search

isl-org / lang-seg

Language-Driven Semantic Segmentation
MIT License
722 stars 91 forks source link

How long does it takes for default training? #31

Closed synsin0 closed 2 years ago

synsin0 commented 2 years ago

System: 4xRTX3090. Training scripts: Default training scripts: python -u train_lseg.py --dataset ade20k --data_path ./datasets --batch_size 1 --exp_name lseg_ade20k_l16 \ --base_lr 0.004 --weight_decay 1e-4 --no-scaleinv --max_epochs 200 --widehead --accumulate_grad_batches 2 --backbone clip_vitl16_384

My system has shown that one epoch takes for 45mins, which is a pretty long time for 200 epochs. Is that a normal procedure? Or we may not need max epochs like that?

Boyiliee commented 2 years ago

Hi @synsin0 ,

Thanks for your interest in LSeg!

As has been mentioned, we used 8 V100 and trained for 200 epochs. I guess, it is because of your batchsize=1, we use batchsize = 4 instead.

Hope this helps.

Best, Boyi