method · first seen 13 Oct 2026
INT4 quantization
Also: W4A4, 4-bit quantization, INT4 weight quantization
Compressing neural network weights to 4-bit integers to reduce memory footprint and enable inference on smaller hardware.
Mentions per week
PapersCodeDiscussion
Earliest sources
- Capacity over TFLOPS: Memory-Centric Deployment of Large Clinical Language Models on Consumer CPUsPapers · 13 Oct 2026 · Eli Yune, Umer Majeed, Mahesh Balan
- HIMX: mixed-format matrix multiplication based on customized HIFLOAT4 representationsPapers · 14 Oct 2026 · Zhiyuan Zhang, Yanzhao Li, Hui Dong, Sifan Cheng, Jingwei Zhang
- FuTransformer: FP8 prefill and INT4 decode for Llama-style inference on HopperPapers · 16 Oct 2026 · Wei Wang
- Knowledge-Guided Floodwater Level Reasoning for Cloud-Edge Collaborative MonitoringPapers · 1 Nov 2026 · Zhengshu Zhou, Shaoyu Wang, Ziyi Geng, Lu Tao, Qian Long
- tuan3w/qwen-warpCode · 1 Nov 2026 · tuan3w · 46 stars
- Gradient-Guided Hardware-Native MIXFP4Papers · 12 Nov 2026 · Jian-Jia Chen, Brian J. Chan, Wen-Hung Lee, Yi-Ting Li, Po-Shen Chen, Kuan-Te Ku…
Latest sources
- Gradient-Guided Hardware-Native MIXFP4Papers · 12 Nov 2026 · Jian-Jia Chen, Brian J. Chan, Wen-Hung Lee, Yi-Ting Li, Po-Shen Chen, Kuan-Te Ku…
- tuan3w/qwen-warpCode · 1 Nov 2026 · tuan3w · 46 stars
- Knowledge-Guided Floodwater Level Reasoning for Cloud-Edge Collaborative MonitoringPapers · 1 Nov 2026 · Zhengshu Zhou, Shaoyu Wang, Ziyi Geng, Lu Tao, Qian Long
- FuTransformer: FP8 prefill and INT4 decode for Llama-style inference on HopperPapers · 16 Oct 2026 · Wei Wang
- HIMX: mixed-format matrix multiplication based on customized HIFLOAT4 representationsPapers · 14 Oct 2026 · Zhiyuan Zhang, Yanzhao Li, Hui Dong, Sifan Cheng, Jingwei Zhang
- Capacity over TFLOPS: Memory-Centric Deployment of Large Clinical Language Models on Consumer CPUsPapers · 13 Oct 2026 · Eli Yune, Umer Majeed, Mahesh Balan