Відео

Glad you liked it!

I'm glad you found it to be helpful.

Sorry to hear you are having issues with this. I usually self host all my models so haven't faced this. But I bet this is super annoying. Do let me know what you finally end up using.

I swear. Hoping and praying that the PR gets merged soon!

but good to show result_validators

That's a great idea. You should be able to use that for such use cases. The goal was to highlight all the important PydanticAI features I love.

I'm sorry to hear that. Do make sure your .env is configured the right way and that the server is running. Inferix does require ollama to be running in the backend

Awww... I really appreciate your kind words. I'm glad you found the video to be helpful!

Go for it!

Absolutely, Langgraph is coming in next.

@YourTechBudCodes 😍😍😍

I'm sorry to hear that. But I'm glad I could help.

Ayi

I'm glad I could help!

Holy smokes! Why didn't I think of that? Lol!

A friend recently suggested ell as well. I've got to check it out too. My first impression was that it's super minimalistic and doesn't force too many opinions/conventions down your throat. I really like that. But I think I have fallen love with Pydantic's request_validator, dependency injection and the ModelRetry exception. I know. It's just two tiny tiny features, but they are so handy!!! Mahn there is so much going on. Such exciting times!

I'll come clean for this one, I haven't checkout CrewAI in a really long time so I won't be able to make a fair comparison. But, I do think it's important to have options. Specially when each option comes with it's own opinion. The key is to find the option which resonates with you. I think that's crew ai for you. But how about this? Let me try crewai out and come up with a stream comparing crewai, langgraph and pydantic. Maybe we can throw autogen in there as well?

Yes I would like to see this comparison

Absolutely. Have added it to my discord!

Sorry about the video quality. I forgot to hit the record button. Still new to this whole streaming business and am trying to figure stuff out. > Why not use langchain? Uhm, I didn't mean to say that. I still use the basic utility functions langchain gives us. It is the best ecosystem out there. This probably got cut out in post. But totally agree with your thoughts there. > kudos, for the Langgraph shout out Oh absolutely. Langgraph is still my favorite. PydanticAI doesn't handle orchestration that well, while Langgraph has no opinion on what you put inside it's nodes. I should probably plan a stream on the differences and how you can use these together. Would love to see you on my streams, I have a feeling you can help me improve the content for everyone!

@@YourTechBudCodes appreciate your response. I'm a dev/enterpreneur building while learning these new paradigm shifts. It can be overwhelmin at times so it's crucial to work together. Ill keep an eye out for your stream cheers!

This makes me happy! Thanks a ton for the kind words, bud!

pgvector for the win!

Glad you found it helpful

Im so sorry to hear that. What problem would that be?

Welcome to capitalism!

So, uv does support the use of custom registries. Ref: github.com/astral-sh/uv/issues/6421 I don't see how things would slow down when you are using a private registry. Sure, maybe the download speeds wouldn't be as great as using a CDN, but uv's speed gains are not entirely because of the use of a custom registry. Most of the time, you would be using packages from a filesystem cache bypassing the registry altogether. And, there are several more optimizations which make uv so awesome. Would highly recommend checking out Charlie's video (link in the description) to know the deets.

@YourTechBudCodes Thanks for the response! I'll definitely look into it and explore the possibility of moving from pip to UV-it’s a significant leap. A lot of things have been streamlined too, which is exciting. I was actually introduced to UV by @arjanacodes, and it's been a game-changer ever since.

Hey, that ain't no reason for forgetting how to spell Pikachu!

Supposed to be... Am I that bad? Ouch!

Haha. Priotities!!!

oh yeah! I have something cool in mind for this, but it will probably take a lil time.

I'm not sure if I'm supposed to take that as a compliment. But thanks.... I guess?

Noted!

Haha. That shall remain a mystery!

Glad you enjoyed it

I'm glad you found it helpful.

Yeah. I've been thinking about doing something on that. Is there any specific use case you are trying to achieve?

Priorities!!!

Sure. Once I reach 50k subscribers. Lol

Thanks for your kind words. I'm glad you like these videos.

@@YourTechBudCodes thanks for your reply..... Can we expect a live youtube video on building a good microsservice arch... Project 🙂🙂

Mahn thats a tough one. Anything specific you are looking for?

@@YourTechBudCodes understood that.... iam expecting a good system design of one project in micro service arch with better solution from your side 🙂

It's a childhood dream of mine. Just can't let go of it.

I'm afraid that is not possible... sorry about that

Haha. I'm glad my cringe-worthy jokes haven't gotten to you yet.

​@@YourTechBudCodes Give it time! It would be amazing to see code walkthrough videos as well though!

Deeplearning.ai has some short courses on Autogen and some other tools you might find helpful. But things are evolving too fast to rely on a bunch of courses. You gotta put on the hat of a researcher. Always remember, these are just tools to help solve a problem. Get into the art of solving problems using the tools available at your disposal. That's what will help grow your career. All the best, buddy!

Wow. Really appreciate the kind words. This helps me keep going. Thanks!

I'm glad you found it helpful

Glad you liked it.

Haha. Glad you liked it. I just posted a part two last week. Do check it out and let me know your thoughts.

@@YourTechBudCodes thank you! I will check it out. I realized that we need our own Open AI key, may I ask why do we need it if we are running our own inference server and open source model?

The openai sdk is annoying. it forces you to provide one. Just put a dummy key and you'll be fine.