I am compiling a custom model to port into ZCU102,
After succesfully calibrating the quantization with quant-mode 1
The vai_q_pytorch tool results error in quant_mode 2
The following trace of the error
(pytorch-xilinx) htic@htic-endo:/workspace/SRHW$ python quantize.py --quant_mode 2
[NNDCT_NOTE]: Loading NNDCT kernels...
---------------------- Model Loaded ----------------------
---------------------- Trace Started ----------------------
---------------------- End of trace ----------------------
---------------------- Start SRHW Quantization ----------------------
---------------------- Dataset Loaded ----------------------
[NNDCT_NOTE]: Quantization test process start up...
[NNDCT_NOTE]: =>Parsing SRHW...
[NNDCT_NOTE]: =>Quantizable module is generated.(quantize_result/SRHW.py)
---------------------- Eval Started ----------------------
. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
---------------------- Eval Complete ----------------------
[NNDCT_NOTE]: =>Converting to xmodel ...
sh: 1: dot: not found
WARNING: Logging before InitGoogleLogging() is written to STDERR
F1201 08:26:19.465744 32177 graph_imp.cpp:277] [UNILOG][Check Failed: result == 0][/proj/rdi/builds/vitis-ai/Vitis-AI-1.2/XIR/1.2/12/src/xir/graph/graph_imp.cpp:277][XIR_OPERATION_FAILED][Fail to execute command!] dot -Tsvg /workspace/SRHW/quantize_result/SRHW_int.svg.dot -o /workspace/SRHW/quantize_result/SRHW_int.svg
*** Check failure stack trace: ***
Aborted (core dumped)
I could not interpret whats happening here.
How to resolve this issue?
Note: I created a new environment and installed all libraries as described in the following link
https://forums.xilinx.com/t5/AI-and-Vitis-AI/Pytorch-Illegal-instruction-core-dumped/td-p/1177479
I am waiting for a reply
Thanks in advance
