Explore Layer during this challenge! Layer will help you track and test your experiments.
Layer will be hosting another challenge soon on Zindi so get your head in the game now!
Create an account for free and get 30 GPU hours per week and unlimited CPU hours.
Layer helps you build, train and track all your machine learning project metadata including ML models and datasets with semantic versioning, extensive artifact logging and dynamic reporting with local↔cloud training.
First, install that you have the latest version of Layer installed:
pip install -U layer
The competition has 5 CSV files that can be fetched as Pandas DataFrames from Layer.
Fetch sample submission file
import layer sample_submission = layer.get_dataset("layer/Zindi-Laduma-Analytics/datasets/sample-submission").to_pandas() sample_submission.head()
Fetch training files
import layer train = layer.get_dataset("layer/Zindi-Laduma-Analytics/datasets/train").to_pandas() train.head() train_game_statistics = layer.get_dataset("layer/Zindi-Laduma-Analytics/datasets/train_game_statistics").to_pandas() train_game_statistics.head()
Fetch test files
import layer test = layer.get_dataset("layer/Zindi-Laduma-Analytics/datasets/test").to_pandas() test.head() test_game_statistics = layer.get_dataset("layer/Zindi-Laduma-Analytics/datasets/test_game_statistics").to_pandas() test_game_statistics.head()
What's the advantage (s) attached to using this layer library or ?
Layer allows you to track your experiments and gives you free GPU access.
Okay, really do not understand the tracking of experiment part!!! I know kaggle gives free limited GPU, colab does too and colab pro gives a larger reasonable amount of GPU and memory allocation. So what make layer different or special?
Hey @Emmanuel360__RAIN, thanks for the question.
This is Mehmet, from Layer. With experiment tracking, you can log anything during your experiments like metrics, plots, datasets even models. Layer automatically version these assets and provide an easy way to compare your experiments.
Here is an example ML model on Layer:
https://app.layer.ai/layer/image-segmentation/models/mask_predictor?w=2.1&w=1.1#val_loss
Looks cool, i created an account already, i just wanted to find out more info about it b4 i start exploring myself. Thanks