sm_120 capability

[nikubesliu]

Hello,
I'm getting these messages:

chatterbox-tts-server-1 | NVIDIA GeForce RTX 5070 Ti with CUDA capability sm_120 is not compatible with the current PyTorch installation.
chatterbox-tts-server-1 | The current PyTorch install supports CUDA capabilities sm_50 sm_60 sm_70 sm_75 sm_80 sm_86 sm_90.

Is there a known workaround (specific PyTorch/cu128 build, nightly, or flags) to run this project on sm_120 today?

If not, are there plans or a timeline to add official sm_120 support (and/or provide a Docker image pinned to a compatible Torch)?

Thanks! Happy to test any suggested build or branch.

[MarkJE]

Getting the same issue on the 5060 Ti. Can't promise any help apart from testing.

I tried working with Grok (as I am clueless here) to implement a workaround involving nightly, but couldn't get round the dependencies.