Image created with OpenAI GPT-Image-1. Image prompt: Cheesy late-night infomercial freeze-frame—oriental scroll unfurling with dragon-bot brand “QWEN QUICK-WISDOM™”; red lantern glow, scan jitter, high-resolution
Running mlx-lm locally with MCP using @huggingface’s tiny-agents is actually pretty easy and works quite well. A demo running Qwen3 4B with an MCP client for the local file system: https://x.com/awnihannun/status/1931755333011349831
RT @menloresearch: Meet Jan-nano, a 4B model that outscores DeepSeek-v3-671B using MCP. It’s built on Qwen3-4B with DAPO fine-tuning, it h…”” / X https://x.com/mervenoyann/status/1934909412117741601
stop using VLMs blindly ✋🏻 compare different VLM outputs on a huge variety of inputs (from reasoning to OCR!) 🔥 > has support for multiple VLMs: Gemma 3, Qwen2.5VL, Llama4 > recommend us new models or inputs, we’ll add 🫡 https://x.com/mervenoyann/status/1935708014645784713
🚀 Excited to launch Qwen3 models in MLX format today! Now available in 4 quantization levels: 4bit, 6bit, 8bit, and BF16 — Optimized for MLX framework. 👉 Try it now! Huggingface： https://x.com/Alibaba_Qwen/status/1934517774635991412
Upgraded from Llama 3 to Qwen3 as my go-to model for research experiments, so I implemented qwen3 from scratch: https://x.com/rasbt/status/1936041873099063333
