python train.py configs/unflow.yaml not working

[DeVriesMatt]

Great repo, thanks for this! However, there are a few small things causing python train.py configs/unflow.yaml to throw errors.

• Dataset path in configs/default.yaml points to owners personal directory (line 3 in configs/default.yaml)
• h5py .value call depreciated in dataset/Humans/scripts/write_sequence_to_obj.py (lines 56 and 57). Needs to be replaced with [()].
• exit_after = args.exit_after (line 75 in train.py) gives an error as no arguments called exit_after are defined in the arguments parser.