Uncertainty is introduced in optimized DNNs through stochastic algorithms, forming specific distributions.

Deep learning has seen remarkable advancements in machine learning, yet it often demands extensive annotated data.

We study a class of optimization problems motivated by automating the design and update of AI systems like coding assistants, robots, and copilots.

In this work, we present a dual-stage low-bit post-training quantization (PTQ) method for image super-resolution, namely 2DQuant, which achieves efficient and accurate SR under low-bit quantization.