Please explain the difference between the two models
#11
by
martjay
- opened
flux1-dev-Q4_0.gguf
flux1-dev-Q4_1.gguf
Please explain the difference between the two models
4.5bpw and 5pbw. Consequently, 4.1 has higher quality, but also requires more VRAM
On my test q4_1 and q5_1 act more like fp8 (while old q5_0 similar to q8 and act like fp16)
On my test q4_1 and q5_1 act more like fp8 (while old q5_0 similar to q8 and act like fp16)
I compared Q4_0 and Q4_1 and found I liked Q4_0 better, although it has lower bit width.