KNOW in few minutes
Просмотров 843 месяца назад
hi everyone I hope you're doing well and welcome back to the AI Tech explained Right channel, a place for AI professinals and entouthiasts to stay updated with AI advances In This video I talk about the Release of llama3 models by Meta Please don't forget to subsctibe to the channel to get notified about new ai tech videos like this, put a thumb up on the video and to leave a comment to give a ...
Try out CLAUDE3, THE SMARTEST AI MODEL IN THE WORLD in few minutes!!
Просмотров 943 месяца назад
hi everyone I hope you're doing well and welcome back to the AI Tech explained Right channel, a place for AI professinals and entouthiasts to stay updated with AI advances In This video I show how you can try out CLAUDE3 models by ANTHROPIC Please don't forget to subsctibe to the channel to get notified about new ai tech videos like this, put a thumb up on the video and to leave a comment to gi...
CLAUDE 3 Release by ANTHROPIC - the SMARTEST AI MODEL in the WORLD!!
Просмотров 823 месяца назад
Hi everyone I hope you're doing well and welcome back to the AI Tech explained Right channel, a place for AI professinals and entouthiasts to stay updated with AI advances In This video I talk about the Release of CLAUDE3 models by ANTHROPIC Please don't forget to subsctibe to the channel to get notified about new ai tech videos like this, put a thumb up on the video and to leave a comment to g...
LangSmith to trace your LLM based application, build with LCEL - LangChain series
Просмотров 803 месяца назад
hi everyone I hope you're doing well and welcome back to the AI Tech explained Right channel, a place for AI professinals and entouthiasts to stay updated with AI advances. Please don't forget to subsctibe to the channel to get notified about new videos like this, put a thumb up on the video and to leave a comment to give a feedback :). ☕ Buy me a Coffee: ko-fi.com/aitechexplainedright In This ...
Retrieval Augmented Generation (RAG) made easier with LCEL - LangChain series
Просмотров 873 месяца назад
Retrieval Augmented Generation (RAG) made easier with LCEL - LangChain series
x.ai Open Release of Grok-1 LLM - all you need to know in few minutes !
Просмотров 994 месяца назад
x.ai Open Release of Grok-1 LLM - all you need to know in few minutes !
Build an LLM powered CHATBOT with RAG and LangChain in few seconds - LangChain series
Просмотров 654 месяца назад
Build an LLM powered CHATBOT with RAG and LangChain in few seconds - LangChain series
Retrieval Augmented Generation (RAG) for LLM based application with LangChain - LangChain series
Просмотров 4464 месяца назад
Retrieval Augmented Generation (RAG) for LLM based application with LangChain - LangChain series
Trace your LLM based application with LangSmith in few seconds - LangChain series
Просмотров 3034 месяца назад
Trace your LLM based application with LangSmith in few seconds - LangChain series
Ask an OpenAI LLM (GPT3.5) with LangChain in few seconds - LangChain series
Просмотров 1414 месяца назад
Ask an OpenAI LLM (GPT3.5) with LangChain in few seconds - LangChain series
OpenAI's Sora for Video Generation is a GameChanger
Просмотров 805 месяцев назад
OpenAI's Sora for Video Generation is a GameChanger
AI inside MIT museum - in person visit
Просмотров 735 месяцев назад
AI inside MIT museum - in person visit
Google's LUMIERE for Video Generation is MINDBLOWING
Просмотров 1475 месяцев назад
Google's LUMIERE for Video Generation is MINDBLOWING
TinyLlama 1.1B parameters language model runs locally, everything to know about the model !!
Просмотров 495 месяцев назад
TinyLlama 1.1B parameters language model runs locally, everything to know about the model !!
Create YOUR OWN custom OPENAI GPT - full tutorial - NO CODING
Просмотров 465 месяцев назад
Create YOUR OWN custom OPENAI GPT - full tutorial - NO CODING
OPENAI CUSTOM GPT s are just AWESOME, 9 cherry picked amazing GPT s applications!!
Просмотров 435 месяцев назад
OPENAI CUSTOM GPT s are just AWESOME, 9 cherry picked amazing GPT s applications!!
OPENAI GPT-4, GPT-4 VISION, 6 INSANE applications
Просмотров 2466 месяцев назад
OPENAI GPT-4, GPT-4 VISION, 6 INSANE applications
Will Google's AI REPLACE DOCTORS - Artificial Medical Intelligence
Просмотров 2986 месяцев назад
Will Google's AI REPLACE DOCTORS - Artificial Medical Intelligence
Google GEMINI beats OpenAI GPT4, GPT4-Vision and Whisper to name a few. EVERYTHING YOU NEED TO KNOW.
Просмотров 2636 месяцев назад
Google GEMINI beats OpenAI GPT4, GPT4-Vision and Whisper to name a few. EVERYTHING YOU NEED TO KNOW.
CODING, PROGRAMMING, GAMING, FOCUS, BOOST PRODUCTIVITY, SUSPENSE, ADVENTURE, MOTIVATION, MUSIC
Просмотров 937 месяцев назад
CODING, PROGRAMMING, GAMING, FOCUS, BOOST PRODUCTIVITY, SUSPENSE, ADVENTURE, MOTIVATION, MUSIC
What is Artificial Intelligence (AI) !!
Просмотров 627 месяцев назад
What is Artificial Intelligence (AI) !!
My left ear appreciates this tutorial :)
Hello ! Thanks for the comment! Please don’t forget to subscribe to the channel if you’re interested in the subject !
What a beast! Enjoy!!
Thanks !
Lmao, waste money, sorry you got scammed.
Thanks for the comment! Please subscribe to the channel if you’re interested in the subject! Have a nice day!
😂 "Nvidia RTX 490". I think you meant Nvidia RTX 4090 but that's a nice computer for sure.
It’s nvidia rtx 4090 🙂! Thanks for the comment! Please subscribe to the channel if you’re interested in the subject ! Have a wonderful day 🙂!
For almost $4,400 USD I would expect some sort of custom cooling and more storage. Beast of a system though
Thanks for the comment! Please don’t hesitate to subscribe to channel if you’re interested in the subject! Thanks in advance !
Ahh, around $4,300 American dollars, makes more sense now. Still a lot of money for a PC, but that is some high end stuff there. Enjoy!!
Thanks for the comment! Please don’t hesitate to subscribe to channel if you’re interested in the subject! Thanks in advance !
GPU only stands for graphics card thats a PC
Definitely !! thanks for the comment !! Please don’t forget to subscribe if the subject interests you ! :)
I looked into vision transformers, which transform a 16x16 image patch "token" (pixel->character, image-> sentence, image patch -> word or subword) into a vector embedding of size 512. The positional encoding is learnable, so that it can learn that the top left patch is next to patches on the right and bottom. (During training, all images have the same size. If the sizes are different during inference, there are some additional calculations to account for the fact that there are more or less image patches) Then we can pass the following sequence to a decoder: text_token_embedding, image_token_embedding, image_token_embedding, text_token_embedding. I am confused by whether they still use the traditional positional encoding for this sequence, given that it won't consider the 2d aspect of the image and that the image_token_embedding already includes a positional encoding from the vision transformer (which does understand the 2d aspect). It's pretty amazing what can be done! Also, having (decent) models on mobile devices that can be used for "free" would be pretty nice. I am guessing that there 3 separate decoders for text, image (Parti has an auto-regressive approach and the vocabulary of image tokens is learned) and sound that rely on the encodings from the shared embedding space. I wonder how it decides which decoder to use. For example, when prompted with "Draw a dog", it will respond with "Sure, here is a drawing of a dog:" and 4 images of dogs. Also, for videos, I wonder how you would represent the fact that the sound and images are occurring at the same time.
Getting this error in my vs code cannot import name 'create_history_aware_retriever' from 'langchain_core'
Hi, do you import create_history_aware_retriever from langchain.chains?
Tres bon trevail
Merci beaucoup 🙂
C bien toujours de l avant
Merci beaucoup 🙂
Bon travail
Merci 🙂
Sorry cant watch this. the music is distracting alot.
Hi, thanks for the feedback, I will consider this in the next videos. Please consider to subscribe to get notified about new videos. Thanks
Coool
Thanks 😊
Super 🎉
Thanks 😊
people inside the building: "Look at us, we are so unique and smart." me: "you do know all that information you are paying thousands of dollars for is 100% free online already right?"
I like the comment :), please subscribe to the channel if you’re interested in the subject. Thanks :)
@@AITechExplainedRight ask and you shall receive ❤
Them: the free resource is our breakfast of our mind.
Very interesting, well done Ghaith👏
Thanks 😊!! Many videos like these are coming !! Stay tuned !!
I thought custom GPTs would just have a custom system prompt, didn't expect the GPT Builder and conversation starters. Pretty cool! Do you know if the capabilities and custom actions are injected into the system message right after the instructions? I remember using openai function calling with langchain.debug = True and not seeing the function descriptions being passed to the llm. I was looking for something like "Use the following functions if necessary {function_schema}" in the prompt.
Hi, thanks for the interest to the subject and for the quesiton in the comments. I defenitely agree that custom GPTs are pretty cool, especially how they make building awesome things so fast!! To answer your question, function calling and tools in general are passed seperately to the assitants (openai) or agents (langchain) among other information like the name of your assistant, instructions and the model to be used, what I would recommand is to create assitant/agent in a custom way, you can even pass many tools, for the same purpose, in the tools field of the assistant, and you will have and assistant/agent per case. Something else that I recommand is to use langsmith, which helps very well to visualize different parts of your chain, when working with langchain, including their inputs and outputs. If you apply for langsmith you might be in the wainting list, myself I was, but you will have access one day:). Here are two links that I find interesting to compare gpts and assitants, and how to create an openai assitant programatically. help.openai.com/en/articles/8673914-gpts-vs-assistants platform.openai.com/docs/assistants/overview Thanks again for the question and please don't hesitate to subscribe to the channel to be notified about new videos like these. Have a good one :)
@AITechExplainedRight Oh right, I was in the LangSmith waiting list and I received an email that they went public recently! I was curious what the final prompt sent to the llm was "System: Tools: User: Assistant: User:", but the details are probably not very "open" (with the openai api, we send an array of messages and tools, it's not clear how they are assembled into a single prompt). Since the model was fine tuned for function calling, maybe there is no need to inject an additional instruction "Use one of the available tools if applicable" in the system message. (The model sees the list of tools and automatically knows what to do with them) Keep up the good work! :)
@@dropinscience6584 Hi again and very sorry for my delayed answer !! Yes exactly I received the notification a little bit late appearently !! I agree that we lack transparency without seeing how the different parts of the prompt are assembled. For the tools exactly no need to ask the model to choose between them in the system prompt, it's kind of done automatically when all the tools are passed in an array !! thanks again for your interest, and don't hesitate to ask any question in the comments of my videos, I would be more than happy to answer !! :)
This is just awesome, it can save lot of time, it has been months and I didn’t finish creating my web page , this application is just what I need. Thanks for sharing
You’re very welcome, stay tuned for more videos like these which are coming !!
Omg 😮, really amazing, thanks for sharing these useful information 👌
My pleasure!! More videos like this are coming 🙂
amazing !! please keep going !!
hopefully something can replace the voice in this so that it can be understood. thumbs down!!
Thanks for the feedback. I included subtitles when creating the video, you can activate them, maybe it could help. Indeed I’m working on improving the sound quality. Your feedback is important to improve the quality of the videos. Please subscribe to the channel if you are interested in the subject. Thanks
By now it's well known that the GEMINI demo was faked to make it look much better than it is, remove delay etc.
thanks for the feedback :). But the way the multimodality has been used in GEMINI and at what extent it improved performace in several taks is extraordinary, it opens and proves many opportunities in academic and industrial fields. please don't forget to subscribe to the channel to have more interesting videos and potentially discuss the subejcts. cheers
Good work !!
Thank you! Cheers!
Very very good job, please more content like this !!!
Thanks !! Yes sure !!
Verry good cotinu
Thank you very much 🤗
Congratulations follow this precious work
thanks !! sure !! many new interesting videos are coming :) !!
Thank you. I learned something new today!
My pleasure!! New videos like these are coming !!
Salut ghaith c fait j en ai profite pour l evoyer a qq amis bravo
Parfait thanks !! 🙂
Thank you, very useful information🎉
You’re very welcome!! 😊
yes please we want to know how to creat a youtube chanel , thanks
Yes sure !!
❤❤❤❤
thanks !! :)
Insightful !! thanks for the video !!
thanks !! my pleasure !!
Insightful !! thanks for the video !!
thanks !! my pleasure !!
Insightful !! thanks for the video !!
thanks !! my pleasure !!
Bravo cher frère
thanks !! :)
Good job
thanks !!
Good job
thanks !!
Good job
thanks !!
Good job😊
Thanks !! 😊
Good job thank you 😊😊
Thanks !! my pleasure !!
Good job :)
thanks :)
👏
Thanks :) !! Please don't hesitate to subscribe to my channel for more videos like this !! :) thanks !!
Good job ❣❣❣thank you
Thanks you !! :)
Bonne continuation
Merci beaucoup !! :)
very useful. Thank you!
my pleasure, thanks for the feedback !! :)
Thanks for the video 🙏🏼🙏🏼
my pleasure !! thanks :) !!
Good job 🎉
thanks man !! :)
Very nice , interesting topic. Thanks a lot.
thanks !! :)
Very Good vidéo good luck
thanks !! :)