Logan Adams

Hi @H9990HH969 - You best bet in this case is probably to build the .whl from source and copy it to the machine without internet access and then install it...

Hi @xu-song, I'll try to repro this and take a look.

At a first glance, it looks like we should be handling that properly [here](https://github.com/microsoft/DeepSpeed/blob/8d53ac0cd3a708d1b9a4281b5e01044b5bab6d61/deepspeed/launcher/runner.py#L389). Could you try setting it this way and let me know if that works for you?...

@xu-song - thanks, I was able to repro and I'm taking a look at this.

@xu-song - if you using the Deepspeed launcher, this isn't supported, but you can specify the nodes this way: https://www.deepspeed.ai/getting-started/#resource-configuration-single-node 

Hi @bmedishe - there are two different triton versions in the requirements, one for sd that is the one you note, and the one for sparse_attn that is 1.0.0. Could...

@bmedishe - thanks, still working on getting these updated, will update this thread when it is complete.

@bmedishe - we need this specific version for now for stable diffusion for now unfortunately.

Hi @SkyAndCloud - the warning you are seeing comes from [here](https://github.com/microsoft/DeepSpeed/blob/master/op_builder/builder.py#L354). Specifically, this comes when the system installed cuda and torch cuda do not match each other. Are you running...

Are there other cuda installations on the system? Since something seems to be triggering that warning. Do you see either of the printouts from [this function ](https://github.com/microsoft/DeepSpeed/blob/master/op_builder/builder.py#LL85C19-L85C19)in your code too?