Fix incompatible data type in DataFrame creation

[seroyer]

Do not specify a dtype when creating a DataFrame when the table has heterogeneous data types. In this case, the table has float and string types.

[seroyer]

This fixes an error when running --all benchmarks that looks like this:

Traceback (most recent call last):
  File "/workspace/jetson_benchmarks/benchmark.py", line 130, in <module>
    main()
  File "/workspace/jetson_benchmarks/benchmark.py", line 49, in main
    benchmark_table = pd.DataFrame(latency_each_model, columns=['GPU (ms)', 'DLA0 (ms)', 'DLA1 (ms)', 'FPS', 'Model Name'], dtype=float)
...
ValueError: could not convert string to float: 'inception_v4'