Fast Matrix Multiplications for Lookup Table-Quantized LLMs / FLUTE (Elhoushi et al., Microsoft)
paper
Fields
id
sources/qlab-flute-paperPublisher
arXiv
Kind
paper
Published
2024-07-15
Accessed
2026-10-03
notes
EMNLP 2024 Findings. LUT engine for 3-bit / non-uniform groupwise weight-only quantization; group size 128 at batch < 32.
Notes
EMNLP 2024 Findings. LUT engine for 3-bit / non-uniform groupwise weight-only quantization; group size 128 at batch < 32.
Cross-references
This record references no other records.