AMDMIGraphX

AMDMIGraphX copied to clipboard

Currently if an onnx file is empty (i.e., contains 0 bytes), MIGraphX will throw a segfault. This is not a useful message to the user. We should try to have...

kahmed10

Get the MIGraphX EP up to parity with the ROCm EP so we can accelerate additional ops via MIGraphX - [ ] Determine deficiencies between EPs - [ ] Make...

TedThemistokleous

Switch to hipblaslt from rocblas

3

This issue is for adding functionality for switch to hipblaslt from rocblas.

ahsan-ca

Write a onnx parser/verify tests for FP8

3

Write a onnx parser test for FP8

ahsan-ca

Add propagate_precision pass

6

pfultz2

In https://github.com/ROCm/AMDMIGraphX/pull/2973, uint8 type has been enabled for MaxPool, however following test cases are failing: - AveragePool with padding and uint8/int8 type - AveragePool with stride that is not default...

nives-vukovic

Cleanup CI to import Onnxruntime prebuild wheel

5

Move our the building of Onnxrt from all our CI and perform this action nightly/when there's an update we pull in for onnxrt Potentially do the same for unit tests....

TedThemistokleous

MIGraphX execution provider for Triton Inference Server

65

Can this be done by leveraging the onnxruntime work we already have as a back end? As a preliminary step, learn to add a Cuda back end, then change it...

bpickrel

Update bert_base_cased squadV1 for int8 and mixed

2

Update test and add dlm pieces - [x] DLM piece - [x] Add change to test

TedThemistokleous

Support SoftmaxCrossEntropyLoss

1

This is a very common ONNX operator. ORT falls back to CPU when encountering this node. The code is at: https://github.com/microsoft/onnxruntime/blob/ee603ee3265dbf6eac112baf273b6b69bf696085/onnxruntime/core/providers/migraphx/migraphx_execution_provider.cc#L1022-L1029 This happens when executing LLama v2, from: https://github.com/microsoft/onnxruntime-inference-examples/blob/main/python/models/llama/LLaMA-2 E2E...

hgaspar