LensVLM-9B by Apple(huggingface.co)
21 points by nthypes | 0 comments
Hugging Face Papers page for an Apple research paper describing LensVLM-9B, a vision-language model.
#vision-language-models on Hacker News
14 stories tagged vision-language-models, newest first.
DeepSeek-v4-flash-vision-exp(api-docs.deepseek.com)
470 points by dares2573 | 144 comments
DeepSeek API documentation page describing the deepseek-v4-flash-vision-exp model, which accepts images alongside text, and how to send images via the API.
Unlimited OCR: One-shot long-horizon parsing(github.com)
447 points by ingve | 102 comments
GitHub repository for Baidu's Unlimited-OCR, an MIT-licensed OCR model for one-shot long-horizon document parsing.
Karpathy on DeepSeek-OCR paper: Are pixels better inputs to LLMs than text?(twitter.com)
381 points by JnBrymn | 150 comments
An X post by Andrej Karpathy discussing the DeepSeek-OCR paper and whether pixels might be better inputs to LLMs than text tokens.
DeepSeek OCR(github.com)
939 points by pierre | 226 comments
GitHub repository for DeepSeek-OCR, DeepSeek's open-source "Contexts Optical Compression" OCR model, with code, paper PDF, and MIT license.
Show HN: Dayflow – A git log for your day(github.com)
400 points by jerryliu12 | 115 comments
GitHub repository for Dayflow, an MIT-licensed macOS app that builds an automatic timeline of your day by analyzing screen content with local or cloud VLMs.
Qwen3-VL(qwen.ai)
411 points by natrys | 134 comments
Qwen team blog post announcing Qwen3-VL, a vision-language model series with improved visual perception, reasoning, and agentic capabilities.
Nanonets-OCR-s – OCR model that transforms documents into structured markdown(huggingface.co)
326 points by PixelPanda | 72 comments
Hugging Face model card for Nanonets-OCR-s, a 3.75B-parameter image-to-markdown OCR model built on Qwen2.5-VL that converts documents into structured markdown.
FastVLM: Efficient vision encoding for vision language models(github.com)
335 points by nhod | 68 comments
Apple's GitHub repository with the official implementation of FastVLM, an efficient vision encoding method for vision language models presented at CVPR 2025.
Qwen2.5-VL-32B: Smarter and Lighter(qwenlm.github.io)
522 points by tosh | 280 comments
Qwen Team blog post announcing Qwen2.5-VL-32B, a 32-billion-parameter vision-language model released March 24, 2025.
A return to hand-written notes by learning to read and write(research.google)
557 points by mfiguiere | 174 comments
Google Research blog post describing a model that converts photos of handwritten notes into digital ink by learning to read and write stroke sequences.
Vision language models are blind(vlmsareblind.github.io)
425 points by taesiri | 172 comments
Project page for the research paper "Vision language models are blind" from Auburn University and University of Alberta, with links to the arXiv paper, code, and dataset.
Llama 3-V: Matching GPT4-V with a 100x smaller model and 500 dollars(aksh-garg.medium.com)
426 points by minimaxir | 67 comments
Medium post by Aksh Garg describing Llama 3-V, an open-source multimodal model claimed to match GPT-4V while being 100x smaller and trained for under $500.
MiniGPT-4(minigpt-4.github.io)
812 points by GaggiX | 256 comments
Project page for MiniGPT-4, a KAUST research model enhancing vision-language understanding with advanced LLMs, with links to paper, code, and video.