Hugging Face

‌Hugging Face (Twitter) RT @AdinaYakup: MiroThinker v1.0 🔥 open research agent released by @miromind_ai https://huggingface.co/collections/miromind-ai/mirothinker-v10 ✨ 8B/30B/72B - MIT license ✨ 256K context, 600 tool calls ✨ Interactive scaling: beyond model size & context

Hugging Face (Twitter) RT @multimodalart: Apply Texture Qwen Image Edit LoRA by tarn59 works with EVERYTHING! 👉🪵🧶, this model trains so well I've built this demo so you can apply *any* texture to *any* object on @huggingface

Hugging Face (Twitter) RT @Xianbao_QIAN: @miromind_ai released their open source latest model on research agent. - interleaved thinking with multi-step analysis, powered by RL - 256k context window - handles up to 600 tool calls per task - comes in 8B, 30B, and 72B for different compute budgets - MiT license Model available on @huggingface , demo and technical report ready.

Hugging Face (Twitter) RT @askOkara: Major open source wins this year: - DeepSeek cracked cheap long context for LLMs - Qwen dropped the best vision model in the world - Llama 4 Scout arrived with a 10M context window - Minimax M2 delivered sonnet 4.5-level performance at 8% of its price - GLM 4.6 brought a 200k token context window and performs with Sonnet 4 - Kimi k2 thinking outperformed GPT-5 and Claude 4.5 Sonnet on several benchmarks while being 10x cheaper than GPT-5 and 20x cheaper than Sonnet 4.5 The gap between closed and open-source AI is closing fast.

Hugging Face (Twitter) RT @JordanNanos: thanks to @huggingface@ClementDelangue for this insight, backing up some of our n=140 ClusterMAX research with real data

‌Hugging Face (Twitter) RT @nyxkrage: "HuggingFace port" of @moondreamai's Moondream 3. Weights are converted to a more conventional HF format with nn.Module-based modeling code. Drops right into trl and peft, enabling QLoRA fine-tunes in < 16GB VRAM. Implementation code is Apache 2.0; model weights remain under BSL. hf.co/NyxKrage/moondream3-hf

‌Hugging Face (Twitter) RT @redtachyon: >Trained for just $7.8K 🤯 >looks inside >finetune every. single. time.

Hugging Face (Twitter) RT @NielsRogge: If there's one @huggingface user you should follow it's @dx8152! He created the amazing Multiple Angles (camera control) LoRa on top of @Alibaba_Qwen Image Edit , the Light Restoration LoRa (below), and many more All his LoRa's are trained using modelscope

‌Hugging Face (Twitter) RT @AdinaYakup: China's twitter - Weibo steps into open source AI🚀 VibeThinker 1.5B 🔥 dense language model from @WeiboLLM https://huggingface.co/WeiboAI/VibeThinker-1.5B ✨ Trained for just $7.8K 🤯 ✨ MIT license ✨ Outperforms DeepSeek R1 in math reasoning (AIME24: 80.3 vs 79.8) ✨ Spectrum to signal principle: diversity driven training → superior reasoning

Hugging Face (Twitter) RT @andimarafioti: Sneak peek of Reachy Mini's conversation capabilities! He can chat fluidly and change languages naturally. You can interrupt him and he can see and react to his environment. They start shipping soon, I'm excited to see what people start building with him!

Hugging Face (Twitter) RT @ClementDelangue: Cool to see transformers number 4 top open-source project by contributors in 2025 in that latest @github octoverse! With vLLM, Codex, Ollama, llama.cpp all deeply integrated with transformers, it's an AI take-over of github this year!

Hugging Face (Twitter) RT @ClementDelangue: You can just open-source things! If you think you don’t have enough experience, support or ressources to release useful open-source, remind yourself that @eddybuild is 18 and released the currently number one trending dataset on Hugging Face, ahead of Nvidia and Meta (respectively valued at 5 and 1.6 trillion dollars). Let's go! https://twitter.com/eddybuild/status/1987951619804414416#m