Tung D. Le

@sorenlassen thanks for the investigation! If `ends` is not a constant we cannot come to the code in ConstProp.cpp, since we have a guard in ConstProp.td: ``` def SliceofConst :...

I examined the model and found that the problem was because of that `ends` was produced during the constant propagation, so it is a constant linked with a buffer. That...

@Connor-XY please feel free to reopen this issue if you still see the error. Thanks!

@zincnode constant propagation is to do computation at compile time for operations whose inputs are constant. So for big inputs , it would take time because constant propagation is done...

@yaochengji > Optionally decompose coarse-grained ONNX operations to fine-grained ONNX-operations FYI, we do the same thing in onnx-mlir, some ONNX operations will be decomposed into other ONNX operations: https://github.com/onnx/onnx-mlir/blob/main/src/Transform/ONNX/Decompose.td. Currenty,...

With `-mcpu=z14` I think compilation will take more time since more instructions are generated for simd. Also, these numerical tests often use a small range for tensor size, there might...

Could we define an unknown element type, say `any`, and use `tensor`? Anyway, I found a thread related to this: https://llvm.discourse.group/t/rfc-tensors-with-unknown-element-types/1039/24

> Technically, if we had transpose(binary-op(transpose(a, X), transpose(b, X), Y) where type of a and b are identical and binary op is an element-wise operation, then we can also do...

> Remember that in practice if you have a variable x that's declared unsigned, then x < 0 will never return true even if you do x = -5 (which...

@gongsu832 thanks! This is the driver code, say, `c-driver.cpp`: ```c #include "OnnxMlirRuntime.h" #include extern "C" OMTensorList *run_main_graph(OMTensorList *); extern "C" OMTensor *omTensorCreateWithOwnership(void *, int64_t *, int64_t, OM_DATA_TYPE, int64_t); extern "C"...