Reproducible experiments #13

New Issue

gitlab-migration · 2023-07-15T14:44:07Z

gitlab-migration commented

2023-07-15 14:44:07 +00:00

Changes required

Take board representation as parameter in main.py
Take learning rate as parameter in main.py
Initial weights should be the same when testing hyperparameters such as learning rate and board representation
There should be a baseline model for our networks, thus allowing all of the models to start at the same place. This will lead to us easily being able to compare
models.

TODO:

Extend train-evaluate-save to take learning rate etc. as parameter

Migrated from GitLab. Created by christoffermadsen

Changes required * [ ] Take *board representation* as parameter in `main.py` * [ ] Take *learning rate* as parameter in `main.py` * [ ] Initial weights should be the same when testing hyperparameters such as *learning rate* and *board representation* * [x] There should be a baseline model for our networks, thus allowing all of the models to start at the same place. This will lead to us easily being able to compare models. TODO: * [ ] Extend `train-evaluate-save` to take learning rate etc. as parameter Migrated from GitLab. Created by christoffermadsen

gitlab-migration added the

Evaluation

label 2023-07-15 14:44:07 +00:00

christoffer was assigned by gitlab-migration

2023-07-15 14:44:07 +00:00

gitlab-migration closed this issue

2023-07-15 14:44:07 +00:00

Sign in to join this conversation.