RuntimeError while executing Alexnet from Pytorch to Caffe2 example

Observed the following error while executing the example at - https://pytorch.org/docs/master/onnx.html

RuntimeError: [enforce fail at reshape_op.h:110] total_size == size. 92160 vs -1613423362598410240. Argument `shape` does not agree with the input data. (92160 != -1613423362598410240)Error from operator:
input: "29" input: "36" output: "37" output: "OC2_DUMMY_1" name: "" type: "Reshape" device_option { device_type: 1 cuda_gpu_id: 0 }
[F context_gpu.h:118] Check failed: error == cudaSuccess driver shutting down.