How to install Torch version that supports RTX 5090 on Windows? - CUDA kernel errors might be asynchronously reported at some other API call #146977
Description
I have purchased RTX 5090 just to test AI apps
Currently getting this error on any app
I need torch for Python 3.10 venv on Windows
I am ok with installing nightly version etc just install command please
Traceback (most recent call last):
File "E:\trellis_v5\TRELLIS\app.py", line 401, in <module>
pipeline = TrellisImageTo3DPipeline.from_pretrained("JeffreyXiang/TRELLIS-image-large")
File "E:\trellis_v5\TRELLIS\trellis\pipelines\trellis_image_to_3d.py", line 56, in from_pretrained
pipeline = super(TrellisImageTo3DPipeline, TrellisImageTo3DPipeline).from_pretrained(path)
File "E:\trellis_v5\TRELLIS\trellis\pipelines\base.py", line 39, in from_pretrained
_models = {
File "E:\trellis_v5\TRELLIS\trellis\pipelines\base.py", line 40, in <dictcomp>
k: models.from_pretrained(f"{path}/{v}")
File "E:\trellis_v5\TRELLIS\trellis\models\__init__.py", line 59, in from_pretrained
model = __getattr__(config['name'])(**config['args'], **kwargs)
File "E:\trellis_v5\TRELLIS\trellis\models\structured_latent_vae\decoder_mesh.py", line 105, in __init__
self.mesh_extractor = SparseFeatures2Mesh(res=self.resolution*4, use_color=self.rep_config.get('use_color', False))
File "E:\trellis_v5\TRELLIS\trellis\representations\mesh\cube2mesh.py", line 68, in __init__
verts, cube = construct_dense_grid(self.res, self.device)
File "E:\trellis_v5\TRELLIS\trellis\representations\mesh\utils_cube.py", line 11, in construct_dense_grid
vertsid = torch.arange(res_v ** 3, device=device)
RuntimeError: CUDA error: no kernel image is available for execution on the device
CUDA kernel errors might be asynchronously reported at some other API call, so the stacktrace below might be incorrect.
For debugging consider passing CUDA_LAUNCH_BLOCKING=1
Compile with `TORCH_USE_CUDA_DSA` to enable device-side assertions.
cc @ezyang @gchanan @zou3519 @kadeng @msaroufim @malfet @seemethere @peterjc123 @mszhanyi @skyline75489 @nbcsm @iremyux @Blackhex @ptrblck @eqy