added GTests for Conv GPU batch>1 inference

[harz05]

Closes #18

This PR adds ONNX models with batch sizes 2, 4, and 8 (same architecture as ConvWithPadding, to test the batch loop and not the Conv op itself) along with reference outputs and TEST_F cases in TestCustomModelsFromONNXForAlpakaCuda.cxx.

Reference outputs are verified against PyTorch F.conv2d. Tested on Google Colab (NVIDIA T4, CUDA 12.x), all 60 tests pass with no regressions.

Steps to reproduce the test on Colab-

!apt-get install -y libprotobuf-dev protobuf-compiler libgtest-dev
!git clone https://github.com/harz05/SOFIE.git && cd SOFIE && git checkout feat/conv-batch-tests
!mkdir build && cd build && cmake -Dtesting=ON -DCMAKE_INSTALL_PREFIX=../install -DCMAKE_BUILD_TYPE=RelWithDebInfo -DENABLE_ALPAKA_TESTS=ON -DALPAKA_BACKEND=cuda .. && cmake --build . --target install -j$(nproc)
!ctest -V