8/11/2023 0 Comments Memory clean 3 1.0.7![]() ![]() ![]() Improved categorical handling in the MLI decision tree surrogate model by using one-hot encoding to encode categoricals by default. Optimized the number of default explainers in MLI to reduce runtime and increase clarity.Īdded support for microseconds in MLI time-series explainer.Įnabled Shapley values for MLI TS when only a training dataset is used. Listing pages now retain previous values for search, pagination and sorting. When importing experiment on project page, user is automatically prompted to download its datasets.Īdded explanatory videos for several MLI explainers. Improved C++ MOJO performance under high CPU load.Īdded a new log tab to allow admin users access to internal services log files.Īdded support for HTTPS SSL key file when encrypted with a passphrase. Sped up BertModel and BertTransformer when data is text-dominated to avoid unnecessary validation repeats for small data. Reduced memory usage when making test set predictions.Īllow control over early_stopping_threshold (relative min_delta) for LightGBM.Īdded stronger overfit protection for recipe ( more_overfit_protection).Īdded support for unsupervised recipes that handle text columns. pkl files.Īutomatically toggle GPU ON/OFF in the experiment setup page based on whether models and transformers perform better on (or must use) GPUs. Prediction frames now contain the original target column name in case the target column name contains special characters that require sanitization.Īdded the ability to ingest pandas sparse frames for pandas. (That is, the number of newly updated historic values for each time period.) Python scorers for lag-based time-series models now keep the target column in the frame to allow test-time augmentation.ĭetails about test-time augmentation are now provided in logs. The health API flag is_idle has been updated to account for large datasets being uploaded from a browser session. You can now navigate to an experiment detail in the H2O MLOps app from the experiment page. ![]() Shift detection is now performed on the final model’s transformed features and target to check generalization.Īdded an Experiment Results Wizard (beta): Shows several details for a given finished experiment.Īdded the Experiment Comparison Wizard for easy comparison of expert settings and knobs.Īdded training data column stats JSON file to scorers (MOJO and Python).Ĭontrol runtime more accurately using runtime estimation.Īdded estimation of CPU memory usage during experiment preview, to help with instance sizing.Īdded CPU memory usage of C++ MOJO to experiment summary and to Deploy Wizard.Īdded a page that lets admin users view system logs.Įxperiments can be assigned to a project from experiment page through Deployments wizard. Note that this new option is distinct from the existing Fit & Transform option. Visualizations of training data, the potential temporal gap until production, and the forecast horizon are now also provided.Īdded the option to select specific leaderboards for IID and time-series experiments.Īdded support to transform a dataset with the experiment’s fitted pipeline (excluding any models). The Experiment Wizard now provides detailed control of time-series validation splits. A visualization of the split is also displayed. You can now split train and test data by a specific date or datetime. You can now use interactive plots created with Plotly. Automated Model Documentation (AutoDoc).Driverless AI License and Version Support. ![]()
0 Comments
Leave a Reply. |
AuthorWrite something about yourself. No need to be fancy, just an overview. ArchivesCategories |