[webgpu] Enable graph capture

[qjia7]

This PR enables graph capture capabilities in the WebGPU provider, which is similar with jsep one #18989.

All limitations are similar with JS/CUDA EP:

1. Models with control-flow ops (i.e. If, Loop and Scan ops) are not supported.
2. Usage of graph capture is limited to models where-in all ops in the model can be partitioned to the WebGPU EP or CPU EP and no memory copy between them.
3. Shapes of inputs/outputs cannot change across inference calls.
4. IOBinding is required. And all inputs/outputs are pre-allocated gpu buffers.

[qjia7]

To reuse the captured buffers, we need to change BufferManager::Create/TryAcquireCachedBuffer to add one additional parameter session_id so that we can check whether this session has been captured. If yes, we can get buffer from captured_buffers_ to reuse the buffer in TryAcquireCachedBuffer. How do you think? @fs-eire