search

justinpinkney / stable-diffusion

MIT License
1.45k stars 266 forks source link

Should we close DDP accelerator on while training with 1 GPU? #62

Open aartykov opened 1 year ago

aartykov commented 1 year ago

Hi,

Should we set the " trainer_config["accelerator"] = "ddp" " part to None while training with only 1 GPU?

WhiteZz1 commented 1 year ago

I ran the main.py script in only one GPU environment, and the following error occurred ValueError: You selected an invalid accelerator name: accelerator='ddp'. Available names are: cpu, cuda, hpu, ipu, mps, tpu.

Are our problems the some ? @artykov1511 @justinpinkney

justinpinkney commented 1 year ago

hmmm, this seems like maybe a pytorch lightning version issue? I used ddp should work fine for 1 gpu on the version i used