-
Based on recent discussions the pad-pack pipeline has allowed us to make progress but the long term solution seems to be that we want to further build the pack pipeline which is currently not e2e enab…
-
For this, we need to lower it from multiple models and multiple frameworks and see "what they do". At least two potential options: `exp(arg)` and `exp(arg-mean)`, might be more. We need to pick one, p…
-
### What happened?
When running the inference of a YOLOv5 model, which has been imported from ONNX and compiled for the CUDA backend, the execution does not terminate. It just keeps on running on the…
-
Here is the crash I am seeing for matmul with accumulate
https://gist.github.com/nirvedhmeshram/32a76d8676b8d4d3ff6e7bbc7a10d05d
related PR: https://github.com/nod-ai/iree-amd-aie/pull/119
-
Error:
```
➜ ninja all iree-test-deps && ctest -j32 --label-exclude '^driver=cuda|metal' --output-on-failure
[0/2] Re-checking globbed directories...
[53/53] Generating kernels_gfx1100.co
FAILED…
kuhar updated
5 months ago
-
I'm trying to install SHARK, but I keep getting these errors:
```
ERROR: Could not find a version that satisfies the requirement torch-mlir (from versions: none)
ERROR: No matching distribution f…
-
We're adding new automated test coverage for compiling and running several PyTorch models in https://github.com/iree-org/iree/pull/17129 and related PRs.
I have two overall goals here:
1. Provide …
-
There are a number of HIP functions that assume a selected device for the current thread and they operate on this device. For example `hipModuleLoadDataEx`.
We need to set the correct HIP device duri…
-
I think this task is already started. Once we have a second machine on the network, what are the steps to integrate it into the CI workflow? TBH the amount of time spent running tests on HW is quite l…
-
This gist https://gist.github.com/Max191/28827ebad403bb21500bc9dc3be37e47 contains 2 func ops with very similar IR, and the only difference being that `%extracted_slice` is used as the thread forall i…