🎉🎉🎉Introducing our latest project: Segment-Anything NeRF, a novel approach for performing interactive 3D-consistent segmentation in NeRF. By leveraging the lightweight SAM decoder, we can achieve 5 FPS (rendering 512x512 image) on a V100. Link: github.com/ashawkey/Segme…

VisionLLM: Large Language Model is also an Open-Ended Decoder for Vision-Centric Tasks paper page: huggingface.co/papers/2305.11…

Welcome to follow our latest work, VisionLLM! 🥰

LGM Large Multi-View Gaussian Model for High-Resolution 3D Content Creation paper page: huggingface.co/papers/2402.05… 3D content creation has achieved significant progress in terms of both quality and speed. Although current feed-forward models can produce 3D objects in seconds,

🔥Large Multi-View Gaussian Model (LGM)🔥 We introduce #LGM, a feed-forward foundation model for text-to-3D and image-to-3D, which generates high-res 3D content in 5s - Project: me.kiui.moe/lgm/ - Code: github.com/3DTopia/LGM - Demo Hugging Face: huggingface.co/spaces/ashawke…

LGM demo is out text and image to 3d in seconds HF demo: huggingface.co/spaces/ashawke… local demo: github.com/3DTopia/LGM?ta…

As the author of Janus, I’m excited to share this work with you! We’ve worked hard to build a model that’s not only flexible and powerful but ready to set new benchmarks in unified multimodal AI. 💡🔥

🚀 Introducint JanusFlow: harmonizing autoregressive LLMs with rectified flow! By adopting the best practices in both fields, JanusFlow excels at both image understanding & generation in a single model. ⚡️ Powerful, simple, flexible, & your next-gen of Janus is here! 🔥 📄

😀😀DeepSeek-VL2 is coming!

Congrats to our team! It's a milestone.🚀

🚀 DeepSeek-R1 is here! ⚡ Performance on par with OpenAI-o1 📖 Fully open-source model & technical report 🏆 MIT licensed: Distill & commercialize freely! 🌐 Website & API are live now! Try DeepThink at chat.deepseek.com today! 🐋 1/n

I tested Janus-Pro-7B from Deepseek, and wow—it’s leagues ahead of DALL-E. Here’s the prompt I used and the mind-blowing result. 👇

🚨Please be vigilant and watch out for fake accounts related to DeepSeek! Stay safe and double-check before engaging. 🔍

Janus-Pro generates images at a resolution of 384x384 and utilizes a 16x downsampling discrete encoder, it may lag behind the latest text-to-image models in terms of image clarity and detail rendering. However, it demonstrates strong instruction-following capabilities.#JanusPro7B

⚠️ Attention The account @XIAOKANG4CHEN is impersonating me. This is a FAKE account and does NOT represent me or my views. Please do not follow or engage with it. #Imposter #FakeAccount

Get Ready for #OpenSourceWeek!

👍🏻

Introducing DeepSeek-V3.1: our first step toward the agent era! 🚀 🧠 Hybrid inference: Think & Non-Think — one model, two modes ⚡️ Faster thinking: DeepSeek-V3.1-Think reaches answers in less time vs. DeepSeek-R1-0528 🛠️ Stronger agent skills: Post-training boosts tool use and