John Mai
@JohnMai_Dev
🚀Building ChatMLX / Go ORM Helper / Jinja Swift | 💻 AI ➢ LLM ➣ MLX ➢ Full Stack
New research from Exo done (in part) with MLX on Apple silicon: An algorithm for distributed training that leverages higher RAM capacity of Apple silicon relative to FLOPs and inter-machine bandwidth.
Paper is out. Link: openreview.net/pdf?id=TJjP8d5…
Hey friends, do you know any local AI apps that can run these models? 😁
Big MLX Swift LLMs release: - New models: Gemma 3, BitNet - Streamlined API - Structured tool use - Other bug fixes and QoL improvements
Quickly fixing the SmolLM3 issue by @DePasqualeOrg 🚀🚀🚀 github.com/johnmai-dev/Ji…
Being handled here by @JohnMai_Dev github.com/johnmai-dev/Ji…
Latest mlx-lm is out! pip install -U mlx-lm Bunch of new models: - SmolLM3 (Hugging Face) - Ernie family (Baidu) - BitNet (Microsoft) - Falcon-E (TII) - Text-only Gemma3n (Google) - MiniCPM4 (OpenBMB) - AFM (Apple) +Performance improvements for DWQ, dynamic quantization, and…
MLX-Swift Transpilation Building a side-by-side dataset of MLX Python -> MLX Swift models. Useful for prompting when converting models to Swift, and for fine-tuning specialized MLX coding models.
MLX Swift BitNet is ready! Thanks to MLX King 👑 @Prince_Canuma github.com/ml-explore/mlx…
MLX Ernie4.5 is ready! github.com/ml-explore/mlx… mlx_lm.generate --model mlx-community/ERNIE-4.5-0.3B-PT-bf16 --prompt "Hello! How are you?"
The ERNIE 4.5 series is now officially open source. This family of models includes 10 variants—from MoE models with 47B and 3B active parameters, the largest having 424B total parameters, to a 0.3B dense model—all available now to the global AI community for open research and…
If the rumors are true and the MLX team is considering leaving, it would mark one of Apple’s worst strategic missteps at a critical moment. Killing MLX wouldn’t just derail Apple’s AI trajectory—it would undercut the entire Studio line and the M-series chips. @tim_cook, nurture…
MLX ERNIE 4.5 coming~
Baidu kept its promise, releasing 10 open models on the very last day of June🚀 Let's meet ERNIE 4.5 🔥 huggingface.co/collections/ba… ✨ From 0.3B to 424B total params ✨ Includes 47B & 3B active param MoE models + a 0.3B dense model ✨ Fully open under Apache 2.0 ✨ 128K context…
Yay this is VLo! Still a Preview version but so interesting! Just edit images as you wish and funny things may happen! Ghibli style, adding a cap, removing stuffs,... Anyway, it is our first step towards a unified image understanding and generation. Give it a shot!
Meet Qwen-VLo, your AI creative engine: • Concept-to-Polish: Turn rough sketches or text prompts into high-res visuals • On-the-Fly Edits: Refine product shots, adjust layouts or styles with simple commands • Global-Ready: Generate image in multiple languages • Progressive…