-
Notifications
You must be signed in to change notification settings - Fork 19
logs3:How to save model
Higepon Taro Minowa edited this page Apr 16, 2018
·
1 revision
- Log 1: what specific output am I working on right now?
- We should define a way to store and saved train model.
- Log 2: thinking out loud - e.g. hypotheses about the current problem, what to work on next
- Information we need
- actual model files
- ipynb
- other train files
- hparams as summary
- Where to store?
- dropbox/dont_remove_models/20180414seq2seq/README.md
- naming rule
- Information we need
- Log 3: record of currently ongoing runs along with a short reminder of what question each run is supposed to answer
- todo
- save model files in dropbox
- save ipynb
- save hparams as text
- todo
- Log 4: results of runs (TensorBoard graphs, any other significant observations), separated by type of run (e.g. by environment the agent is being trained in)
- store it in https://mega.nz/.