An atlas for the bit-crunching research flood

What it does This repository is a curated reading list of research papers on model quantization for efficient deep learning. It collects work from major AI conferences, journals, and arXiv, then sorts it into categories like language transformers, vision transformers, and convolutional networks. Many entries carry methodological tags so you can scan for specific techniques rather than guessing from titles alone.

The interesting bit Instead of dumping every paper into one alphabetical heap, the list splits literature by model architecture and downstream task—image classification, object detection, super resolution, even point clouds. The maintainers label entries with shorthand like PTQ (post-training quantization) and Extreme (binary/ternary), which turns the index into a quick filter for practitioners hunting a specific constraint.

Key highlights

• Covers both the current LLM quantization craze and older CNN-centric work.
• Papers grouped by architecture: Language Transformers, Vision Transformers, Visual Generation, and CNN task variants.
• Methodology tags include PTQ, Non-uniform, MP (mixed-precision), and Extreme.
• Includes survey papers for newcomers who need orientation before diving into the deep end.
• Actively maintained with additions from recent venues such as NeurIPS, ICML, ICLR, CVPR, and ECCV.

Caveats

• It is strictly a bibliography: no code, no reproduced benchmarks, and no side-by-side accuracy comparisons.
• The update log contains some odd date/venue pairings (for example, a March 2024 entry claiming ICLR 2025 papers), so the timeline may require a grain of salt.

Verdict Bookmark this if you are researching quantization methods or writing a literature review and need a map of the territory. Skip it if you are looking for ready-to-run quantization frameworks or copy-paste training scripts.