Automated Model Improvement Recommendations

The Roboflow Model Evaluation tool now makes recommendations on how to improve your model. These recommendations are based on the results from your model confusion matrix.

The Model Evaluation tool can make recommendations based on:

• How to improve a model that predicts many false negatives.

• How to improve a model that predicts many false positives.

• What classes are often confused (mis-identified).

• What classes need more data to improve accuracy.

• When a test or validation set may be too small.

• And more.

You can see this feature by clicking the "Model Evaluation" button on any model trained on or uploaded to Roboflow.