Can you add Q4 families?
#1
by sdve - opened
Could you also upload the Q4 quantized versions shown in the quantization perplexity graph you provided?
prithivMLmods changed discussion status to closed
@prithivMLmods why is the old Q4 is deleted? can you upload it again?
https://huggingface.co/prithivMLmods/chandra-ocr-2-GGUF/commits/main
Hey @uptonking ,
I noticed some bugs in it.
So I’ll revise and upload soon.
Q8 → F32
Works fine as usual.