Automate/ skip Aug_Strength variable when resuming from previous pkl

[js20222022]

Derrick - (Sorry if this isn't in the right place, this is my first time submitting a PR... i asked about this feature on the Slack channel and you mentioned to start a PR).

I noticed when training multiple models for a number of weeks - retyping the aug_strength (from the log.txt) gets somewhat redundant. I ask to please automate this step. I'm not sure if this value is recorded in the pkl file itself or not. But whatever the value is , if one just wants to resume and keep training the model - it would be much more stremlined if you can just put in the pkl file path you want to resume from and start training.

Thank you so much for your help!

resume_from = 'content.pkl'
aug_strength = 1.374 (<<Skip this step, or automate)
train_count = 100
mirror_x = False
mirror_y = False