Closed gngdb closed 9 years ago
Train a net based on the maxout cigar yaml. Tune learning rate and analyse results in notebook. Submit results to the kaggle leaderboard.
Test results weren't good, so probably not going to submit this model.
Log loss was ~3 and the model was not trained on 10 classes either.
Train a net based on the maxout cigar yaml. Tune learning rate and analyse results in notebook. Submit results to the kaggle leaderboard.