mehrdadh commented on issue #13013:
URL: https://github.com/apache/tvm/issues/13013#issuecomment-1274973295

   @alanmacd I have tested this on AOT with C++ and it doesn't show any error. 
It might also be a problem with CRT.
   To reproduce the error, you can use this test: 
https://github.com/apache/tvm/blob/0b034d77bd611ca7642b7211fa34110a87271f25/tests/python/relay/aot/test_cpp_aot.py#L54
   
   If you apply this modification, then the test fails because the output is 
incorrect but you don't see any error regarding passing data to the wrong input 
name.
    
   ```
   diff --git a/tests/python/relay/aot/test_cpp_aot.py 
b/tests/python/relay/aot/test_cpp_aot.py
   index b67bc90d3..44d336d88 100644
   --- a/tests/python/relay/aot/test_cpp_aot.py
   +++ b/tests/python/relay/aot/test_cpp_aot.py
   @@ -90,7 +90,7 @@ def test_conv2d(enable_usmp, target_kind):
        weight_data = np.random.randint(1, 255, 
shape_dict["weight"]).astype(type_dict["weight"])
        input_data = np.ones(shape_dict["data"]).astype(type_dict["data"])
        params = {"weight": weight_data}
   -    inputs = {"data": input_data}
   +    inputs = {"data1": input_data}
        ref_outputs = generate_ref_data(ir_mod, inputs, params)
    
        with tvm.transform.PassContext(
   ```
   


-- 
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.

To unsubscribe, e-mail: [email protected]

For queries about this service, please contact Infrastructure at:
[email protected]

Reply via email to