Machine learningOther
Efficient Post-training Quantization with FP8 Formats
A study finds that FP8 data formats are superior to INT8 in post-training quantization, offering better workload coverage, model accuracy, and versatility across various network architectures.
Featured in No. 43 on 3 Apr 2024 · · 51 citations today · published in Conference on Machine Learning and Systems
- Released
- 26 Sep 2023
- First featured
- No. 43 · 3 Apr 2024
- Citations (Semantic Scholar)
- 51
- Influential citations
- 6
- Published in
- Conference on Machine Learning and Systems
- Shares when featured
- 100
- Identifier
- arXiv:2309.14592
Citations and venue from Semantic Scholar (ODC-BY), refreshed weekly. Summary: Quant Letter (CC BY 4.0).