Compare the GPU memory requirements between TorchSharp-phi and Microsoft.ML.OnnxRuntimeGenAI

[GeorgeS2019]

No description provided.

[LittleLittleCloud]

@GeorgeS2019 thanks for the feedback. We are working on migration this project to Microsoft.ML.GenAI.Phi and we will add the GPU memory comparsion between this package and phi model in onnxruntime package afterwards

[GeorgeS2019]

I have problems with using GPU for onnxruntime, onnxruntime-training and GenAI

The requirements are not clear and the examples provided not updated

All examples focus only on cpu, making using GPU unclear

[LittleLittleCloud]

For the GenAI, are you referring to this sample? If that case would you mind sharing the error log you have, I'm happy to help out.