additional model attributes

[triwahyuu]

closes #5

Type of changes

Please check the type of change your PR introduces: - [ ] Bugfix - [x] Feature - [ ] Code style update (formatting, renaming) - [ ] Refactoring (no functional changes, no api changes) - [ ] Build related changes - [ ] Documentation content changes - [ ] Other (please describe): ## What is the current behavior?

• close #5

Related issues :

What is the new behavior?

• update checkpoint model format
• support for resume training in training pipeline
• reading class_names attribute from checkpoint in prediction and export pipeline
• change class_names in dataset to optional
• adding model checkpoint update script in script/update_model.py

Checkpoint Model Format

default (required) member in checkpoint:

• epoch --> number of epochs the model have been trained
• state_dict -> trained model's state dict, this is the same as the entire data as the old format
• optimizer_state --> optimizer state in training
• config --> configuration used to train the model

additional member (if any):

• class_names --> model's output class names map
• metrics --> per epoch training metrics
• scheduler_state -> state_dict for trainer's lr scheduler

Checklist

• [x] Unit tests pass locally with my changes
• [x] I have added tests that prove my fix is effective or that my feature works
• [x] I have added necessary documentation (if appropriate)
• [x] Any dependent changes have been merged and published in downstream modules
• [x] I have filled the [Unreleased] Changelog

[triwahyuu]

ready for review cc @alphinside