Recent 3D generation models typically rely on limited-scale 3D ‘gold-labels’ or 2D diffusion priors for 3D content creation. However, their performance is upper-bounded by constrained 3D priors due to the lack of scalable learning paradigms. In this work, we present See3D, a visual-conditional multi-view diffusion model trained on large-scale Internet videos for open-world 3D creation. The model aims to Get 3D knowledge by solely Seeing the visual contents from the vast and rapidly growing video data — You See it, You Got it. To achieve this, we first scale up the training data using a proposed data curation pipeline that automatically filters out multi-view inconsistencies and insufficient observations from source videos. This results in a high-quality, richly diverse, large-scale dataset of multi-view images, termed WebVi3D, containing 320M frames from 16M video clips. Nevertheless, learning generic 3D priors from videos without explicit 3D geometry or camera pose annotations is nontrivial, and annotating poses for web-scale videos is prohibitively expensive. To eliminate the need for pose conditions, we introduce an innovative visual-condition - a purely 2D-inductive visual signal generated by adding time-dependent noise to the masked video data. Finally, we introduce a novel visual-conditional 3D generation framework by integrating See3D into a warping-based pipeline for high-fidelity 3D generation. Our numerical and visual comparisons on single and sparse reconstruction benchmarks show that See3D, trained on cost- effective and scalable video data, achieves notable zero-shot and open-world generation capabilities, markedly outperforming models trained on costly and constrained 3D datasets. Additionally, our model naturally supports other image-conditioned 3D creation tasks, such as 3D editing, without further fine-tuning. Please refer to our project page at: https://vision.baai.ac.cn/see3d.

Sketch animation, the transformation of static sketches into dynamic experiences, is an essential tool for visual expression. Current methods rely on global-level optimization and neglect instance-level priors, resulting in optimization mismatch and overfitting across multiple sketches. To solve this problem, information about the attention map, representing different instances is needed to guide the optimization. In this work, we propose AnimateSketches, a novel optimization-based framework that focuses on animating complex vectorized sketches. We introduce prompt-guided instance-aware mask generation (PGIM), which leverages attention maps from a pretrained diffusion model to guide the optimization of individual sketches. In addition, we use mask-based score distillation sampling (MSDS) to maintain the integrity of untargeted sketches. Quantitative and qualitative evaluations show the superiority of our approach over baseline methods in terms of visual quality and instance-level prompt-guided correspondence.

Existing encoder-free vision-language models (VLMs) are rapidly narrowing the performance gap with their encoder-based counterparts, highlighting the promising potential for unified multimodal systems with structural simplicity and efficient deployment. We systematically clarify the performance gap between VLMs using pre-trained vision encoders, discrete tokenizers, and minimalist visual layers from scratch, deeply excavating the under-examined characteristics of encoder-free VLMs. We develop efficient strategies for encoder-free VLMs that rival mainstream encoder-based ones. After an in-depth investigation, we launch EVEv2.0, a new and improved family of encoder-free VLMs. We show that: (i) Properly decomposing and hierarchically associating vision and language within a unified model reduces interference between modalities. (ii) A well-designed training strategy enables effective optimization for encoder-free VLMs. Through extensive evaluation, our EVEv2.0 represents a thorough study for developing a decoder-only architecture across modalities, demonstrating superior data efficiency and strong vision-reasoning capability. Code is publicly available at: https://github.com/baaivision/EVE.

This paper presents a novel approach that enables autoregressive video generation with high efficiency. We propose to reformulate the video generation problem as a non-quantized autoregressive modeling of temporal frame-by-frame prediction and spatial set-by-set prediction. Unlike raster-scan prediction in prior autoregressive models or joint distribution modeling of fixed-length tokens in diffusion models, our approach maintains the causal property of GPT-style models for flexible in-context capabilities, while leveraging bidirectional modeling within individual frames for efficiency. With the proposed approach, we train a novel video autoregressive model without vector quantization, termed NOVA. Our results demonstrate that NOVA surpasses prior autoregressive video models in data efficiency, inference speed, visual fidelity, and video fluency, even with a much smaller model capacity, i.e., 0.6B parameters. NOVA also outperforms state-of-the-art image diffusion models in text-to-image generation tasks, with a significantly lower training cost. Additionally, NOVA generalizes well across extended video durations and enables diverse zero-shot applications in one unified model. Code and models are publicly available at https://github.com/baaivision/NOVA.

Sketch is an abstract visual representation that can be recovered as natural photographs in the human mind. Many researchers are drawn to work on translating abstract sketches to natural photographs. Since conventional sketch-to-image models are designed to generate images with a single object as the subject, generating scene image with multiple classes of objects is a tricky problem. To tackle this challenge, we propose the first scene sketch-to-image generation method based on diffusion models. Our model uses an encoder to summarize the contour and class features of the scene sketch into a latent variable, and a decoder to reconstruct scene images from it. In scene sketch-to-image generation tasks, our method outperforms the state-of-the-art methods. Experiments also show that our model beats other methods in zero-shot general sketch-to-image generation. It demonstrates our model’s potential for full-domain image generation.

Channel estimation in millimeter-wave (mmWave) massive MIMO systems is challenging, owing to the large number of antennas and the sparsity of mmWave channels. In this paper, we apply a distributed computing technique and design a channel estimation algorithm inspired by accelerated projection-based consensus (APC). The algorithm can thus run in parallel fashion. Simulation results show that the distributed algorithm has high reliability and high convergence rate under different noise levels.