The AI community building the future. https://nitter.cf/t.co/TpiXQMQ9rZ

NYC and Paris and 🌏
Joined September 2016
Hugging Face retweeted
Meet Qwen-Image-2.1, the most balanced and cost-effective image generation model in the Qwen-Image series! Now open weights! 🎨 A unified model for both generation and editing, delivering top-tier quality in a lightweight package. Highlights: 👀 - Compact & exceptionally fast: A lightweight 7B architecture that outperforms most closed-source models, with drastically accelerated inference for multi-image inputs. - Native transparency: Natively generates and edits RGBA layers, enabling seamless compositing and text editing within transparent images. - Versatile, high-fidelity editing: Supports up to 10 reference images and precise local control while preserving strict fidelity for portraits and products. - Broad coverage & stunning aesthetics: Excels at panoramas, infographics, and virtual try-ons, delivering realistic textures and elegant typography. Start to create your next masterpiece with Qwen-Image-2.1! 🖼️ - Blog: qwen.ai/blog?id=qwen-image-2… - GitHub: github.com/QwenLM/Qwen-Image… - Model Scope: modelscope.cn/models/Qwen/Qw… - Hugging Face: huggingface.co/Qwen/Qwen-Ima…
98
310
163
2,535
261,255
Hugging Face retweeted
2 new OCR models landed on @huggingface today - tencent/WeVisDoc (2B and 4B), Apache 2.0 - jinaai/jina-ocr-v1, non-commercial license Which one should you pick? Based on the OmniDocBench v1.6 benchmark on Papers with Code, WeVisDoc-4B leads with an overall score of 95.38, but the current SOTA is NaviDC-OCR
9
21
152
16,536
Hugging Face retweeted
Found a faster kernel? You shouldn’t need to rewrite your model to use it. With 🤗 Kernels, you can choose which kernel runs a supported layer and replace its forward() with an optimized implementation. Here’s how 🧵 1/5
5
4
1
45
11,601
Hugging Face retweeted
Introducing the Papers with Code MCP server! To celebrate its release, I've asked Claude Code to research the architecture of Jev by @typesafeai using the `search_papers` tool. Here's what it came up with: "If Jev is architecturally anything public, it's a large schema-conditioned bidirectional encoder with parallel label-query heads (GLiNER/ML-Decoder shape) trained with a proper-scoring-rule RL objective (RLCR shape), scaled far past the ~150M-param range those papers operate in — "neither small nor an LLM" fits that. The 40–200× speedup is consistent with removing autoregression, not with any exotic mechanism."
17
6
1
89
19,496
Hugging Face retweeted
Today, we’re announcing Ternary Bonsai 2 27B. Based on Qwen3.8 27B, Bonsai 2 27B is 9x smaller than its full-precision counterpart while retaining 98.2% of its aggregate benchmark performance. Two months after the first Bonsai 27B release, the biggest change is quality. The footprint remains 5.9 GB, but the gap to full precision has narrowed materially, with particularly strong gains in agentic coding, multimodal reasoning, and long-horizon tool use. Ternary Bonsai 2 27B is available today under Apache 2.0.
628
1,544
755
15,385
5,186,650
too many open jev claims and reproductions which one works? which one can you run on your laptop? i created this tracker that categorizes 38 artefacts (github repos, hugging face models) and separates signal from noise huggingface.co/spaces/multim…
17
31
5
245
26,979
Hugging Face retweeted
I only wear ties when I'm forced to go to DC 😅 (should we do HF ties btw?). Great chat with @alexanderburns @politico. We need more transparency and open-source in AI! politico.com/video/2026/09/1…
33
21
1
199
27,282
Hugging Face retweeted
So we actually built this! Paper: arxiv.org/abs/2606.31567 Demo: ai-reports.org/ We are actually looking for an org to run this fulltime as researchers in their individual capacity don't have bandwidth. If you are an independent org interested in maintaining a registry of flaw/incident reports and follow up with model providers pls reach out!
i think we need to create a "model card" equivalent for reporting misalignment incidents, this would guarantee a certain level of transparency and help build a better understanding over time. some ideas for what the fields could be: - date of the incident/detection/report (already present in the examples oai reported!) - frequency: how often does this behavior happen (number or % of rollouts affected) - stage: does this happen during eval or RL training. if training: do we expect this behavior to be reinforced by RL? evolution of % of rollouts affected over time - detection: was this incident caught by the current monitoring system? - task category: broad description of the tasks where the misalignment happened (cyber, research, web search, basic Q&A, biology etc.) - model family: what model family is affected (Sol, Astra etc.) - novelty: is this an issue we were already aware of or not? - external impact: did the incident have an external impact (i.e. wiki incident would have been yes) this is just some random ideas i had (more in thread that are a bit more "complex"), we need to add more that would contribute to increase transparency and understanding. but it's also very important that this does NOT slow down the process of reporting misalignment behavior! some examples from the incident reported by oai recently
9
19
130
25,718
The AI revolution will not be centralized. 🤗 📢 Calling all AI builders to San Francisco on Friday, October 16, one month from today! We’ll kick off Open Source AI Week with Open Together: a huge meetup, community demos, and a dance party. Come meet the people building with open models. Everyone should be able to build AI they can control. That’s the gift of open models, and a future worth celebrating together. I’d love to see you there! Who’s coming? Tell me what you’re building in the replies 👇 Register here: luma.com/OpenTogether
🤖 Made with AI
10
24
4
86
17,077
Hugging Face retweeted
hum anything, get a finished song 🎤 Hum-to-Song is an YuE2 LoRA that turns humming into full songs 🎼 hum in, a produced track, verse, chorus, and the tune you hummed coming back around ▶️ on Spaces hf.co/spaces/hugging-apps/yu…
5
37
5
291
29,836
Hugging Face retweeted
We presented our work Flash-BoN on the 11th Sept at #ECCV26. It was a fulfilling experience, to say the least! @RawalRuchit told me about the idea in Hawai'i during ICCV'25, and I was immediately like, let's go! The origins of the work started with a curiosity: Change the metric for inference-time scaling algos in diffusion from number of function evals (NFE) to something more bounded, like wall-clock time. We decided to spend that (inference-time) precious compute exploring more candidates WITHOUT busting the tanks. It's so cool that Flash-BoN works across the board: T2I, T2V; different model scales; complements other techniques like BFS, ReflectionFlow; and even improves the convergence of Flow-GRPO. If you haven't checked it out yet, here's the link: flash-bon.github.io/
3
8
98
16,889
Hugging Face retweeted
As the first publicly disclosed agent cyberattack victim, we've had a front-row seat to this new risk. I formalized my thinking about it below. I'll be in DC tomorrow to share more with policymakers and at decoded summit by @politico!
161
224
63
1,132
167,533
Hugging Face retweeted
Tencent just dropped AuK on @huggingface. It is the nano banana for audio: edit content, voice, emotional tone of any audio ▶️ on Spaces hf.co/spaces/tencent/AuK
12
48
1
515
51,769
RT @ClementDelangue: It's now clear that: - alignment is critical to making AI safe - alignment won't be solved behind the closed doors of…
This quoted post is unavailable.
26
4,674
Hugging Face retweeted
This is so cool! One of our community members built a Space on @huggingface with step by step build instructions for SO-ARM101. It includes nice 3D visualizations of each step with detailed description. Here is a link to the space: huggingface.co/spaces/KitSmi… Join our Discord and share what you built: discord.com/invite/s3KuuzsPF…
9
28
250
19,521
Hugging Face retweeted
New in Gradio 6.27: Streaming audio chunks are treated as if they were a single AAC file, preserving codec state throughout the stream. In other words, no more buffering in your audio demos!
10
8
42
11,720
Hugging Face retweeted
Along with everything else, we've been improving Gradio and just released 6.27! In this release, you'll see a bunch of nice bug fixes, especially around audio streaming and workflows! github.com/gradio-app/gradio…
🤖 Made with AI
4
27
11,301
The real tokenmaxxing move: use frontier agents to build small classifiers for large-scale data curation. Small classifiers already help curate the data used to train LLMs. I wanted to see how far an agent could take me in building one. I used Astra, SetFit and @huggingface Jobs to turn 200 agent-labelled examples into a reusable document-purpose classifier, reviewing the categories and tricky cases along the way. It classified 191,724 FinePDFs-Edu documents for ~$0.70 in inference compute, versus an estimated $13–26 to label the same document excerpts with low-cost batch LLMs. Training experiments added ~$2.90 in compute. Workflow, mistakes, reusable model and a prompt to try on your own data: danielvanstrien.xyz/posts/20…
8
16
3
207
39,023
Hugging Face retweeted
🤗 Novita now supports DeepSeek-V4.1-Flash on @huggingface. • 552B backbone parameters • Native image and text input • Up to a 1M-token context window • Continuously controllable reasoning effort
2
6
35
9,878
Hugging Face retweeted
For folks wondering what YOCO means, you can simply ask it on Papers with Code :) It powers the new DeepSeek-V4.1-Flash architecture. It's not an encoder-decoder, but rather a decoder-decoder architecture. It's all about reducing GPU memory and prefill latency. Find the chat here: paperswithcode.co/share/a943…
3
8
1
28
15,316