search

artidoro / qlora

QLoRA: Efficient Finetuning of Quantized LLMs
https://arxiv.org/abs/2305.14314
MIT License
9.96k stars 820 forks source link

question about quantization #243

Open fxb392 opened 1 year ago

fxb392 commented 1 year ago

image

Does absmax(XFP32) represent the maximum value of current input x or the maximum value of the entire FP32 data type？