Hugging Face (Twitter)

RT @reach_vb: BOOM! Latest Qwen 30B A3B 2507 running blazingly fast on Mac powered by MLX 💥

mlx_lm.chat --model "lmstudio-community/Qwen3-30B-A3B-Instruct-2507-MLX-4bit"

That's it, try it out today!

Hugging Face (Twitter)

RT @roo_code: Got a favorite @huggingface model? Now it lives in your editor. 🤗

Roo Code makes it easy to connect your API key, choose from 90+ models, and select your preferred inference provider in just a few clicks.

Watch the quick tutorial and explore more: https://docs.roocode.com/providers/huggingface

Hugging Face (Twitter)

RT @vanstriendaniel: I just processed 1000s of prompts using Qwen3-235B-A22B-Instruct-2507 across 4 GPUs!

How? Everyone plays their part:
@astral_sh UV handles dependencies
@huggingface Jobs handles GPUs
@Alibaba_Qwen handles the model
@vllm_project handles inference

One command. Zero complexity!

Hugging Face (Twitter)

RT @lhoestq: > hf jobs is just out and damnnnn I love the uv integration 💛

@huggingface made their scripts uv-ready to run them on HF infra without setting up docker or dependencies.

E.g.
run DPO locally > uv run dpo․py
run DPO on HF > hf jobs uv run dpo․py

Bonus: --flavor for GPUs🔥

Hugging Face (Twitter)

RT @NielsRogge: Efficient LoFTR was just integrated into @huggingface Transformers!

It improves upon LoFTR, a detector-free image matcher, by being 2.5x faster. It can even surpass the SOTA efficient sparse matching pipeline SuperPoint + LightGlue.

Now available in a few lines of code!

Hugging Face (Twitter)

RT @jandotai: Jan v0.6.6 is out: Jan now runs fully on llama.cpp.

- Cortex is gone, local models now run on @ggerganov's llama.cpp
- Toggle between llama.cpp builds
- @huggingface added as a model provider
- Hub enhanced
- Images from MCPs render inline in chat

Update Jan or grab the latest.

Hugging Face (Twitter)

RT @sleenyre: Happy to have released my first model. A lot of work went into making Krea 1 open source. https://twitter.com/krea_ai/status/1950921488871408075#m

‌Hugging Face (Twitter)

RT @NVSWSourcer: 👀 We just opened over 26M lines of synthetic data that was used to train the Llama Nemotron Super v1.5 model.

🔢 Find them on Hugging Face 🤗 bit.ly/4l8DIc7

Hugging Face (Twitter)

RT @ClementDelangue: let's open superintelligence for all!

Hugging Face (Twitter)

RT @NVIDIAAIDev: 👀 We just opened over 26M lines of synthetic data that was used to train the Llama Nemotron Super v1.5 model.

🔎 This transparency into our model training also helps you build your own models -- without expending the effort and time required to produce your own datasets.

🔢 Find them on @HuggingFace 🤗 https://huggingface.co/datasets/nvidia/Nemotron-Post-Training-Dataset-v1

Hugging Face (Twitter)

RT @ClementDelangue: If you're a researcher or engineer releasing open science papers & open models and datasets, I bow to you 🙇🙇🙇

From what I'm hearing, doing so, especially in US big tech, often means fighting your manager and colleagues, going through countless legal meetings, threatening to quit or taking a lower paycheck, and sometimes the result is only that you'll get scolded when what you shared is used by competitors.

But, please remember: research papers and open models and datasets is how progress happens! Your efforts are pushing AI toward a more open and collaborative future. Thanks to openness, your research or models get a chance to be noticed, seen & built upon by people you respect to accelerate progress, grow your network & accelerate your impact.

It might be tough right now but open science will ultimately prevail as it always did! The researchers & engineers that we'll remember in ten years are the ones who share what they build, not the ones that keep it behind closed-doors for company profit maximization.

Please keep fighting for openness. We see you and we thank you! 💚💛 💙💜

Hugging Face (Twitter)

RT @Xianbao_QIAN: Step 3 has just been released. It proposed a new infra level optimization of Attention, FFN disaggregation.

Model & Infra co-design is the way forward!

Model: https://huggingface.co/stepfun-ai/step3
Technical paper: arxiv.org/abs/2507.19427

Hugging Face (Twitter)

RT @nickfrosst: cohere vision model :)

weights on huggingface

https://huggingface.co/blog/CohereLabs/introducing-command-a-vision-07-2025

Hugging Face (Twitter)

RT @victormustar: Black Forest Labs did a great job here, really like the vibe of the outputs here.

👇free demo is available on Hugging Face https://twitter.com/bfl_ml/status/1950920537741336801#m

Hugging Face (Twitter)

RT @bfl_ml: Today we are releasing FLUX.1 Krea [dev] - a new state-of-the-art open-weights FLUX model, built for photorealism.

Developed in collaboration with @krea_ai, this model is focused on images with unique aesthetics. No “AI look”, no blown-out highlights, just natural detail.

Hugging Face (Twitter)

RT @reach_vb: New favourite model: Flux.1 Krea Dev by @bfl_ml 🔥

Focused on aesthetics - nails prompt guidance too! - You can run for free via ZeroGPU! 🤗

Hugging Face (Twitter)

RT @multimodalart: I've built a demo to allow you to navigate some of the immersive worlds generated by HunyuanWorld 🌎 https://twitter.com/TencentHunyuan/status/1949288986192834718#m

‌Hugging Face (Twitter)

RT @Akashi203: I released the first Arabic reasoning dataset, designed to help train and fine-tune AI models for reasoning tasks in Arabic.

It’s open-sourced on Hugging Face:

https://huggingface.co/datasets/Jr23xd23/Arabic-Optimized-Reasoning-Dataset

Hugging Face (Twitter)

RT @eliebakouch: If you’re a researcher working on RL, you should definitely try SmolLM3-3B and get another data point besides Qwen3-3B.

1) We didn’t have time to try RL during post training, so I think there’s still some room to build an even better version of smollm!

2) We released the intermediate checkpoints from post training, so you can use our model at different stages (base, mid training, SFT, APO, merging,) and see if it changes RL perf.

3) The model is also pretty good at long context, you can probably push it past 128k thanks to NoPE and yarn.

Hugging Face (Twitter)

RT @julien_c: 50 (!) LLMs released these past 2-3 weeks.

But the real kicker is when you think of this:

It is the most releases we’ve seen so far, but the least releases we’ll see in the future 🤯

Hugging Face (Twitter)

RT @ClementDelangue: Every tech company can and should train their own deepseek R1, Llama or GPT5, just like every tech company writes their own code (and AI is no more than software 2.0).

This is why we're releasing the Ultra-Scale Playbook. 200 pages to master:
- 5D parallelism (DP, TP, PP, EP, FSDP)
- ZeRO
- Flash Attention
- Compute/communication overlap and bottlenecks

All with accessible theory intros and 4,000+ scaling experiments.