This repository was archived by the owner on Aug 7, 2025. It is now read-only.

Description
🚀 The feature
TorchServe is currently using CUDA 11.7 by default.
Change the default version to CUDA 11.8 and upgrade all the test infra for the same.
Motivation, pitch
Maintenance
Alternatives
No response
Additional context
No response
### Tasks
- [ ] https://github.com/pytorch/serve/issues/2487
- [ ] https://github.com/pytorch/serve/pull/2489