| project_name: police_case_reports_json |
| tasks: |
| - name: sentiment |
| type: sentiment |
| data_path: /home/naynek/Desktop/TujengeAI/Update/JengaAI/backend/uploads/datasets/16ced0930e7e_police_case_reports.json |
| heads: |
| - name: sentiment |
| num_labels: 4 |
| weight: 1.0 |
| dropout: 0.1 |
| text_column: report_text |
| label_column: sentiment |
| label_maps: |
| sentiment: |
| 0: negative |
| 1: neutral |
| 2: positive |
| 3: urgent |
| - name: main_category |
| type: single_label_classification |
| data_path: /home/naynek/Desktop/TujengeAI/Update/JengaAI/backend/uploads/datasets/16ced0930e7e_police_case_reports.json |
| heads: |
| - name: main_category |
| num_labels: 10 |
| weight: 1.0 |
| dropout: 0.1 |
| text_column: report_text |
| label_column: main_category |
| label_maps: |
| main_category: |
| 0: Child Abuse |
| 1: Commendation |
| 2: Drug Abuse |
| 3: Fraud |
| 4: Human Trafficking |
| 5: Missing Person |
| 6: Theft |
| 7: Traffic |
| 8: Vandalism |
| 9: Violence |
| - name: subcategory |
| type: single_label_classification |
| data_path: /home/naynek/Desktop/TujengeAI/Update/JengaAI/backend/uploads/datasets/16ced0930e7e_police_case_reports.json |
| heads: |
| - name: subcategory |
| num_labels: 24 |
| weight: 1.0 |
| dropout: 0.1 |
| text_column: report_text |
| label_column: subcategory |
| label_maps: |
| subcategory: |
| 0: ATM Fraud |
| 1: Abandoned Child |
| 2: Abandoned Vehicle |
| 3: Bank Robbery |
| 4: Breaking and Entering |
| 5: Child Labour |
| 6: Consumer Fraud |
| 7: Domestic Servitude |
| 8: Domestic Violence |
| 9: Drug Distribution |
| 10: Extortion |
| 11: Gunfire |
| 12: Hit and Run |
| 13: Identity Theft |
| 14: Illicit Brewing |
| 15: Missing Child |
| 16: Mugging |
| 17: Officer Commendation |
| 18: Phone Scam |
| 19: Property Destruction |
| 20: Robbery |
| 21: Sexual Abuse |
| 22: Tourist Scam |
| 23: Vehicle Theft |
| - name: priority |
| type: single_label_classification |
| data_path: /home/naynek/Desktop/TujengeAI/Update/JengaAI/backend/uploads/datasets/16ced0930e7e_police_case_reports.json |
| heads: |
| - name: priority |
| num_labels: 4 |
| weight: 1.0 |
| dropout: 0.1 |
| text_column: report_text |
| label_column: priority |
| label_maps: |
| priority: |
| 0: critical |
| 1: high |
| 2: low |
| 3: medium |
| - name: intervention |
| type: single_label_classification |
| data_path: /home/naynek/Desktop/TujengeAI/Update/JengaAI/backend/uploads/datasets/16ced0930e7e_police_case_reports.json |
| heads: |
| - name: intervention |
| num_labels: 8 |
| weight: 1.0 |
| dropout: 0.1 |
| text_column: report_text |
| label_column: intervention |
| label_maps: |
| intervention: |
| 0: Counseling |
| 1: Investigation |
| 2: Legal Aid |
| 3: Medical |
| 4: None Required |
| 5: Police Dispatch |
| 6: Search and Rescue |
| 7: Welfare Services |
| - name: entity_extraction |
| type: ner |
| data_path: /home/naynek/Desktop/TujengeAI/Update/JengaAI/backend/uploads/datasets/16ced0930e7e_police_case_reports.json |
| heads: |
| - name: entity_extraction |
| num_labels: 15 |
| weight: 1.0 |
| dropout: 0.1 |
| text_column: report_text |
| label_column: entities |
| label_maps: |
| ner_head: |
| 0: O |
| 1: B-DATE |
| 2: I-DATE |
| 3: B-LOCATION |
| 4: I-LOCATION |
| 5: B-NAME |
| 6: I-NAME |
| 7: B-ORG |
| 8: I-ORG |
| 9: B-TIME |
| 10: I-TIME |
| 11: B-VEHICLE |
| 12: I-VEHICLE |
| 13: B-WEAPON |
| 14: I-WEAPON |
| model: |
| base_model: distilbert-base-uncased |
| hidden_size: 768 |
| dropout: 0.1 |
| fusion: |
| type: attention |
| dropout: 0.1 |
| use_residual: true |
| num_attention_heads: 1 |
| gate_init_value: 0.5 |
| freeze_encoder_layers: 0 |
| gradient_checkpointing: false |
| tokenizer: |
| max_length: 512 |
| padding: max_length |
| truncation: true |
| training: |
| output_dir: ./results/police_case_reports_json |
| learning_rate: 2.0e-05 |
| batch_size: 4 |
| eval_batch_size: 4 |
| num_epochs: 4 |
| weight_decay: 0.01 |
| warmup_steps: 50 |
| max_grad_norm: 1.0 |
| gradient_accumulation_steps: 1 |
| use_amp: false |
| device: auto |
| task_sampling: round_robin |
| temperature: 2.0 |
| early_stopping_patience: 3 |
| metric_for_best_model: eval_loss |
| greater_is_better: false |
| logging: |
| service: mlflow |
| experiment_name: police_case_reports_json |
| tracking_uri: http://localhost:5000 |
| log_every_n_steps: 10 |
| checkpoint: |
| save_every_n_epochs: 1 |
| save_best: true |
| max_checkpoints: 2 |
| data: |
| test_size: 0.08 |
| seed: 42 |
| num_workers: 0 |
| pin_memory: true |
| pii_redaction: |
| enabled: false |
| detect_types: |
| - email |
| - phone |
| - url |
| - card |
| - name |
| - org |
| log_detections: true |
|
|
