A cheat sheet with everything I’ve learned about FFmpeg github.com/rendi-api/ffmp…

A #kinetic #sculpture, moving with wind but #balanced by #science.

Qwen3.6 27B landed yesterday, so I ran it on 4 setups side-by-side to see how they stack up: 🔴 RTX 4090 — 45.59 tok/s, TTFT 525ms 🟢 RTX 5090 — 51.83 tok/s, TTFT 752ms ⚫️ M2 Ultra — 22.30 tok/s, TTFT 216ms 🟣 DGX Spark — 11.08 tok/s, TTFT 319ms This is a standard test: no

Qwen-Image-2.0-Pro is now live 🚀🚀 We’ve pushed image quality, multilingual text rendering, and instruction following to a new level, while making performance much more consistent across styles.🌅🌃 Ranked #9 worldwide for Text-to-Image on Arena.ai 🔗Try it now on ModelScope:

Check out the official Open Automatic Speech Recognition (ASR) leaderboard on the hub! These list the best speech-to-text (STT) models huggingface.co/datasets/hf-au…

MOSS-Audio：开源统一音频理解模型 4B/8B各有Instruct和Thinking版本，语音识别、说话人分析、情感检测、环境声理解、音乐理解、时间戳ASR全部集成在一个模型里。时间戳ASR精度碾压级领先，远高于Gemini-3.1-Pro。 模型：huggingface.co/collections/Op…

Today we're releasing Nemotron 3 Nano Omni. Audio, Video, Image, Text ➡️ Text Ask questions about all your data. Amazing efficiency powered by the Nemotron Hybrid SSM MoE architecture. State of the art multimodal intelligence.

🚀 Introducing FlashQLA: high-performance linear attention kernels built on TileLang. ⚡ 2–3× forward speedup. 2× backward speedup. 💻 Purpose-built for agentic AI on your personal devices. 💡Key insights: 1. Gate-driven automatic intra-card CP. 2. Hardware-friendly algebraic

Mistral releases Mistral Medium 3.5, a new vision reasoning model. 🔥 Mistral-Medium-3.5-128B offers highly competitive performance for models 6x its size. Run locally on ~64GB RAM. Guide: unsloth.ai/docs/models/mi… GGUFs: huggingface.co/unsloth/Mistra…

Qwen just released their interpretability toolkit on Hugging Face Qwen-Scope adds Sparse Autoencoders to Qwen3.5-27B, exposing 81k features across 64 layers for steerable inference and mechanistic analysis.

Alibaba's Qwen3.6 27B is the new open weights leader under 150B parameters scoring 46 on the Artificial Analysis Intelligence Index, but uses ~3.7x the output tokens and costs ~21x more than Gemma 4 31B (39) to run the full Intelligence Index Qwen has released two open

All three leading open weights models were released last week. Progress continues for open weights models alongside proprietary ones, with the gap to GPT-5.5, the leading proprietary model, sitting at 6 points on the Artificial Analysis Intelligence Index Kimi.ai’s Kimi

⁉️So get this, AMD is making a bold move to own the affordable personal inferencing market by launching a Mini PC in June, a 128GB Shared Memory Inferencing Box 🎇 They call it the ⬭ Halo Box. 🧾 It's a Ryzen AI MAX+ 395 (16 Zen 5 cores + 40 RDNA 3.5 CUs + XDNA 2 NPU) ✅ Up

holy shit. Someone just open-sourced a diagnostic that runs 32 inspections on any AI agent and tells you exactly where it's misaligned. It's called iFixAi. You point it at OpenAI, Anthropic, Gemini, Bedrock, or your own agent. Five minutes later you get a scorecard graded A

Ok, #vibejam is over, and here's a video of my submission, Tron: Light Cycles. Built on Cursor with Three.js. Play it here: tron.ganji.me It was a fun experience overall and I learned a bunch of things. Thanks to @levelsio for organizing it. Costs breakdown: -

honestly Granite 4.1 8b might be the best model to run at this size huggingface.co/blog/ibm-grani…

China has become one of the most active and fastest-growing regions for open-source model development, said a recent report released by the world's largest open-source AI platform Hugging Face, which showed 41% of large language model downloads on the platform over the past year

用 DeepSeek-V4 架构从头训练一个 110M 参数的小模型，方便上手实验 MLA、MoE、Hyper-Connections、MTP 这些新特性。 github.com/huggingface/na… Nanowhale 是 HuggingFace 做的 110M 参数小模型，把 DeepSeek-V4 的架构特性都搬过来了：MLA（8 头 1 KV 头）、MoE（4 路由 + 1