[Bug] Error converting operator Resize: Mismatched type on argument

[coffezhou]

Expected behavior

TVM should run the model correctly.

Actual behavior

For the following model,

it can be executed by onnxruntime, the results are as follows:

ONNXRuntime:
 [array([[[[0.6251511 , 0.6180997 , 0.6039968 , 0.4824833 , 0.25355908,
          0.17335546, 0.2418724 , 0.27613088],
         [0.47064707, 0.46871254, 0.46484345, 0.37809238, 0.20845926,
          0.15066189, 0.20470022, 0.23171937],
         [0.16163902, 0.16993824, 0.18653667, 0.16931047, 0.11825964,
          0.10527475, 0.13035582, 0.14289635],
         [0.34176996, 0.3465182 , 0.3560147 , 0.42180583, 0.5438916 ,
          0.47625098, 0.21888396, 0.09020043],
         [1.0110399 , 0.9984524 , 0.9732775 , 1.1355784 , 1.4853553 ,
          1.2635906 , 0.4702846 , 0.0736316 ],
         [1.4519522 , 1.4102529 , 1.3268545 , 1.4443552 , 1.7627555 ,
          1.5264107 , 0.7353209 , 0.3397761 ],
         [1.6645073 , 1.5819199 , 1.4167454 , 1.3481362 , 1.3760922 ,
          1.2647111 , 1.013993  , 0.88863397],
         [1.7707846 , 1.6677535 , 1.4616909 , 1.3000267 , 1.1827606 ,
          1.1338614 , 1.153329  , 1.1630629 ]],

        [[1.2571493 , 0.9908935 , 0.4583817 , 0.2656508 , 0.41270074,
          0.7101021 , 1.1578549 , 1.3817313 ],
         [0.99612385, 0.79088306, 0.3804013 , 0.23428433, 0.35253203,
          0.6584027 , 1.1518965 , 1.3986433 ],
         [0.47407275, 0.390862  , 0.22444047, 0.17155136, 0.23219463,
          0.555004  , 1.1399796 , 1.4324675 ],
         [0.25609106, 0.2502444 , 0.23855111, 0.2727924 , 0.35296828,
          0.5743203 , 0.9368483 , 1.1181124 ],
         [0.34217882, 0.36903027, 0.4227332 , 0.53800744, 0.714853  ,
          0.7163514 , 0.5425026 , 0.4555782 ],
         [0.76455665, 0.7480468 , 0.7150272 , 0.7738706 , 0.9245769 ,
          0.8096743 , 0.4291628 , 0.23890704],
         [1.5232246 , 1.387294  , 1.1154332 , 0.98038185, 0.98213995,
          0.85428894, 0.5968289 , 0.46809882],
         [1.9025584 , 1.7069178 , 1.3156362 , 1.0836375 , 1.0109215 ,
          0.8765963 , 0.6806619 , 0.5826947 ]]]], dtype=float32)]

However, the onnx frontend of TVM cannot import it:

File "/home/carla/Documents/tvm/python/tvm/relax/frontend/onnx/onnx_frontend.py", line 3925, in from_onnx
    return g.from_onnx(graph, opset)
           ^^^^^^^^^^^^^^^^^^^^^^^^^
  File "/home/carla/Documents/tvm/python/tvm/relax/frontend/onnx/onnx_frontend.py", line 3556, in from_onnx
    self._construct_nodes(graph)
  File "/home/carla/Documents/tvm/python/tvm/relax/frontend/onnx/onnx_frontend.py", line 3736, in _construct_nodes
    raise err
  File "/home/carla/Documents/tvm/python/tvm/relax/frontend/onnx/onnx_frontend.py", line 3731, in _construct_nodes
    op = self._convert_operator(op_name, inputs, attr, self.opset)
         ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
  File "/home/carla/Documents/tvm/python/tvm/relax/frontend/onnx/onnx_frontend.py", line 3831, in _convert_operator
    sym = op_function(self.bb, inputs, attrs, [self._nodes, self._params])
          ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
  File "/home/carla/Documents/tvm/python/tvm/relax/frontend/onnx/onnx_frontend.py", line 2147, in _impl_v18
    return relax.op.image.resize2d(
           ^^^^^^^^^^^^^^^^^^^^^^^^
  File "/home/carla/Documents/tvm/python/tvm/relax/op/image/image.py", line 116, in resize2d
    return _ffi_api.resize2d(  # type: ignore
           ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
  File "tvm/ffi/cython/./function.pxi", line 228, in tvm.ffi.core.Function.__call__
TypeError: Mismatched type on argument #2 when calling: `relax.op.image.resize2d(0: ir.RelaxExpr, 1: ir.RelaxExpr, 2: Array<ir.FloatImm>, 3: ffi.String, 4: ffi.String, 5: ffi.String, 6: ffi.String, 7: float, 8: int, 9: float, 10: Optional<DataType>) -> ir.RelaxExpr`. Expected `Array<ir.FloatImm>` but got `relax.expr.Call`

Environment

OS: Ubuntu 20.04
TVM: 0.22.dev0 (c6969d7)
onnxruntime: 1.21.0

Steps to reproduce

This bug can be reproduced by the following code with the model in the attachment. As shown in the code, the model can be executed by onnxruntime. However, TVM cannot import this model.

import sys

import numpy as np
import onnx
import onnxruntime

import tvm
from tvm import relax
from tvm.relax.frontend.onnx import from_onnx

import pickle

            
def main():
    onnx_model = onnx.load("111.onnx")
    
    with open("inputs.pkl", "rb") as fp:
        inputs = pickle.load(fp)

    try:
        ort_session = onnxruntime.InferenceSession(
            onnx_model.SerializeToString(), providers=["CPUExecutionProvider"]
        )
        ort_output = ort_session.run([], inputs)
    except Exception as e:
        print(e)
        sys.exit(1)
        
    print("ONNXRuntime:\n", ort_output)   

    # Convert the onnx model into relax through the onnx importer.
    tvm_model = from_onnx(onnx_model, keep_params_in_input=True)
    

    
if __name__ == "__main__":
    
    main()

testcase.zip

Triage

Please refer to the list of label tags here to find the relevant tags and add them below in a bullet format (example below).

• needs-triage