New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Jump to bottom

[Question] semantic description of different quantization methods #3088

Open

phgcha opened this issue Jan 9, 2025 · 0 comments

Labels

question

phgcha commented Jan 9, 2025

❓ General Questions

Hi,

I'd love to know about trends in different quantization methods supported by MLC
For example (I made this up) ,

slowest-fastest: q0f32, q3f16_0, q4f16_0, q4f16_1, q4f32_1, q4f16_2, q4f16_autoawq
least-most memory consumption: q0f32, q3f16_0, q4f16_0, q4f16_1, q4f32_1, q4f16_2, q4f16_autoawq
response quality scoring (highest-lowest) : q0f32, q3f16_0, q4f16_0, q4f16_1, q4f32_1, q4f16_2, q4f16_autoawq

It'd be also really nice if there is existing benchmark results (of any model on any platform)

I'm particularly interested in Llama3.1-8B model on Jetson AGX orin hardware.

The text was updated successfully, but these errors were encountered:

phgcha added the question label

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment