search

Ree1s / IDM

298 stars 22 forks source link

About training on a gpus #28

Open hjdihs opened 11 months ago

hjdihs commented 11 months ago

So is that the start of training? But after an hour, how could it still be like this. How long does it take to train an epoch? 屏幕截图 2023-12-06 163622

23-12-06 07:29:46.861 - INFO: Model [DDPM] is created. 23-12-06 07:29:46.861 - INFO: Initial Model Finished 23-12-06 07:29:46.861 - INFO: Resuming training from epoch: 0, iter: 0.

404NautFond commented 10 months ago

I ran into the same problem. I followed the data prepare instruction of FFHQ and found that the dataloader can get no sr_img_bytes so that the process enter infinite loop at 144 line in LRHR_dataset.py. The solution might be running the prepare_data.py without -l parameter to avoid lmdb mode and get png files instead. Be reminded to comment break command at 145 line in prepare_data.py or it will just process one instance.