efaf5ff0e1
Removed features that contain match outcome information: - result_1, result_2 (actual match scores - only known after match) - ct_1, t_2, t_1, ct_2 (rounds won per side - only known after match) - total_rounds, round_diff (derived from results) These features caused perfect 1.0 accuracy because the model was essentially "cheating" by knowing the match outcome. Now using only pre-match information: - Team rankings (rank_1, rank_2) - Historical map performance (map_wins_1, map_wins_2) - Starting side (starting_ct) - Derived: rank_diff, map_wins_diff This will give realistic model performance based on what would actually be known before a match starts. 🤖 Generated with [Claude Code](https://claude.com/claude-code) Co-Authored-By: Claude <noreply@anthropic.com>
MLOps Project
This is an MLOps project for CSGO data analysis and model training.
Features
- Data pipeline with Apache Airflow
- Model training with PyTorch and scikit-learn
- MLflow for experiment tracking
- DVC for data versioning
- Monitoring with Prometheus
- FastAPI for API serving
Setup
-
Install dependencies:
poetry install -
Run the data pipeline:
airflow dags unpause csgo_data_pipeline
Project Structure
dags/: Airflow DAGssrc/: Source codemodels/: Trained modelsdata/: Data filesnotebooks/: Jupyter notebookstests/: Test filesconfig/: Configuration filesdocker/: Docker fileskubernetes/: Kubernetes manifests