AdaLLM is an NVFP4-first inference runtime for Ada Lovelace (RTX 4090) with FP8 KV cache and custom decode kernels. This repo targets NVFP4 weights and keeps the entire decode path in FP8
☆155Feb 15, 2026Updated 8 months ago
Alternatives and similar repositories for NVFP4-on-4090-vLLM
Users that are interested in NVFP4-on-4090-vLLM are comparing it to the libraries listed below. We may earn a commission when you buy through links labeled 'Ad' on this page.
Sorting:
- asmsense: a small model that names and describes functions in stripped x86-64 binaries. Open dataset, baselines, and LoRA experiments.☆17Oct 18, 2026Updated 2 weeks ago
- writing really fast kernels☆21Updated this week
- ☆20Apr 16, 2025Updated last year
- Fast & lightweight interface for running Nano Banana LLM and compact local models on your PC. Optimized for high-speed AI interactions.☆44Mar 17, 2026Updated 7 months ago
- ☆13Sep 9, 2026Updated last month
- Deploy to Railway using AI coding agents - Free Credits Offer • AdUse Claude Code, Codex, OpenCode, and more. Autonomous software development now has the infrastructure to match with Railway.
- Qwen4-Exp (Qwen3.8-Flash-Next) SWA + MTP + PLE-on-iswa — bounded deep-context decode w/ long-range recall, TBQ4 4.25 bpv KV, DSpark + Rot…☆95Sep 24, 2026Updated last month
- Qwen3.8-27B-GSQ3 on one RTX 4080: 100k context with rk4v4-e8, up to 2720 tok/s prefill and 262 tok/s generation☆33Oct 9, 2026Updated 3 weeks ago
- A GUI tool to overclock NVIDIA GPUs on Wayland and X11. Also supports fan curves,oc profiles, and adjusting power limits.☆85Updated this week
- llama.cpp fork with additional SOTA quants and improved performance☆22Oct 21, 2026Updated last week
- SonicBoom is a compiler-first PyTorch alternative for language, vision, audio, and training. It builds on proven kernels and pushes perfo…☆32Updated this week
- Python app created with the purpose of speeding up and greatly facilitating the task of cleaning and adjusting Booru-style tags, aimed at…☆12Dec 2, 2023Updated 2 years ago
- Thinking blocks that open during generation: a gate decides when to think, a tree of candidate thoughts is explored, and the best one sta…☆18Oct 11, 2026Updated 3 weeks ago
- Classic PC demoscene productions running in the browser☆23Oct 8, 2026Updated 3 weeks ago
- A small A/B test of a laconic brevity directive on three LLMs: 54 replies to safety decisions, blinded two-scorer consensus against pre-r…☆17Aug 10, 2026Updated 2 months ago
- Simple, predictable pricing with DigitalOcean hosting • AdAlways know what you'll pay with monthly caps and flat pricing. Enterprise-grade infrastructure trusted by 600k+ customers.
- 100.000 links, 50.000 artworks dataset. Includes source code that used to scrape data.☆10May 29, 2021Updated 5 years ago
- ☆20Apr 26, 2026Updated 6 months ago
- recipe for running Qwen3.8-Flash-Next on a single DGX Spark (fork)☆51Oct 22, 2026Updated last week
- ☆20Aug 19, 2024Updated 2 years ago
- A pure-Python implementation of the Nvidia CuTe layout algebra intended to be approachable and easy to learn.☆257Jun 29, 2026Updated 4 months ago
- A high performance authentication and access-control gateway for LLM API backends☆17Jun 11, 2026Updated 4 months ago
- Source-available C++/CUDA inference engine for Qwen3.8-27B INT4 on 3x RTX 3090☆36Updated this week
- ☆13Oct 9, 2024Updated 2 years ago
- A Kubernetes controller and webhook implementation that enables safe, staged rollouts of DaemonSets☆16Jul 31, 2025Updated last year
- Deploy on Railway without the complexity - Free Credits Offer • AdConnect your repo and Railway handles the rest with instant previews. Quickly provision container image services, databases, and storage volumes.
- CLI secret management☆16Sep 24, 2026Updated last month
- How much of your code is written by agents?☆24Apr 24, 2026Updated 6 months ago
- ☆18Jul 1, 2025Updated last year
- ☆17Jun 18, 2026Updated 4 months ago
- 🎓The path to LLM mastery is paved with broken embeddings and resurrected gradients. (CHINESE NOTES)☆13Jul 16, 2026Updated 3 months ago
- Research: vGPU unlock on consumer NVIDIA RTX 5090 (Blackwell/GB202). 19 binary patches, full CPU-side pipeline working, GSP firmware bloc…☆34Apr 1, 2026Updated 7 months ago
- ☆64Jul 21, 2024Updated 2 years ago
- bandwidth limiting middleware plugin for Traefik that provides fine-grained control over data transfer rates. This plugin supports per-ba…☆16Oct 22, 2026Updated last week
- a minimal paged attention implementation☆20Jan 30, 2026Updated 9 months ago
- Managed hosting for WordPress and PHP on Cloudways • AdManaged hosting for WordPress, Magento, Laravel, or PHP apps, on multiple cloud providers. Deploy in minutes on Cloudways by DigitalOcean.
- learn mojo☆10May 6, 2026Updated 5 months ago
- Fun playground and adventure game☆13Updated this week
- Code for the papers: “Four Over Six: More Accurate NVFP4 Quantization with Adaptive Block Scaling” and “Adaptive Block-Scaled Data Types”☆211Apr 21, 2026Updated 6 months ago
- ☆12Mar 18, 2025Updated last year
- ☆10May 1, 2024Updated 2 years ago
- Tensor-Core selector unlock for RTX 4070 Ti Super☆29Oct 20, 2026Updated 2 weeks ago
- ☆14May 4, 2026Updated 5 months ago