arXiv:2605.08371v2 Announce Type: replace Abstract: Multi-view geometry transformers are feed-forward 3D foundation models that jointly predict depth maps, point maps, and camera poses for N images i...
arXiv:2610.35706v1 Announce Type: cross Abstract: Attention over tens of thousands of visual tokens dominates the inference of video diffusion transformers (DiTs). Existing methods reduce this cost t...
DistillPath-KS16 is a 22‑million‑parameter ViT‑S/16 pathology encoder distilled from larger teachers ranging from 86 M to 1.1 B parameters. By training only on the teachers’ final class and patch tokens across 6,000 public slides, it avoids costly pretraining heads and large tile corpora, yet surpasses the kaiko baseline on EVA, HEST, and PLISM benchmarks. The strongest variant, DistillPath-KS16‑Virchow2, achieves a mean EVA score of 0.795—just 0.015 points below the top model—while being 29× smaller and 25× faster.
arXiv:2610.36671v1 Announce Type: cross Abstract: Diffusion vision-language models (DVLMs) process their entire input at every denoising step. Since visual tokens make up 89-97% of this input on the...
arXiv:2610.07758v1 Announce Type: cross Abstract: Training-free token reduction accelerates vision transformers by removing redundant tokens across layers, recovering most of the original accuracy at...
LeVJEPA is a video encoder that eliminates the need for architectural asymmetries, exponential-moving-average target encoders, stop-gradients, and capacity-limited predictors used in prior self‑supervised methods. It trains a single encoder with an invariance loss over global and local views, regularized by SIGReg to prevent collapse, and achieves strong performance with far less pretraining compute. The approach also allows block‑causal attention, making temporal ordering a property of the encoder itself, and matches or surpasses state‑of‑the‑art baselines on both appearance‑centric and motion‑centric benchmarks.
The paper investigates whether keeping non‑spatial register tokens in vision transformer decoders improves image reconstruction and generation. Using representation autoencoders with frozen register‑equipped encoders, the authors show that joint self‑attention over both patch and register tokens yields better reconstruction quality on ImageNet‑1K, with lower FID scores and higher PSNRs compared to patch‑only baselines. For generation, a multi‑head attention pooling predictor recovers register features from sampled patches, further reducing FID by nearly 10% over the patch‑only approach.
arXiv:2610. 23796v1 Announce Type: new Abstract: Transformer prefill reuses weights across tokens and is dominated by matrix multiplication and activation data movement; single-token decode has limited reuse and is dominated by weight and KV-cache reads.
TileSkipper introduces a region‑adaptive tile pruning strategy for 3D Gaussian splatting rasterizers, assigning a static per‑Gaussian contribution cutoff policy that is calibrated on a frozen checkpoint. The method partitions Gaussians into 64 groups, uses a single‑byte policy per Gaussian, and requires no runtime updates or additional kernels. Across 13 benchmark scenes, TileSkipper achieves up to 1.238× speedups with negligible PSNR loss, and integrates with existing opacity‑aware bounds for further compiler‑only gains.
arXiv:2610.10524v1 Announce Type: new Abstract: Highly compressed video autoencoders offer an effective way to accelerate video diffusion models, as the Diffusion Transformer (DiT) operates on far fe...
TileMix is a tile‑centric mixed‑precision attention kernel that routes score‑tile groups within fused dense attention to either FP16 or INT8 computation, using compact bitmasks to decide precision per tile. By partitioning the attention matrix into hardware‑aligned tiles and updating a shared online‑softmax state, TileMix preserves dense token connectivity without requiring training and supports grouped‑query attention, variable‑length batches, and INT8 key/value caches. Benchmarks on LLaMA, Qwen, and Vicuna show that TileMix restores long‑context quality lost with uniform INT8 and improves prefill throughput over FP16, offering a controllable accuracy‑efficiency trade‑off across model families.
arXiv:2606.19617v2 Announce Type: replace-cross Abstract: We present GB-LSR (Global-Bandwidth Local Spectral Representation), a fixed-grid local spectral representation for continuous image decoding....