George S

Hi @demuxin - thanks for the report - we likely need to add `getter`/`setter` methods to toggle this value in C++ as well: https://github.com/pytorch/TensorRT/blob/07c5b07b5ca81165682be90049a6bc67cff9928e/core/runtime/runtime.cpp#L10 Similar to the following functions: https://github.com/pytorch/TensorRT/blob/07c5b07b5ca81165682be90049a6bc67cff9928e/core/runtime/register_jit_hooks.cpp#L121-L122...

These may actually already be accessible in C++, prior to inference, could you try adding the line: `torch_tensorrt::core::runtime::set_multi_device_safe_mode(true)`

Does `torch::ops::tensorrt::get_multi_device_safe_mode()` exist, or does this also cause a compilation error?

@HolyWu - thanks for this detail. I wanted to add the `release` specification to the nightly builds in advance of getting your PR #2806 merged, and to get this change...

Hi - thanks for the report. I think this may be related to the following lowering pass, where it's possible that both inputs are upcasted integers, so we accidentally construct...

Yes, this appears to be bug and we can work on a fix for this. Do you have a reproducer script or model we could use to recreate the error?

Hello - thanks for the details. I am unable to access the model at that link, is the model available elsewhere? Also, could you provide the full output debug log...