Poetica

This model has been #1 trending for 3 weeks now. It's Qwen3.5-27B fine-tuned on distilled data from Claude-4.6-Opus (reasoning). Trained via Unsloth. Runs locally on 16GB in 4-bit or 32GB in 8-bit. Model: huggingface.co/Jackrong/Qwen3…

Local AI is free, fast & secure! So today we're introducing hf-mount: attach any storage bucket, model or dataset from @huggingface as a local filesystem. This is a game changer, as it allows you to attach remote storage that is 100x bigger than your local machine's disk. This is also perfect for Agentic storage!! Let's go!

PinchBench results for Qwen3.5 27B using @UnslothAI K_XL quants, best of 3, thinking enabled. TL;DR: Q3 KXL (14.5GB) or Q4 KXL (18GB) While overall the "best" results showed little degradation, if you dig into mean/std Q4_K_XL overall was the best at ~84% on average. Q3 seems viable, while Q2 is the the lowest performing, of course.

Archangel 🪽

@tejessrivalsan been waiting for this

excited to announce that we’re open sourcing EGO-SNAKE the largest dataset of egocentric snake pov footage to train the next generation of autonomous vipers comment for a data sample

@goo_vision super cool

Testing Midjourney v8

@MaziyarPanahi @julien_c @UnslothAI will be available very soon🙂

@julien_c @UnslothAI Can you run Unsloth Studio on HF spaces? 😆

OK it's training!!! (A100 80GB, there was no H100 available on Colab Pro) Yay @UnslothAI Studio

Pewdiepie reveals how to break free from the algorithm “A lot of this is going to sound crazy but you’ve gotta hear me out, it’s a step by step process. I’m not saying you should do all of it but you should try some of it” “Step 1 is creating friction. I put all social media and attention hungry apps in a second profile and I can’t understate how much this changed my life. Those 5-6 seconds it takes to switch profiles stops me every time and makes me think, is this what I want to be doing?” “The second thing I did was self hosting. The effect that had on me is I’m not the product anymore. The things I use are mine and because they’re not free, I’m not paying with my privacy. I think the main difference is ads and news don’t reach me” “Next thing I did was disable Shorts, I like YouTube but I hate how Shorts is everywhere I can’t escape it” “Then I unfollowed everyone. You don’t have to do this, this is definitely a me thing, I just got really fed up” “Next, get a DNS blocker. You can remove ads completely, most of it won’t even reach your device” “I think you owe it to yourself to take some time today and start building your tech fence” “These tech companies don’t care about you, so you’ve got to care about yourself. The cheat code is building some friction and filtering out the noise, that’s your defence and your cure”

Transform PDFs, CSV, DOCX, TXT or any file into a structured synthetic datasets via Unsloth Data Recipes. Build and edit your datasets visually via a graph-node workflow and use them for fine-tuning. Powered by @NVIDIA DataDesigner.

Introducing Unsloth Studio ✨ A new open-source web UI to train and run LLMs. • Run models locally on Mac, Windows, Linux • Train 500+ models 2x faster with 70% less VRAM • Supports GGUF, vision, audio, embedding models • Auto-create datasets from PDF, CSV, DOCX • Self-healing tool calling and code execution • Compare models side by side + export to GGUF GitHub: github.com/unslothai/unsl… Blog and Guide: unsloth.ai/docs/new/studio Available now on Hugging Face, NVIDIA, Docker and Colab.

@UnslothAI woah!🔥data designer looks super fun

@UnslothAI @nvidia @NVIDIAAIDev @NVIDIA_AI_PC NeMo Gym is fire 🔥

We collaborated with @NVIDIA to teach you about Reinforcement Learning and RL environments. Learn: • Why RL environments matter + how to build them • When RL is better than SFT • GRPO and RL best practices • How verifiable rewards and RLVR work Blog: unsloth.ai/blog/rl-enviro…

@yacineMTB Grok 4

5.4 xhigh

Introducing the Synthetic Data Playbook: We generated over a 1T tokens in 90 experiments with 100k+ GPUh to figure out what makes good synthetic data and how to generate it at scale huggingface.co/spaces/Hugging…

when people at my company ask how I'm hitting all these 16 node jobs

🚀 vLLM v0.17.0 is here! 699 commits from 272 contributors (48 new!) This is a big one. Highlights: ⚡ FlashAttention 4 integration 🧠 Qwen3.5 model family with GDN (Gated Delta Networks) 🏗️ Model Runner V2 maturation: Pipeline Parallel, Decode Context Parallel, Eagle3 + CUDA graphs 🎛️ New --performance-mode flag: balanced / interactivity / throughput 💾 Weight Offloading V2 with prefetching 🔀 Elastic Expert Parallelism Milestone 2 🔧 Quantized LoRA adapters (QLoRA) now loadable directly

agentic RL hackathon this weekend! mentors from @PyTorch, @huggingface , and @UnslothAI will guide you to build agentic environments to win from a pool of $100K prizes 🏆 + free compute and token credits just for attending! lock in mar 7-8 in SF.

Qwen3.5 can now be fine-tuned locally with Unsloth via LoRA using only 10GB VRAM. You can then export to GGUF for llama.cpp, Ollama, LM Studio inference! Unsloth also supports Qwen3.5‑35B‑A3B LoRA, using ~74GB VRAM (1x H100)