The Flow has not summarised this story yet — read it at arXiv Computer Vision.
arXiv:2606. 14990v1 Announce Type: cross Abstract: Sparse autoencoders (SAEs) are standard tools for mechanistic interpretability, but current SAE families are constrained by fixed encoder nonlinearities such as ReLU, JumpReLU, and TopK.
arXiv:2608.28730v1 Announce Type: cross Abstract: Latest JPEG restoration systems achieve strong quality with large models, yet often remain too slow and expensive for efficient on-device deployment....
The paper introduces SANA‑SR, an efficient one‑step diffusion restoration framework for real‑world image super‑resolution. It uses a deep compression autoencoder to reduce latent tokens by a 32× ratio, preserving essential structures and textures, and applies a linear‑attention DiT with LoRA fine‑tuning for linear‑complexity token mixing. Experiments show competitive or superior performance on benchmark datasets, and the pruned model runs in 0.019 s with 407.95 G MACs and 344 M parameters, indicating strong potential for mobile deployment.
The paper presents a two‑stage speed‑limit sign recognizer for automotive SoCs that compresses a YOLO‑family detector via channel replay pruning and implements a lightweight five‑layer CNN reader using 4‑bit weights and 8‑bit activations. Replay pruning removes over 60% of parameters while matching parent accuracy, outperforming other pruning methods on YOLOv5n and YOLOv8n. The resulting system, with 0.72 M parameters, achieves 95.2% recognition accuracy at 94 fps on a Jetson AGX Orin, fitting within tight memory, arithmetic, and energy budgets.
The paper revisits the use of contrastive loss for post‑training CLIP models, showing that the reported catastrophic forgetting is mainly due to an overly large temperature parameter in the InfoNCE objective. By setting a smaller temperature, contrastive post‑training can actually improve CLIP. The authors introduce ComCLIP, a lightweight single‑epoch recipe that freezes the text encoder, trains the vision encoder with a properly tempered contrastive loss, an MSE anchoring loss, and a relational distillation loss, achieving comparable zero‑shot classification to CLIP‑Refine while significantly enhancing visual feature transferability and maintaining downstream compatibility with LLaVA.
arXiv:2606. 27321v1 Announce Type: cross Abstract: Sparse autoencoders (SAEs) have become a leading tool for interpreting the representations of vision foundation models, decomposing their polysemantic activations into a larger set of sparse, more monosemantic features.