search

LiheYoung / UniMatch

[CVPR 2023] Revisiting Weak-to-Strong Consistency in Semi-Supervised Semantic Segmentation
https://arxiv.org/abs/2208.09910
MIT License
478 stars 60 forks source link

About GPUs #56

Closed ChengxiHAN closed 1 year ago

ChengxiHAN commented 1 year ago

Hi, bro How can I start my training for Number 2 and 3 GPU not from starting Number 0 GPU, if I set the of 2. I know the program is starting from the default GPU '0'.

Many thanks. Sigma

LiheYoung commented 1 year ago

Add environment variableCUDA_VISIBLE_DEVICES=2,3 before your command, such as:

CUDA_VISIBLE_DEVICES=2,3 sh scripts/train.sh 2 20024
ChengxiHAN commented 1 year ago

many thanks bro! Can I add your wechat bro ? (My wechat is: whuhanchengxi)

LiheYoung commented 1 year ago

Okay.