https://github.com/bunneyster/tots

Tensors of the Storm!
https://github.com/bunneyster/tots

heroes-of-the-storm hots machine-learning tensorflow

Last synced: 7 days ago
JSON representation

Tensors of the Storm!

• Host: GitHub
• URL: https://github.com/bunneyster/tots
• Owner: bunneyster
• Created: 2018-01-01T06:32:16.000Z (about 7 years ago)
• Default Branch: master
• Last Pushed: 2018-01-14T08:16:09.000Z (about 7 years ago)
• Last Synced: 2024-12-06T07:29:45.421Z (about 2 months ago)
• Topics: heroes-of-the-storm, hots, machine-learning, tensorflow
• Language: Python
• Size: 1.8 MB
• Stars: 0
• Watchers: 0
• Forks: 0
• Open Issues: 0
• Metadata Files:
  • Readme: README.md

Awesome Lists containing this project

README

        
## Model #1

### Parameters

* training dataset size: 235,738

* testing dataset size: 28,124

* hidden units: [100, 75, 50, 25]

### Results

* accuracy: 0.610902

* accuracy_baseline: 0.5

* auc: 0.657483

* auc_precision_recall: 0.644627

* average_loss: 0.655231

* global_step: 235882

* label/mean: 0.5

* loss: 26.1943

* prediction/mean: 0.524407

## Model #2

### Parameters

* training dataset size: 800,000

* testing dataset size: 88,888

* hidden units: [23, 13, 3]

### Results

* accuracy: 0.583858

* accuracy_baseline: 0.5

* auc: 0.619257

* auc_precision_recall: 0.611224

* average_loss: 0.673288

* global_step: 21142445

* label/mean: 0.5

* loss: 13.4655

* prediction/mean: 0.499873

## Model #3

Using just my own replays.

### Parameters

* total replays: 782

  * training dataset size: 1,404

  * testing dataset size: 160

* hidden units: [23, 13, 3]

### Results

* accuracy: 0.5875

* accuracy_baseline: 0.5

* auc: 0.589453

* auc_precision_recall: 0.625588

* average_loss: 0.722218

* global_step: 589682

* label/mean: 0.5

* loss: 1.44444

* prediction/mean: 0.486236