2026-10-19

Looks quite interesting. Lots of numbers presented, but for a single model only. Highlights: File sizes are the same as Q-quants. KLD improves substantially over Q-quants. Performance suffers slightly Creating IQ quants is somewhat slow. Asterisks apply... Appologies to Mr. Addario if I mangled anything.

Some of y'all may be familiar with AutoTrust from Hugging Face. Many of their models were trending on the front page of HF and they claimed some very impressive results. But if you looked at the org's models and clicked into likes, you may notice something a bit off: These accounts sure do look awfully realistic, no? Well let's take a look at a few of these accounts (sorry for my very poor graphic design skills): Across all these accounts, we can see a pattern: a fake, realistic name seemingly f

It's an extremely lightweight speech recognition model. 949K parameters using a weight-shared Conformer-CTC architecture Just 1.06 MB for INT8 weights Around 436× real-time on a single CPU thread 11.43% WER on LibriSpeech test-clean, using greedy decoding with no external language model Apache 2.0 licensed Any feedback, question or criticism is welcome!

Just something I've been testing over few week since Embedding Gemma 2 released which is about 330MB at Q8. Not sure if this would count as 'resource' flair... Image shows cosine similarity changes as I refine the text query (aka text embedding vector) over time. Shows how well it work & quirks it has at same time. I'll leave tl;dr version for now, 3:30 am here. If anyone is interested I'll do few more tests and make example query image like above - still have 7 more sets totaling 22448 screensh

Doska is a kanban board web and desktop app that supports offline file storage (folders with .md files) or self-hosting a sync server. When there’s a sync server, collaboration features are also supported. The self-hosted setup includes versioned Docker containers for the web server and the sync server. It looks good on mobile (PWA), and I’ve put a lot of work into making it fast, responsive and pleasant to use. There’s no login gate or any other kind of gate: it’s fully MIT open source. GitHub:

Video world models turn actions and camera motion into long visual rollouts, making video a natural interface for interactive simulation. Such interaction requires responsive generation, yet coherent rollouts must retain visual memory of what has already been generated. Chunked video diffusion models usually keep this context in growing KV caches, so later chunks become progressively slower. To address this bottleneck, we introduce Velox, a training-free attention router for region-level context

Inline rendering with document​.currentScript Lean Rada describes a neat trick that takes advantage of the document.currentScript property. You can use that to drop in a tiny inline element that replaces itself when the page loads: The year is document.currentScript.replaceWith( new Date().getFullYear()) If you define a shortcut function like this: $=(...n)=>document.currentScript.replaceWith(...n) You get a tiny template language: The year is $(new Date().getFullYear()) Tags: javascript

While generalist vision-language-action (VLA) models have enabled diverse open-world manipulation, real-world robotic assistance requires two critical capabilities they currently lack: personalization and reliable execution. First, executing an instruction like "make me a sandwich" requires grounding subtask decomposition in persistent user preferences rather than relying on a single generic solution. Second, reliable deployment demands explainable safety assurance that proactively prevents haza

Generative simulation can automate robot environment construction, but physically valid environments are not necessarily useful for training. The challenge is to identify environments that address a policy's failures and supervision that helps it overcome them. We introduce CoDA, a framework in which a simulation data engine and an action policy co-evolve through a shared heterogeneous memory of environment construction, policy failures, and learning progress. The memory compiles validated diagn

I use Portainer across my whole stack but I’ve been looking to switch. Any ideas on what people recommend I was looking into Komodo but looking to hear some opinions.