General Filter & Portrate Mode Filter - train model [MVP]

[XiaoSong9905]

Info

1. each model corrsponding to a specific filter
2. trained model weight is stored in local computer, model is packed into servable and upload to zenhub. See "pack model into servavle card"
3. *Train model do not have any code dependency since model weight are not upload to github. In order to met with the github grading rubrics, a log file for training process is upload to `src/backend/TrainLog/` for grading purpose**

Progress

1. Done

Link to Feature

1. Code https://github.com/DynamicProgrammingEECS441/PicassoXS/tree/master/src/backend/train
2. Model weight : model weigth are too large to be put on github, it's packed into tf server and upload to docker hub directly

List of Current Support Model

1. arbitary_style
2. francoise
3. kirchner
4. munch
5. roerich 6.cezanne
6. gauguin
7. monet
8. paul
9. van-gogh
10. el-greco
11. kandinsky
12. morisot
13. peploe

User Story & Acceptance Criteria

General Filter: https://docs.google.com/document/d/1YU66XGAY_96yoz0vlgMPM-049zOeC4QVQgMvJiivYsE/edit?usp=sharing

Portrate Mode Filter : https://docs.google.com/document/d/1BMszGmVqfQa7ZOoLGFx_vB3N_HeCxCEe_6KgkWVSmRo/edit?usp=sharing

Additional Acceptance Criteria:

1. model generate decent result
2. weight can be used to packed into tf server

[XiaoSong9905]

FOR GRADER, LOOK HERE ----------↓↓↓-----------

Train model do not have any code dependency since model weight are not upload to github. In order to met with the github grading rubrics, a log file for training process is upload to src/backend/TrainLog/* for grading purpose

----------↑↑↑-----------