Abed Khooli

Here is a 2nd batch of April architecture drops. What a month! - Ant Ling 2.6 1T - Minimax M2.7 - Xiaomi MiMo V2.5 - Poolside Laguna XS.2 - Tencent Hy3-preview - IBM Granite 4.1

6 Open-Source Libraries to FineTune LLMs 1. Unsloth GitHub: github.com/unslothai/unsl… → Fastest way to fine-tune LLMs locally → Optimized for low VRAM (even laptops) → Plug-and-play with Hugging Face models 2. Axolotl GitHub: github.com/OpenAccess-AI-… → Flexible LLM fine-tuning configs → Supports LoRA, QLoRA, multi-GPU → Great for custom training pipelines 3. TRL (Transformer Reinforcement Learning) GitHub: github.com/huggingface/trl → RLHF, DPO, PPO for LLM alignment → Built on Hugging Face ecosystem → Essential for post-training optimization 4. DeepSpeed GitHub: github.com/microsoft/Deep… → Train massive models efficiently → Memory + speed optimization → Industry standard for scaling 5. LLaMA-Factory GitHub: github.com/hiyouga/LLaMA-… → All-in-one fine-tuning UI + CLI → Supports multiple models (LLaMA, Qwen, etc.) → Beginner-friendly + powerful 6. PEFT GitHub: github.com/huggingface/pe… → Fine-tune with minimal compute → LoRA, adapters, prefix tuning → Best for cost-efficient training Save this for future use.

shoutout to the man @kylejeong for cooking this up 🔥 every agent can now get a dedicated browser subagent in seconds to navigate sites, fill out forms, click through things, you get it Deep Agents x Browserbase goes hard

Ecosystem-first approach continued! Ling-2.6-1T officially landed on @huggingface and the official inference is now live via @novita_labs. Experience the efficiency of Ling-2.6-1T for yourself, front and center on HF model card page! 🔥

@bclavie Is this the start of edge computing era (less reliance on cloud computing) ?

Tomorrow I’m getting 10gbps FTTH installed. Please suggest things I can do as a bandwidth-rich person

Second Call For Papers is out! - Abstract submission deadline: June 18th, 2026 - Full paper submission deadline: June 25th, 2026 Visit our website for more information on submission instructions: arabicnlp2026.sigarab.org

Update: it's out. 🔥 OpenAI's privacy-filter retrained on @nvidia's Nemotron-PII data. 8 → 50+ labels. Healthcare + enterprise + demographic + biometric coverage. Apache 2.0. 8-bit MLX version coming this week. huggingface.co/collections/Op…

The data you need to benchmark enterprise RAG is the data no one shares. So we built it. Today we're releasing EDiTh on @huggingface , an open benchmark for enterprise retrieval, designed around the questions executives actually ask. What's inside: 📄 1,004 PDFs across 6 languages and 3 formats 🎯 36 use cases with full answer keys 🎭5.5% plausible distractors (the kind that fool real systems) 🏢 All built around Véracier Industries, a fictional €1.8B industrial group Each use case mirrors a real executive question, the kind that usually takes a team, a week, and a stack of PDFs to answer. If you're building, buying, or evaluating RAG for the enterprise, EDiTh gives you something the public benchmarks don't: realistic documents, realistic noise, realistic stakes. Built by Adèle Guignochau and @IgorCarron at @LightOnIO . Read the release 👉🏻 lighton.ai/lighton-blogs/… Dataset on Hugging Face 👉🏻 huggingface.co/datasets/light… #EnterpriseAI #RAG #SovereignAI 📄

مستقبل الذكاء الاصطناعي: الفجوات تتقلص بسرعة عالية linkedin.com/posts/akhooli_…

April was a pretty strong month for LLM releases: - Gemma 4 - GLM-5.1 - Qwen3.6 - Kimi K2.6 - DeepSeek V4 All are now added to the LLM Architecture Gallery. More details once I am fully back in May!

The Future of AI: Gaps are Shrinking and Shrinking Fast linkedin.com/posts/akhooli_…

ChatGPT image 2 is good with Arabic - dummy test:

Just out: A new fact sheet and an updated map on access restrictions affecting 3.4 million Palestinians in the #WestBank – 925 obstacles as of December 2025, the highest number on a 20-year record. Learn more: ochaopt.org/content/moveme…

OpenAI Privacy Filter in Arabic (not perfect): Demo: huggingface.co/spaces/akhaliq… HF Model: huggingface.co/openai/privacy…

The mystery is solved! The model previously known as “Elephant” on @OpenRouter is now officially revealed as Ling-2.6-flash 🐘⚡ Powered by Novita, it's already: → Top 8 on OpenRouter by total calls this week → Dominated the trending list during its anonymous run And now it gets even bigger: 👉 Novita is the exclusive hosting provider 👉 1 week FREE API access for all devs Why everyone is switching: • Up to 340 tokens/s (peak, 4× H20) • ~10× more token-efficient (on AA benchmarks) • Built for real agent workloads (not only benchmarks) This is what happens when you optimize for execution > verbosity. Huge shoutout to @AntLingAGI for pushing the frontier.

Lucky me! Zero Meta forever.

new open-source Bonsai models are out 🔥 > ternary weights in 8B (1.75 GB), 4B (0.86 GB), and 1.7B (0.37 GB) > comes in MLX, ONNX weights and WebGPU browser demo 😍 > a2.0 licensed 👏

if you're using Replit, Antigravity or other vibe-building tools 👋🏻 simply adding @huggingface Skills to your setup gives your agent access to ~3M open models, 500k+ local AI apps and ~1M datasets agent will pick and build with the best model for your use case and hardware