Display all input types on type error

Nobody has claimed this yet.

**Is your feature request related to a problem? Please describe.**
While trying to convert PyTorch model, I was frustrated trying to understand the types of input tensors after the conversion.
ONNX Runtime only shows an error for the specific input without displaying other inputs.

For example, I would get the following output:
```
onnxruntime.capi.onnxruntime_pybind11_state.InvalidGraph: [ONNXRuntimeError] : 10 : INVALID_GRAPH : This is an invalid model. Type Error: Type 'tensor(int32)' of input parameter (3230) of operator (Clip) in node () is invalid.
```
And from it, I couldn't understand which type is expected and why.

**System information**
- ONNX Runtime version (you are using): 1.3.0

**Describe the solution you'd like**
A clear indication of the expected type and explanation would be great.

**Describe alternatives you've considered**
It will help even if the error shows only types for every input node.

**Additional context**
For now, the only workaround I found is to split the graph and run every part separately. It is time consumable and error-prone method from my perspective.

The issue does not identify files, tests, or an entry point. Start by locating ONNX Runtime's invalid-graph type-error reporting and reproduce the PyTorch conversion example; done means the error identifies the expected type and provides the relevant input types clearly.

Written by the indexing model from the issue text.