Efficient real-world deployments of large language models (LLMs) rely on Key-Value (KV) caching for processing and generating long outputs, reducing the need for repetitive computation. For large contexts, Key-Value caches can take up tens of gigabytes of device memory, as they store vector representations for each token and layer. Recent work has shown that the cached vectors can be compressed through quantization, pruning or merging, but these techniques often compromise quality towards higher compression rates. In this work, we aim to improve Key & Value compression by exploiting two observations: 1) the inherent dependencies between keys and values across different layers, and 2) the existence of high-compression methods for internal network states (e.g. attention Keys & Values). We propose AQUA-KV, an adaptive quantization for Key-Value caches that relies on compact adapters to exploit existing dependencies between Keys and Values, and aims to "optimally" compress the information that cannot be predicted. AQUA-KV significantly improves compression rates, while maintaining high accuracy on state-of-the-art LLM families. On Llama 3.2 LLMs, we achieve near-lossless inference at 2-2.5 bits per value with under 1% relative error in perplexity and LongBench scores. AQUA-KV is one-shot, simple, and efficient: it can be calibrated on a single GPU within 1-6 hours, even for 70B models.

This .zip archieve contains 10 scenes and 5 styles for each scene. Also there are several folders with experimental data: 1-3 indestinguishable runs run with reduced control run with specified region run on weekend all runs combined in one data Each data contains raw .tsv file with all data extracted from voting procedure. Also there is some procecced data that contains scores of each image, nuber of votes for each image an some extra information.

A camera’s auto-white-balance (AWB) module operates under the assumption that there is a single dominant illumination in a captured scene. AWB methods estimate an image’s dominant illumination and use it as the target "white point" for correction. However, in natural scenes, there are often many light sources present. We performed a user study that revealed that non-dominant illuminations often produce visually pleasing white-balanced images and, in some cases, are even preferred over the dominant illumination. Motivated by this observation, we revisit AWB to predict a distribution of plausible illuminations for use in white balance. As part of this effort, we extend the Cube+ + illumination estimation dataset [12] to provide ground truth illumination distributions per image. Using this new ground truth data, we describe how to train a lightweight neural network method to predict the scene’s illumination distribution. We describe how our idea can be used with existing image formats by embedding the estimated distribution in the RAW image to enable users to generate visually plausible white-balance images.

This paper presents a review of the NTIRE 2023 challenge on night photography rendering. The goal of the challenge was to find solutions that process raw camera images taken in nighttime conditions conditions, and thereby produce a photo-quality output images in the standard RGB (sRGB) space. Unlike the previous year’s competition, participants were not provided with a large training dataset for the target sensor. Instead, this time they were given images of a color checker illuminated by a known light source. To evaluate the results, a sufficient number of viewers were asked to assess the visual quality of the proposed solutions, considering the subjective nature of the task. The highest ranking solutions were further ranked by Richard Collins, a renowned photographer. The top ranking participants’ solutions effectively represent the state-of-the-art in nighttime photography rendering. More results can be found at https://nightimaging.org/