cant convert multiple inputs model to be flexible

## multiple inputs to be flexible error

- when convert multiple inputs to be flexible, some error comes
'NotImplementedError: Image output 'output_img' has symbolic dimensions in its shape'
- is there any sample of multi input model to be flexible, here is my code
```
input1_shape = ct.Shape(shape=(1,
256,
ct.RangeDim(lower_bound=16, upper_bound=136, default=52),
ct.RangeDim(lower_bound=16, upper_bound=240, default=120)))
input1_shape = ct.Shape(shape=(1,
128,
ct.RangeDim(lower_bound=4, upper_bound=34, default=17),
ct.RangeDim(lower_bound=4, upper_bound=60, default=30)))
input_1=ct.TensorType(name="input_1", shape=input_1_shape)
input_2=ct.TensorType(name="input_2", shape=input_2_shape)
outputs=ct.ImageType(name="output_img",color_layout=ct.colorlayout.RGB)
mlmodel = ct.convert(
trace_model,
inputs=[input_1, input_2],
outputs=[outputs],
)
```

Start with the ct.convert call in the issue, focusing on the two flexible TensorType inputs and the ImageType output that raises the symbolic-dimensions NotImplementedError. Reproduce the conversion with the supplied shapes, then inspect the converter path handling multiple flexible inputs and image outputs; done means the behavior is fixed or a documented limitation and working multi-input example are provided.

Written by the indexing model from the issue text.