Spaces:
Build error
Build error
Commit History
-
Fix import
371767b -
Add k-fold cross-validation to prediction pipeline
Introduces a --kfold argument to main.py to enable 3-fold cross-validation. Implements run_kfold_cv in pipeline.py, using event-based splits and MLflow for experiment tracking and model registration. Refactors imports and typing for consistency, and moves configuration constants to config.py for better modularity.eb615ca
-
Startup model check: 2025-08-03 13:02:33
f972c61
-
Revert "(CAREFUL!!!!!!) Refactor argument parsing and prediction pipeline"
This reverts commit 7fcaffe1b17b52b1451fe966549d2708dc218ffc.ffd453e -
(CAREFUL!!!!!!) Refactor argument parsing and prediction pipeline
Moved command-line argument parsing to a dedicated src/args.py module and updated main.py and predict/main.py to use these functions. Improved model management logic and modularized pipeline execution. Enhanced feature engineering and preprocessing in predict/preprocess.py and refactored model classes in predict/models.py for consistency and maintainability.7fcaffe
-
Refactor imports to use absolute paths and clean up scripts
Standardized all internal imports to use absolute paths for better maintainability and to avoid import errors. Removed unnecessary sys.path manipulations and direct script execution blocks from modules, making the codebase more modular and suitable for package usage.9678fdb
-
Update README and UFC data, retrain models
Expanded and clarified the README with detailed usage instructions for scraping, prediction, and pipeline execution. Updated ufc_fights.csv with new event results, added output/last_event.json, and refreshed model artifacts and results to reflect retraining on the latest data.5271c2e
-
Add win probability to fight predictions
Updated prediction logic in models and pipeline to return both the predicted winner and the associated win probability. Adjusted app.py and predict_new.py to display the probability alongside the winner. This provides more informative prediction results for users and improves model evaluation reporting.f3ecc65 -
Add grappling features to fighter stats and ML pipeline
Introduces takedown accuracy and submission attempts per minute as new features in both the model and preprocessing pipeline. Updates requirements.txt with pinned versions and adds new dependencies for data processing and visualization. Enhances fighter history stats calculation to include grappling metrics, improving model input feature richness.1999f4d
-
Add new ML models and CLI tools for prediction
AlvaroIntroduces several new machine learning models (XGBoost, LightGBM, SVC, RandomForest, BernoulliNB) to the prediction pipeline by refactoring model logic into a shared base class. Adds requirements for new dependencies. Implements two new CLI scripts: save_model.py for training and saving models, and predict_new.py for predicting outcomes of hypothetical fights using saved models. Improves preprocessing robustness for date parsing.2aed0aa -
Add logistic regression model and ML preprocessing
AlvaroIntroduces a new LogisticRegressionModel for fight outcome prediction, including a comprehensive ML preprocessing module (preprocess.py) for feature engineering and data preparation. Refactors model interfaces to accept fight dictionaries, updates the pipeline and main entrypoint for model selection and reporting, and improves data cleaning in the scraping and preprocessing steps. Also fixes fighter name formatting and adds pandas and scikit-learn as dependencies.bf7e729 -
Refactor prediction pipeline and modularize models
AlvaroReplaces the monolithic predict.py with a modular prediction pipeline. Adds main.py to orchestrate model evaluation, models.py for model abstractions and the ELO baseline, and pipeline.py for data loading, splitting, evaluation, and reporting. Updates process_fights_for_elo to accept either a file path or pre-loaded data for improved flexibility.e012a04