I followed the instructions. However, I need a clarification on the setup.
Clone this repo or download and extract Chatbot-with-RyzenAI-1.0.zip into the root directory where you have installed the Ryzen AI SW.
This means the "example\transformers\" subfolder in the RyzenAI-SW installation, right?
And for
Activate the conda environment created when RyzenAI was installed.
It is required to activate the conda environment created with the instructions for the transformers example, right?
I get the following error for step 3
I20240519 16:53:06.503293 7336 vitisai_compile_model.cpp:361] present.23.value : [-1x32x-1x64]
C:\tools\anaconda3\envs\ryzenai-transformers\lib\site-packages\voe\passes\fuse_GEMM.py:62: RuntimeWarning: invalid value encountered in divide
weight_q = np.clip(np.round(weight_f / weight_scale), -128, 127).astype(np.int8)
C:\tools\anaconda3\envs\ryzenai-transformers\lib\site-packages\voe\passes\fuse_GEMM.py:62: RuntimeWarning: invalid value encountered in cast
weight_q = np.clip(np.round(weight_f / weight_scale), -128, 127).astype(np.int8)
Couldn't find dll file libGemmQnnAie_8x2048_2048x2048.dll,
F20240519 16:53:07.404875 7336 pass_main.cpp:68] SystemExit: <EMPTY MESSAGE>
At:
C:\tools\anaconda3\envs\ryzenai-transformers\lib\site-packages\voe\passes\fuse_GEMM.py(100): action
C:\tools\anaconda3\envs\ryzenai-transformers\lib\site-packages\voe\rule_ext\__init__.py(81): _action
C:\tools\anaconda3\envs\ryzenai-transformers\lib\site-packages\onnxruntime\capi\onnxruntime_inference_collection.py(435): _create_inference_session
C:\tools\anaconda3\envs\ryzenai-transformers\lib\site-packages\onnxruntime\capi\onnxruntime_inference_collection.py(383): __init__
C:\tools\anaconda3\envs\ryzenai-transformers\lib\site-packages\optimum\onnxruntime\modeling_ort.py(369): load_model
C:\tools\anaconda3\envs\ryzenai-transformers\lib\site-packages\optimum\onnxruntime\modeling_decoder.py(536): _from_pretrained
C:\tools\anaconda3\envs\ryzenai-transformers\lib\site-packages\optimum\modeling_base.py(372): from_pretrained
C:\tools\anaconda3\envs\ryzenai-transformers\lib\site-packages\optimum\onnxruntime\modeling_ort.py(622): from_pretrained
c:\tools\RyzenAI-SW\example\transformers\run.py(106): <module>
[Vitis AI EP] No. of Operators : CPU 2106
I20240519 16:53:07.467839 7336 vitisai_compile_model.cpp:376] Catch fatal exception, skip this subgraph. Set XLNX_ENABLE_SKIP_FATAL=0 to stop skip.
attrs_create_cif```
There is a file libGemmQnnAie_8x2048_2048x2048.dll in the "RyzenAI-SW\example\transformers\dll\phx\qlinear\". There are multiple files in other folders under the dll folder. I tried to add them to the PATH, but I keep getting this issue.
The results is that it runs only on CPU.
I followed the instructions. However, I need a clarification on the setup.
This means the "example\transformers\" subfolder in the RyzenAI-SW installation, right?
And for
I get the following error for step 3