Fix requirements

[erickgalinkin]

Change Requirements.txt to requirements.txt; fix incompatible requirement versions.

The provided Requirements.txt file contains incompatible dependencies. This PR fixes those dependencies and updates a number of libraries to their current versions. It also loosens the dependencies for tensorflow and related libraries. I've also changed requirements.txt to be all lowercase, as is usually expected for Python packages.

Note that several tests fail, including the test for the fixed Remove function that now seems to always return False (or more appropriately, TrinaryEnum.FALSE). While I don't believe these changes in package version should be breaking, I have no good way to validate. I've included the short test summary below:

============================================================================================================================================= short test summary info =============================================================================================================================================
FAILED test_sim/test_Actions/test_BlueActions/test_blue_remove.py::test_blue_remove_killchain_sim_expected_output[1] - assert <TrinaryEnum.FALSE: 3> == True
FAILED test_sim/test_Actions/test_BlueActions/test_blue_remove.py::test_blue_remove_killchain_sim_expected_output[22447] - assert <TrinaryEnum.FALSE: 3> == True
FAILED test_sim/test_Actions/test_BlueActions/test_blue_remove.py::test_blue_remove_killchain_sim_expected_output[3242] - assert <TrinaryEnum.FALSE: 3> == True
FAILED test_sim/test_Actions/test_BlueActions/test_blue_remove.py::test_blue_remove_killchain_sim_expected_output[4] - assert <TrinaryEnum.FALSE: 3> == True
FAILED test_sim/test_Actions/test_MSFActions/test_UpgradeToMeterpreter.py::test_upgrade_msf_shell - assert 1.0 == 0
FAILED test_sim/test_Actions/test_MSFActions/test_UpgradeToMeterpreter.py::test_interesting_sim_behaviour - assert 1.0 == 0
FAILED test_sim/test_Actions/test_MSFActions/test_ms17_010_killchain.py::test_MS17_010_killchain_sim_expected_output - assert 1.0 == 0
FAILED test_sim/test_sim_Cyborg.py::test_agent_train[Scenario1b] - AttributeError: 'ExploitRemoteService' object has no attribute 'sub_action'
FAILED test_sim/test_sim_Cyborg.py::test_training_example[Scenario1b] - AttributeError: 'ExploitRemoteService' object has no attribute 'sub_action'
FAILED test_sim/test_wrappers/test_wrappers.py::test_flat_fixed_wrapper_observation_space[Scenario1b] - AttributeError: 'ExploitRemoteService' object has no attribute 'sub_action'
====================================================================================================================== 10 failed, 1022 passed, 117 skipped, 2 warnings in 111.87s (0:01:51) =======================================================================================================================

[erickgalinkin]

It looks like the test_blue_remove test failures are largely due to randomness -- perhaps something isn't being fully controlled, as I can successfully run the Remove action in the tutorial (with some light modifications)