SNIP: An Adaptive Mixed Precision Framework for Subbyte Large Language Model Training
Open Access
- 22 March 2026
- conference paper
- Published by Association for Computing Machinery (ACM)
- p. 1815-1831
Abstract
No abstract availableAll Related Versions
This publication has 13 references indexed in Scilit:
- PyTorch 2: Faster Machine Learning Through Dynamic Python Bytecode Transformation and Graph Compilation2024
- BitSET: Bit-Serial Early Termination for Computation Reduction in Convolutional Neural NetworksACM Transactions on Embedded Computing Systems, 2023
- Egeria: Efficient DNN Training with Knowledge-Guided Layer Freezing2023
- Stochastic rounding: implementation, error analysis and applicationsRoyal Society Open Science, 2022
- Efficient large-scale language model training on GPU clusters using megatron-LM2021
- WinoGrandeCommunications of the ACM, 2021
- ZeRO: Memory optimizations Toward Training Trillion Parameter Models2020
- HAQ: Hardware-Aware Automated Quantization With Mixed Precision2019
- High-Dimensional Probability2018
- Neural Information Processing2017