Üretim4
görselvideo·reddit-stablediffusion·11 saat önce
Hi! V-chan here! We got another BIIIG update, and you now have some new toys to play with. New models, transparent sprites, and a Character Creator makeover. We even recruited a video model for sprite duty. Hehe! https://preview.redd.it/fb19npkwtkth1.png?width=4296&format=png&auto=webp&s=c3e33b56d0bf5b83dd2597503df4c5aaec98bff2 If you are new here: VNCCS is a ComfyUI pipeline for creating characters and turning them into sprite sets with different poses, outfits, and expressions. For visual novels, games, or whatever your little creative brain is plotting. Here is the fun stuff in 3.2.0…
model3D·reddit-stablediffusion·az önce
During VFX Week last week, we released seven open-weight capabilities for LTX-2.5, covering high-res editing, restoration, HDR, CG-guided generation and compositing: Native Resolution : Run AI edits on 4K and 8K plates and beyond without downscaling. Overlapping tiles keep fine detail, on a single GPU. Refine : Refine generates the fine detail a standard resize leaves soft, without drifting from the source. Restore : Takes archival footage from as low as 540p to 4K, enhances color and gives it a modern digital-camera look. SDR to HDR : Rebuilds shadow and highlight detail in SDR footage and…
görsel·hf-papers·5 gün önce·51 oy
World action models (WAMs) have emerged as a promising paradigm for robotic control by jointly predicting future visual dynamics and actions from an initial observation and instruction. However, existing WAMs struggle with long-horizon prediction, as generating dense video rollouts is highly inefficient. Some recent WAMs address this by predicting a single future frame without generating the full video, but this approach neglects how to progress toward the goal. We present ProWAM, a progressive world action model that jointly predicts actions and an ordered sequence of sparse visual…
video·hf-papers·6 gün önce·71 oy
Long-horizon video generation requires models to effectively leverage an increasingly long generation history. As the generated history grows, retaining all previous content becomes increasingly expensive and redundant, making effective historical selection essential. Existing approaches often determine historical relevance based on the current content. However, information relevant to the present is not necessarily useful for future generation, while seemingly less relevant history may become important later. Our key insight is that historical information should be selected according to its…
Modeller & Araştırma4
↑ yükselişte · yeniden15 oy/saataraçmodelşirket·simonwillison·2 gün önce·615 oy·+1 kaynak
Here's a product feature which the world is going to need a whole lot more of over the coming months and years: default hard budget caps . I'm talking about the feature of pay-by-usage services and APIs that lets you say "after $X/month, cut this thing off and return errors". These need to be hard limits. Soft caps, "after $X/month, send me a warning email", will not cut it. Coding agents, and personal agents (coding agents wrapped in a less threatening UI), greatly reduce the friction of spinning up code that can do useful things. Sometimes those things cost money - calls to paid APIs, or…
↑ yükselişte · yenidenHN'de GPT-6.1 0→8 hikâyemodel·openai·6 gün önce
Meet GPT-6.1 Sol: near-Astra intelligence for coding, computer use, and professional work at one-fifth of Astra’s standard API input and output token prices.
↑ yükselişte33 oy/saat · HF trend 132araçmodelşirket·hackernews·1 gün önce·865 oy
↑ yükselişteHN'de GPT-6.1 0→8 hikâyemodel·marktechpost·18 saat önce
Astra leads computer use, Argon leads legal and finance work, and Sol wins on price for coding agents. The post GPT-6 Astra vs GPT-6.1 Sol vs Gemini 4 Argon vs Claude Fable 5.1: Which Frontier Model Fits Which Job appeared first on MarkTechPost .