catan2001

Performance drop on RX 9070XT with BERT Large

3

# Issue with Bert Large RX 9070XT While running benchmarks with the [MLPerf BERT reference implementation](https://github.com/mlcommons/inference/tree/master/language/bert), I noticed unusual behavior on the RX 9070XT. Compared to the other GPUs I...

# Llama 3.1 8b Server and SingleStream Scenario fail I observed that **Server** and **SingleStream** scenarios fail in the **MLPerf** reference implementation, while the **Offline** scenario runs fine. This appears...

[Issue]: ResNet50 1.5v Accuracy Drop on Radeon GPUs

3

### Problem Description # ResNet 50 1.5 Accuracy drop on Radeon When running the **MLPerf Inference** benchmark (**ResNet50**, **ONNX Runtime**, **reference implementation**) on several AMD GPUs (**gfx1031**/RX 6700XT, **gfx1100**/RX 7900XT/7900XTX,...