Hugging Face

Hugging Face (Twitter) RT @awnihannun: Over 10k MLX models have been uploaded to the @huggingface Hub. Models for every modality from the latest LLMs, to speech recognition and synthesis, video generation, and more. Thanks to the amazing community!

Hugging Face (Twitter) RT @Gradio: All this through a new Python library called daggr!🔥 Build AI workflows that can connect ML models, custom Python functions, and AI apps together🤯

Hugging Face (Twitter) RT @abidlabs: To see how useful (and fun) building Daggr flows can be, you have to try it yourself. Luckily the Python library is very simple to get started -- an intuitive API for users and plenty of logging for LLMs to use easily. Install and star Daggr today: github.com/gradio-app/daggr

Hugging Face (Twitter) RT @abidlabs: After working on Gradio for more than 6 years, I'm finally launching a new library to build AI workflows called: Daggr 🔥 We created it to solve the BIGGEST pain point we heard from users: how do you build reliable, complex AI apps from unreliable, simple models?

Hugging Face (Twitter) Read the blog: huggingface.co/blog/daggr Check features apps: https://huggingface.co/collections/ysharma/daggr-hf-spaces daggr on GitHub: github.com/gradio-app/daggr🤗

Hugging Face (Twitter) Introducing daggr: a new way of building apps 🔥 daggr combines best of all worlds, mix-and-match model endpoints, Gradio apps, functions programmatically, inspect the pipeline visually 🙌🏻 Try it out, build and share to get featured!

Hugging Face (Twitter) RT @cgeorgiaw: This breakthrough model from @GoogleDeepMind can evaluate DNA sequences of up to 1 million base pairs + predict at single base-pair resolution for: 🧬 gene expression 🧬 splicing patterns 🧬 chromatin features 🧬 contact maps All on @huggingface! Future of science is open 🤗https://twitter.com/GoogleDeepMind/status/2016542480955535475#m

Hugging Face (Twitter) RT @Alibaba_Qwen: Qwen3-ASR and Qwen3-ForcedAligner are now open source — production-ready speech models designed for messy, real-world audio, with competitive performance and strong robustness. ● 52 languages & dialects with auto language ID (30 languages + 22 dialects/accents) ● Robust in noisy and complex settings (yes, singing and songs too) ● Long audio support: up to 20 minutes per pass ● Word/phrase-level timestamps: high-precision alignment for 11 languages via Qwen3-ForcedAligner, stronger than MFA/CTC/CIF-style aligners Also included: a full open-source inference & finetuning stack with vLLM batch, streaming, and async serving. GitHub: github.com/QwenLM/Qwen3-ASR Hugging Face: https://huggingface.co/collections/Qwen/qwen3-asr ModelScope: https://modelscope.cn/collections/Qwen/Qwen3-ASR Hugging Face Demo: https://huggingface.co/spaces/Qwen/Qwen3-... Перейти на оригинальный пост

Hugging Face (Twitter) RT @Xianbao_QIAN: It hasn't been noticed by many but @UnitreeRobotics has open sourced the weights of UnifoLM-VLA model on @huggingface together with a few robotics datasets collected using G1 in @LeRobotHF format in Apache license Let's rock! Check them out: https://huggingface.co/unitreerobotics

Hugging Face (Twitter) RT @Nik__V__: MapAnything V1.1 Release is live! 🚨 ✅ Improved Checkpoints ✅ Model factory to test & train many models ✅ Profiling ✅ New COLMAP demos & voxelization tooling ✅ WAI format Benchmarking Data Time to update the CVPR subs 😉 Comparisons to DA3 (1.1), Pi3X & more info in 🧵👇

‌Hugging Face (Twitter) RT @RisingSayak: Two papers made it to #ICLR26. Happy 🤗 StructBench: arxiv.org/pdf/2510.05091 NoiseRefine: arxiv.org/pdf/2412.03895 Grateful to my collaborators and HF for letting me in on the fun 🚗

Hugging Face (Twitter) RT @ben_burtenshaw: We got Claude to teach open models how to write CUDA kernels. This blog post walks you through transferring hard capabilities (like kernel writing) between models with agents skills. Here's the process: - get a powerful model (like Claude Opus 4.5 or OpenAI GPT-5.2) to solve a hard problem - convert that trace into an agent skill - transfer it to open-source, cheaper, or local model - measure if it actually helps We tested this on a gnarly task: writing CUDA kernels for diffusers. The results? Some open models saw +45% accuracy improvements with the right skill. But the skill didn't help every model equally. Some even degraded performance, or used way more tokens. If you're transferring skills, you should evaluate. We used upskill, a new tool for generating and evaluating agent skills. It works like this: uvx upskill generate "write nvidia kernels" --from ./trace.md