search

NVlabs / few-shot-vid2vid

Pytorch implementation for few-shot photorealistic video-to-video translation.
Other
1.79k stars 276 forks source link

What are the best configurations to run the "few-shot-vid2vid" model #87

Open mamunctg opened 2 years ago

mamunctg commented 2 years ago

Could anyone please inform me the best configurations to train the model ? I try to use the following configurations but it shows errors for apex package. The apex packege can be solve if i use torch version = 1.6 but in the same time ,the flownet shows errors.

I used following configurations: python3.6, pytorch 1.2.0, torchvision 0.4.0, cuda 10.0, gcc4.9 in ubuntu-20.

##################ERROR######

(few_vid2vid) @ubuntu-:$ python train.py --name face --dataset_mode fewshot_face --adaptive_spade --warp_ref --spade_combine --batchSize 8 /home/*/anaconda3/envs/few_vid2vid/lib/python3.6/site-packages/apex-0.1-py3.6.egg/apex/pyprof/init.py:5: FutureWarning: pyprof will be removed by the end of June, 2022 warnings.warn("pyprof will be removed by the end of June, 2022", FutureWarning) Traceback (most recent call last): File "train.py", line 13, in from models.models import create_model File "/home//few-shot-vid2vid/models/models.py", line 11, in from models.networks.sync_batchnorm import DataParallelWithCallback File "/home//few-shot-vid2vid/models/networks/init.py", line 13, in from models.networks.discriminator import * File "/home/**/few-shot-vid2vid/models/networks/discriminator.py", line 14, in from models.networks.architecture import actvn as actvn File "/home//few-shot-vid2vid/models/networks/architecture.py", line 13, in from models.networks.normalization import SPADE, SynchronizedBatchNorm2d File "/home//few-shot-vid2vid/models/networks/normalization.py", line 15, in from apex.parallel import SyncBatchNorm as SynchronizedBatchNorm2d File "/home/*/anaconda3/envs/few_vid2vid/lib/python3.6/site-packages/apex-0.1-py3.6.egg/apex/init.py", line 24, in from . import transformer File "/home//anaconda3/envs/few_vid2vid/lib/python3.6/site-packages/apex-0.1-py3.6.egg/apex/transformer/init.py", line 1, in from apex.transformer import amp File "/home//anaconda3/envs/few_vid2vid/lib/python3.6/site-packages/apex-0.1-py3.6.egg/apex/transformer/amp/init.py", line 1, in from apex.transformer.amp.grad_scaler import GradScaler File "/home//anaconda3/envs/few_vid2vid/lib/python3.6/site-packages/apex-0.1-py3.6.egg/apex/transformer/amp/grad_scaler.py", line 21, in class GradScaler(torch.cuda.amp.GradScaler): AttributeError: module 'torch.cuda' has no attribute 'amp'

wangjinbo0929 commented 2 months ago

Did you fix this issue?I'm also getting the issue of AttributeError: module 'torch.cuda' has no attribute 'amp'

wangjinbo0929 commented 2 months ago

Did you fix this issue?I'm also getting the issue of AttributeError: module 'torch.cuda' has no attribute 'amp' @mamunctg.Thank you