❓ [Question] Support for CUDA 11.6?

[alercelik]

Does latest version support CUDA 11.6❓

Pytorch officially supports CUDA 11.6, however docs say torch_tensort supports CUDA 11.3 at max. But in some issues it is said that CUDA version 11.6 is used. Is CUDA 11.6 officially supported by torch_tensorrt?

Environment

• PyTorch Version (e.g., 1.0): any
• CPU Architecture:
• OS (e.g., Linux): Linux
• How you installed PyTorch (conda, pip, libtorch, source): pip
• Build command you used (if compiling from source):
• Are you using local sources or building from archives:
• Python version: 3.8 or 3.9
• CUDA version: 11.6
• GPU models and configuration:
• Any other relevant information:

[narendasan]

Do you mean CUDA 11.6?

[alercelik]

Sorry, yes. CUDA 11.6

Correcting the issue for further reads.

[narendasan]

This will come with Torch-TensorRT 1.2.0 which targets PyTorch 1.12.0