onnxruntime

2f4be9a1 - Add check that tensor sizes match in DataTransferManager::CopyTensors (#27008)

Add check that tensor sizes match in DataTransferManager::CopyTensors (#27008) ### Description <!-- Describe your changes. --> Add check that tensor sizes match in DataTransferManager::CopyTensors before calling the IDataTransfer implementation so that the check is done in one place. We check the sizes match in DataTransferManager::CopyTensor[Async] so this makes things consistent when a batched copy is done. It is not required for DataTransferManager::CopySparseTensors. The default implementation of IDataTransfer::CopySparseTensors is not overridden by any EP so all sparse tensor copies (single or batched) end up going via SparseTensor::Copy which has size checks. ### Motivation and Context <!-- - Why is this change required? What problem does it solve? - If it fixes an open issue, please link to the issue here. --> TRT RTX had a bug and was returning an output value that was an incorrect size. When pre-allocated outputs on a different device were provided we hit DataTransferManager::CopyTensors which had no check the sizes matched, leading to a heap checker violation.