nngen
nngen copied to clipboard
Fix input_scale_factors
I found a bug in tests/onnx_matrix_mul/onnx_matrix_mul.py. The outputs of torch model and np.eval does not match. So I fix the input_scale_factors as following.