apple / ml-quant

Research publication code for "Least Squares Binary Quantization of Neural Networks"

Summary

		Main Code: 1,874 LOC (30 files) = PY (98%) + TOML (1%) Secondary code: Test: 1,223 LOC (23); Generated: 0 LOC (0); Build & Deploy: 0 LOC (0); Other: 1,793 LOC (36);
		Duplication: 10%
		File Size: 0% long (>1000 LOC), 74% short (<= 200 LOC)
		Unit Size: 0% long (>100 LOC), 84% short (<= 10 LOC)
		Conditional Complexity: 0% complex (McCabe index > 50), 91% simple (McCabe index <= 5)
		Logical Component Decomposition: primary (8 components)
		Goals: Keep the system simple and easy to change (4)

generated by sokrates.dev (configuration) on 2022-01-30