Hugging Face

Hugging Face (Twitter) RT @remi_or_: Late Christmas gift for RL people 🎄 We are adding support for parallel decoding in transformers continuous batching! You can now decode as many streams as you want from one prompt, which changes the game for long contexts 📜 but since a picture is worth 1024 words:

Hugging Face (Twitter) RT @interconnectsai: Latest open artifacts (#17): NVIDIA, Arcee, Minimax, DeepSeek, Z ai and others close an eventful year on a high note Ending the year with a bang. https://www.interconnects.ai/p/latest-open-artifacts-17-nvidia-arcee

‌Hugging Face (Twitter) RT @ClementDelangue: Gemini open-source would kill the game and definitely set GDM as the #1 in AI!

Hugging Face (Twitter) RT @AdinaYakup: Youtu-LLM 🔥 2B model with 128K context and strong agentic abilities from @TencentHunyuan https://huggingface.co/tencent/Youtu-LLM-2B

Hugging Face (Twitter) RT @mervenoyann: TII just dropped Falcon H1R-7B 🙌🏻 a new reasoning model outperforming others in math and coding with only 7B params with 256k context window 🔥 it's a mamba-transformers hybrid model so more efficient per throughput and memory too 🤩

Hugging Face (Twitter) RT @Xianbao_QIAN: @miromind_ai Miro Thinker 1.5 is out - Post-trained on top of qwen3 - Available in both 30A3B and 235A22B - Claimed to have great result on BrowserComp - Technical report coming soon - MiT license Official demo: dr.miromind.ai/ HF link: https://huggingface.co/collections/miromind-ai/mirothinker-v15https://twitter.com/Xianbao_QIAN/status/1988929067203326322#m

Hugging Face (Twitter) RT @victormustar: MiniMax just published their 2026 TODOs on Hugging Face, interesting times ahead 💫 https://huggingface.co/blog/MiniMaxAI/multilingual-and-multi-task-coding-with-strong-gen

Hugging Face (Twitter) RT @0xSero: https://huggingface.co/0xSero/MiniMax-M2.1-REAP-50-W4A16 Guys here you have it MiniMax-M2.1 with full context on 96gb vram weights are 58gb Enjoy

Hugging Face (Twitter) RT @GenrobotAI: 10Kh RealOmni-Open Dataset is now live! 10K+ hrs, 1M+ clips, 30+ skills, 3,000+ real households The largest in hours, most generalizable, largest-scale per skill,open-source embodied AI dataset. First batch: 4 scenarios, 12 skills @huggingface https://huggingface.co/datasets/genrobot2025/10Kh-RealOmin-OpenData

Hugging Face (Twitter) RT @GalaxeaDynamics: We just open-sourced G0 Plus VLA model & launched "Pick Up Anything" demo. See our robot perform diverse real-world tasks through pure language. No specialized training needed. That's zero-shot embodied intelligence. #VLA#Robotics#OpenSource 🔗Try now：https://github.com/OpenGalaxea/GalaxeaVLA

Hugging Face (Twitter) RT @TencentHunyuan: Now Tencent-HY-MT1.5-1.8B is the #1 trending model on Hugging Face! 🥇📈 Huge thanks to the community for the support! https://twitter.com/TencentHunyuan/status/2005908069239447988#m

Hugging Face (Twitter) RT @MiniMax__AI: #1 on @huggingface, once again! 171,000+ builders are already coding into 2026 with M2.1🔥