-
Notifications
You must be signed in to change notification settings - Fork 274
Description
my cuda 12.4,
pytorch3d 0.7.8
torch 2.5.1+cu124
torchaudio 2.5.1
torchvision 0.20.1+cu124
use text_to_3d_lite_separately, after generating views.jpg:
DEFAULT_RENDERING_KWARGS
{'ray_start': 'auto', 'ray_end': 'auto', 'box_warp': 1.2, 'white_back': True, 'disparity_space_sampling': False, 'clamp_mode': 'softplus', 'sampler_bbox_min': -0.6, 'sampler_bbox_max': 0.6}
SVRMModel has 458.69 M params.
Load model successfully
view2mesh model {'Total': 458688965, 'Trainable': 0}
Hunyuan3D-1\infer\utils.py:74: FutureWarning: torch.cuda.amp.autocast(args...) is deprecated. Please use torch.amp.autocast('cuda', args...) instead.
with autocast():
outputs\test
Hunyuan3D-1\svrm\predictor.py:144: FutureWarning: torch.cuda.amp.autocast(args...) is deprecated. Please use torch.amp.autocast('cuda', args...) instead.
with torch.cuda.amp.autocast():
=====> Triplane forward time: 47.709824562072754
Error occurred during execution.
how should i do?
Thanks!