paper-with-me

홈 › Papers

NF4 Isn't Information Theoretically Optimal (and that's Good)

2023-06-12 · Davis Yoshida

This note shares some simple calculations and experiments related to absmax-based blockwise quantization, as used in Dettmers et al., 2023. Their proposed NF4 data type is said to be information theoretically optimal for representing normally distributed weights. I show that this can't quite be the case, as the distribution of the values to be quantized depends on the block-size. I attempt to apply these insights to derive an improved code based on minimizing the expected L1 reconstruction error, rather than the quantile based method. This leads to improved performance for larger quantization block sizes, while both codes perform similarly at smaller block sizes.

📄 PDF Abstract BibTeX arXiv:2306.06965

Code (1)

davisyoshida/abnormal-floats 공식 구현 jax

Tasks

Quantization

Similar Papers 제목 키워드 기반

Value Pursuit Iteration

2012-12-01 · NeurIPS 2012 12 · Amir M. Farahmand, Doina Precup

Value Pursuit Iteration (VPI) is an approximate value iteration algorithm that finds a close to optimal policy for reinforcement learning and planning problems with large state spaces. VPI has two main features: First, …

Reinforcement LearningReinforcement Learning (RL)

Near-Optimal Comparison Based Clustering

2020-10-08 · NeurIPS 2020 12 · Michaël Perrot, Pascal Mattia Esser, Debarghya Ghoshdastidar

The goal of clustering is to group similar objects into meaningful partitions. This process is well understood when an explicit similarity measure between the objects is given. However, far less is known when this inform…

Clustering

Mode Connectivity in Auction Design

2023-05-18 · NeurIPS 2023 11 · Christoph Hertrich, Yixin Tao, László A. Végh

Optimal auction design is a fundamental problem in algorithmic game theory. This problem is notoriously difficult already in very simple settings. Recent work in differentiable economics showed that neural networks can e…

Good Semi-supervised VAE Requires Tighter Evidence Lower Bound

2019-09-25 · Haozhe Feng, Kezhi Kong, Tianye Zhang, Siyue Xue 외

Semi-supervised learning approaches based on generative models have now encountered 3 challenges: (1) The two-stage training strategy is not robust. (2) Good semi-supervised learning results and good generative performan…

4k

A connection between bacterial chemotactic network and optimal filtering

2020-05-27 · Kento Nakamura, Tetsuya J. Kobayashi

The chemotactic network of Escherichia coli has been studied extensively both biophysically and information-theoretically. Nevertheless, the connection between these two aspects is still elusive. In this work, we report …