Math & Code Benchmark/Testing for GGUFs

#3
by bobchenyx - opened

Hi, thanks for releasing such great quantization versions.

I would like to ask if there are any open source frameworks/tools that could be used to test code/math benchmark for GGUF models?

Thanks!

Unsloth AI org

Hi, thanks for releasing such great quantization versions.

I would like to ask if there are any open source frameworks/tools that could be used to test code/math benchmark for GGUF models?

Thanks!

You could use elethur ai's lm harness

Hi, thanks for releasing such great quantization versions.

I would like to ask if there are any open source frameworks/tools that could be used to test code/math benchmark for GGUF models?

Thanks!

You could use elethur ai's lm harness

Did you use that framework to test these GGUF models?
Caused I'v tried before and they didn't seems to be working.
and they are truly slow.

Unsloth AI org

Hi, thanks for releasing such great quantization versions.

I would like to ask if there are any open source frameworks/tools that could be used to test code/math benchmark for GGUF models?

Thanks!

You could use elethur ai's lm harness

Did you use that framework to test these GGUF models?
Caused I'v tried before and they didn't seems to be working.
and they are truly slow.

Yes we did. However they did not match official MMLU scores so we needed to make our own custom evaluation framework

Sign up or log in to comment