I just created my own LaTeX-OCR app using Llama... 3.2 Vision! Upload the LaTeX code as an image, and it gives you the corresponding LaTeX code using Llama 3.2 multimodal! Here's what I used: - Ollama for serving Llama 3.2 vision locally - Streamlit for the UI Everything is just 50 lines of code! Find the code in the next tweet. -- Every day, I share tutorials and insights on DS, ML, LLMs, and RAGs. Find me → Avi Chawlashow more

Advent of Claude Day 31 - Claude Agent SDK... I saved the best for last. The same agent loop, tools, and context management that power Claude Code is available as an SDK. Build agents that work like Claude Code in as little as 10 lines of code. This is just the beginning. 🎇🥂show more

Introducing Dexter. An open source financial agent in ~200... lines of code. Think: Claude Code, but for finance. What Dexter does today: • plans tasks • runs on its own • validates its work • researches stocks It’s built with OSS tools like for managing LLMs, prompts, outputs, etc. No coding experience needed to run. Goal: build the best financial agent and share every line of code along the way. Next: evals, starting withshow more

I just customized the scrollbar on my latest Framer... project using Chat GPT. It took me 3 minutes and I know less about code than a drywall. Totally transformed the feel of my site. Here's a step-by-step of how I did it 👇🏼show more

Here's the Rive source file and the code behind... it. The Rive file is available on the Rive Rive Marketplace, and I'll also share the code below. Apologies in advance for the messy file structure and naming conventions. At some point I stopped naming things properly and chose chaos instead.show more

it's day 6 of sharing code + tutorials for... my computer vision projects "fun with computer vision dot com" 🪬 thank you to everyone for the incredible support on my launch we're up to 126 customers now next steps: > "Floor is Lava" code + tutorial will be released this week > making more video content and cross-sharing on yt / tt / ig (thank you jack friks for the inspiration to post widely and post often) > creating more structure and engagement in our discord servershow more

here's how the whole thing works. claude code doesn't... care what's behind the API. it just sends requests and expects responses. so i pointed it at my own machine instead of anthropic's servers. llama-server runs the model locally. LiteLLM sits in between and translates the API format. claude code thinks it's talking to claude. it's talking to qwen on localhost. the setup: 2x 3090s, 38 layers on GPU, 10 on CPU. 128K context window. generation is only 7 tok/s but the tradeoff is worth it. 128K means the agent can hold an entire project in memory without losing context midtask. claude code alone loads a 17.5K token system prompt on every request. tool definitions, safety rules, agent behavior. that's your baseline before you even say hello. pushed as far as i could tonight. what surprised me most wasn't the speed. it was the iteration quality. first prompt gave me a working particle sim. second prompt, the model read its own 564 lines, understood the architecture, and added trails, explosions, gravity wells, bloom effects. no handholding. 4bit quantized. 45GB on two consumer cards. running a full coding agent autonomously. detailed article coming. full benchmarks, hardware breakdowns, engine debugging, code quality. everything from setup to what broke and why.show more

Fine-tune DeepSeek-OCR on your own language! (100% local) DeepSeek-OCR... is a 3B-parameter vision model that achieves 97% precision while using 10× fewer vision tokens than text-based LLMs. It handles tables, papers, and handwriting without killing your GPU or budget. Why it matters: Most vision models treat documents as massive sequences of tokens, making long-context processing expensive and slow. DeepSeek-OCR uses context optical compression to convert 2D layouts into vision tokens, enabling efficient processing of complex documents. The best part? You can easily fine-tune it for your specific use case on a single GPU. I used Unsloth to run this experiment on Persian text and saw an 88.26% improvement in character error rate. ↳ Base model: 149% character error rate (CER) ↳ Fine-tuned model: 60% CER (57% more accurate) ↳ Training time: 60 steps on a single GPU Persian was just the test case. You can swap in your own dataset for any language, document type, or specific domain you're working with. I've shared the complete guide in the next tweet - all the code, notebooks, and environment setup ready to run with a single click. Everything is 100% open-source!show more

You used to see these on the wall of... every computer science lab, but I'd never written one... so I wrote one this morning! It's under 200 lines, code in the comments. And if you like these littlle code excursions, follow me, and I'll keep doing them!show more

Next generation flight tracking app has just dropped. It... has, ADSB, ATC, ACARS, CODLC, routes, navigation charts, emergency squawk code alerts, and much more! I love this app It’s called FlightDeck. The FlightDeck team has a special promo code for all my followers: FLIGHTDECK51 There is a free three day trial and it’s only available right now for the iPhone. ✈️🤓show more

If you want new income streams, read this. 💰👀.... Mom life leaves almost no time to make extra $$$… but I built a whole mobile app during nap time 😭💜 I just typed my idea and the AI built the entire app for me, NO CODE! If you want a side hustle, comment “build” and I’ll share what I used.show more

I finally got to shoot my Latex Misty Cosplay... by water and it makes me so happy! 🥰 Please tell me you checked out the full video, and I didn't just make it for no reason?! 👀⬇️ Latex sent to me to shoot in by @catalystlatexshow more

Just tried something cool on my own app. I... let my app generate a workout for me using the new beta mode… and it actually built a solid session based on my goals. The idea is simple.. instead of guessing what to train, you hit generate and the app builds the workout for you. Still early days but I’m keen for people to test it and tell me what you think. Try it here:show more

A few people have asked how I make these... animations. The answer is it's just code and a screen recording. I’ll make little jigs like this interactive timeline to dial things in, but for me I find this is still the fastest way to create and refine exactly what I want.show more

My project has 39,205 lines of code, and Cursor... can't answer questions about it. Cursor's context seems to be capped at around 10,000 tokens. Unfortunately, this is not enough for any decent-sized project. If you have a large codebase, check out Augment Code. This thing is faaaast! I'm currently using their Visual Studio Code plugin, but you can also use them on JetBrains, Neovim, and even Vim. (I'm a Neovim fan, but Copilot's implementation for Neovim is nowhere as good as Augment Code.) Augment Code was gracious enough to sponsor this post. After you install their extension and run it for the first time, it will index your entire codebase. This is why it can answer questions as fast as it does, regardless of the size of your codebase. Augment Code supports chat and completions like every other AI coding assistant, but its killer feature is "Next Edit." When you make a change, two things happen: 1. The model analyzes the change to determine the ripple effects across your *entire* codebase. 2. The model suggests everything you need to update to ensure everything works correctly. This is pretty wild!show more

The future of AI in web development? This app... generates code and images to convert a screenshot into HTML/Tailwind CSS. I just gave it a screenshot of my X feed, and look what it produced (Notes & link in the comments):show more

If you are using an if statement inside a... for loop, did you know there's a better way of doing this is Swift? I'm talking about the where clause. Using the where clause in a for loop can provide several benefits over using an if statement inside a for loop: Readability: It explicitly states the condition that must be met for an element to be processed, making the purpose of the loop clearer to anyone reading the code. Efficiency: When using a where clause, the filtering of elements is done at the level of the sequence, which can be more efficient than filtering within the loop with an if statement. Brevity: Using a where clause often results in more concise code, as it eliminates the need for an if statement and associated braces. This can make the code more manageable and easier to maintain.show more

Claude Code can now go find the right skill... itself, instead of you searching for one. it's called find-skills. a small package that plugs into Claude Code, and instead of you hunting for rest, you just describe the task and it searches the whole open skills ecosystem, finds the ones that fit, and installs them for you. > tell it what you're trying to do, in plain english > it scans the skills registry and maps your task to real skills > it pulls the right ones in and sets them up half the time you don't even know a skill exists for what you're doing. now you don't have to.show more

Update: reached a milestone of 10 tutorials published on... "fun with computer vision dot com" 🪬 - 185 customers - $3.9k revenue - >1 million views it's been 21 days since I launched this project, and I'm grateful for your incredible support and kindness 💗 thank you to Three.js and mediapipe for the magic behind these computer vision / augmented reality experiments this project has been incredibly fun and I have lots more I want to create more experiments - demos - tutorials - code, coming soon to a tweet near you :]show more