Jiawei Zhang comments

Repositories
Issues
Comments

Results 3 comments of


                                            Jiawei Zhang

DDIM Timesteps Offset

I am also confused about this and met the same problem... I can not sample if I set the ddim_num_steps to 1000.

What is the variable "a" in def attack ?

you can check https://github.com/AI-secure/QEBA/blob/8ee4a586a1ac5b172da92cc1c8efa9f0e543269c/foolbox-based/foolbox/attacks/bapp_custom.py#L176 => then you can see a is the "input_or_adv", namely, the original correctly classified input.

【bug】Exception raised from c10_cuda_check_implementation at /opt/conda/conda-bld/pytorch_1720538435607/work/c10/cuda/CUDAException.cpp:43 (most recent call first):

same here, and i have tried `export VLLM_ATTENTION_BACKEND=XFORMERS` and `export VLLM_USE_V1=0`. Both did not work. But this error will not appear if i only use one gpu.