Pretrain, finetune ANY AI model of ANY size on 1 or 10,000+ GPUs with zero code changes.
ci: move torch-tensorrt to GPU-only test requirements (#21575)
* ci: move torch-tensorrt dependency to GPU-only requirements torch-tensorrt is GPU-only (requires CUDA) and all its tests use `@RunIf(min_cuda_gpus=1)`, so installing it in CPU CI provides no value while blocking PyTorch version upgrades due to wheel compatibility lag. * ci: add GPU-specific test requirements to setup
B
Bhimraj Yadav committed
283ce7733ec23786d41751f185093eac83c0ef8d
Parent: c05cadb
Committed by GitHub <[email protected]>
on 3/9/2026, 4:18:56 PM