[Submitted on 22 Jun 2026] · arXiv.org
Recently, end-to-end OCR models, exemplified by DeepSeek OCR, have once again thrust OCR into the spotlight. A widely held view is that employing a large language model (LLM) as the decoder allows the model to leverage the prior distribution of language, leading to improved OCR performance. However, the downside is equally evident: as the output sequence lengthens, the accumulated KV cache drives up memory consumption and progressively slows down generation. This stands in stark contrast to huma
Piekniewski's blog
Introduction I read a lot of deep learning papers, typically a few/week. I’ve read probably several thousands of papers. My general problem with papers in machine learning or deep learning is…
GitHub
The most customizable Wayland compositor with TypeScript(tsx). - bea4dev/ShojiWM
Liquid AI · Liquid AI
Meet LFM2.5-230M, Liquid AI’s smallest model yet: a fast, open-weight foundation model for fine-tuning, edge deployment, tool use, and data extraction.
GitHub
Contribute to googlecolab/google-colab-cli development by creating an account on GitHub.
GitHub
A library of practical AI-agent loops and an installable skill for finding, adapting, and designing repeatable agent workflows. - Forward-Future/loopy
github.com
Source link only
animejs.com
A fast, multipurpose and lightweight JavaScript animation library
huggingface.co
We’re on a journey to advance and democratize artificial intelligence through open source and open science.
GitHub
Contribute to jumpjet-runtime/jumpjet development by creating an account on GitHub.
www.esptoolkit.hu
Source link only
r3f.docs.pmnd.rs
Source link only
Flue
Build and deploy Flue agents on Cloudflare Workers.
knight-l.github.io
Source link only
laminalabs.ai
Drop a document. Get a clear whiteboard explainer in seconds. Built by Lamina Labs.
GitHub
Powerful and easy-to-use embedded GUI with many widgets, advanced visual effects (opacity, antialiasing, animations) and low memory requirements (16K RAM, 64K Flash). - ms-rtos/littlevgl
ugfx.io
uGFX library has been designed to be the smallest, fastest and most advanced embedded library for display and touchscreen solutions
GitHub
CREPE: A Convolutional REpresentation for Pitch Estimation -- pre-trained model (ICASSP 2018) - marl/crepe
GitHub
Train voice styles for Supertone/supertonic-3 model. - saurabhv749/supertonic3-voice-clone
OHF-Voice · GitHub
Fast and local neural text-to-speech engine. Contribute to OHF-Voice/piper1-gpl development by creating an account on GitHub.
GitHub
agent multiplexer that lives in your terminal. Contribute to ogulcancelik/herdr development by creating an account on GitHub.
GitHub
ESP32 Camera motion capture application to record JPEGs to SD card as AVI files and stream to browser as MJPEG. If a microphone is installed then a WAV file is also created. Files can be uploaded v...
www.interfacecraft.dev
A working library for those committed to designing with uncommon care.
docs.ufpr.br
Source link only
spiced-self-play.com
Project website for spiced self-play in Human-like autonomy emerges from self-play and a pinch of human data.