hatsunexym opened a new issue #6120: URL: https://github.com/apache/incubator-tvm/issues/6120
I successfully used assert command by Python to make sure the **tvm_output** is similar to the **torch_output** with 1e-3 when using TVM to do model transform. Then I deploy the **param, graph and lib** on my C++ SDK, and use the same input to test the result. But the output of C++ is too different from the original tvm_output on python. (e.g: the tvm_output on python is [0.7237, 0.4792, ..., 0.6381], while the C++ tvm_output is [0.5127, 0.2638, ..., 0.2417]) MY QUES: Why the C++ output is different from Python's Thx ---------------------------------------------------------------- This is an automated message from the Apache Git Service. To respond to the message, please log on to GitHub and use the URL above to go to the specific comment. For queries about this service, please contact Infrastructure at: us...@infra.apache.org
