Logan Adams

@Dingjz - any updates on trying with the Hybrid Engine disabled? We should have proper AMD scripts and directions up soon.

@Dingjz - that's good, we're continuing to work on this, will re-open and update this when we have more updates on AMD.

You'll need to have a torch version that matches the cuda version you have installed. This would be easiest to do by installing cuda 11.8 to match your torch-cuda version,...

@ggyggy666 - were you able to test if that worked for you?

@ggyggy666 - let us know if you have a chance to try that, or other questions on trying to get it to run on your machine, we're happy to help...

Closing as the issue appears to be resolved.

@Eichhof - you can pre-build the ops with the directions [here](https://www.deepspeed.ai/tutorials/advanced-install/#pre-install-deepspeed-ops). When you install deepspeed, `DS_BUILD_OPS=1 pip install deepspeed/./etc.` regardless of installing via pip or from source.

@Eichhof - if you're seeing issues with this, feel free to re-open this issue or open a new one with the specific issue you're seeing, but installing the ops should...

Hi @helloworld1 - does this only occur if you pre-compile the ops, could you let me know if this happens if you just use JIT to build the ops or...

@helloworld1 - thanks, I was able to repro on my side as well now. I'll take a look at fixing this.