facebookresearch / diffq

DiffQ performs differentiable quantization using pseudo quantization noise. It can automatically tune the number of bits used per weight or group of weights, in order to achieve a given trade-off between model size and accuracy.

Summary

	PYX IN CFG
		Main Code: 1,000 LOC (13 files) = PY (92%) + PYX (6%) + IN (<1%) + CFG (<1%) Secondary code: Test: 386 LOC (5); Generated: 0 LOC (0); Build & Deploy: 0 LOC (0); Other: 2,676 LOC (26);
		Duplication: 9%
		File Size: 0% long (>1000 LOC), 57% short (<= 200 LOC)
		Unit Size: 0% long (>100 LOC), 68% short (<= 10 LOC)
		Conditional Complexity: 0% complex (McCabe index > 50), 66% simple (McCabe index <= 5)
		Logical Component Decomposition: primary (2 components)
		9 months old 0% of code older than 365 days 0% of code not updated in the past 365 days
		0% of code updated more than 50 times Also see temporal dependencies for files frequently changed in same commits.
		Goals: Keep the system simple and easy to change (4)

Commits Trend

Latest commit date: 2021-12-15

0 commits (30 days) 0 contributors (30 days)

Commits 66 Contributors 2 2021

show commits trend per language

Commits	2022 2021 27 py pyx 10 pyx in 1 in cfg cfg
Contributors	2022 2021 2 py pyx 2 pyx in 1 in cfg cfg