Track individual model files instead of single multitask model

The training script creates separate model files for each task
(match_winner, map_winner, score_team1, score_team2, round_diff, total_maps)
so DVC needs to track each file individually.

🤖 Generated with [Claude Code](https://claude.com/claude-code)

Co-Authored-By: Claude <noreply@anthropic.com>

dvc.yaml

@@ -26,7 +26,12 @@ stages:
       - train.max_depth
       - train.random_state
     outs:
-      - models/multitask_model.pkl
+      - models/model_match_winner.pkl
+      - models/model_map_winner.pkl
+      - models/model_score_team1.pkl
+      - models/model_score_team2.pkl
+      - models/model_round_diff.pkl
+      - models/model_total_maps.pkl
     metrics:
       - models/metrics.json:
           cache: false