{"id":29910914,"url":"https://github.com/mosesab/yolov12-comic-panel-detection","last_synced_at":"2025-08-02T01:43:44.485Z","repository":{"id":302128354,"uuid":"1011333549","full_name":"mosesab/YOLOV12-Comic-Panel-Detection","owner":"mosesab","description":"A complete workflow for fine-tuning a YOLOv12 model to detect panels in comic book pages","archived":false,"fork":false,"pushed_at":"2025-06-30T17:42:52.000Z","size":14111,"stargazers_count":0,"open_issues_count":0,"forks_count":0,"subscribers_count":0,"default_branch":"main","last_synced_at":"2025-06-30T18:43:31.299Z","etag":null,"topics":[],"latest_commit_sha":null,"homepage":null,"language":"Jupyter Notebook","has_issues":true,"has_wiki":null,"has_pages":null,"mirror_url":null,"source_name":null,"license":"mit","status":null,"scm":"git","pull_requests_enabled":true,"icon_url":"https://github.com/mosesab.png","metadata":{"files":{"readme":"README.md","changelog":null,"contributing":null,"funding":null,"license":"LICENSE","code_of_conduct":null,"threat_model":null,"audit":null,"citation":null,"codeowners":null,"security":null,"support":null,"governance":null,"roadmap":null,"authors":null,"dei":null,"publiccode":null,"codemeta":null,"zenodo":null}},"created_at":"2025-06-30T16:48:04.000Z","updated_at":"2025-06-30T17:42:55.000Z","dependencies_parsed_at":"2025-06-30T18:53:33.851Z","dependency_job_id":null,"html_url":"https://github.com/mosesab/YOLOV12-Comic-Panel-Detection","commit_stats":null,"previous_names":["mosesab/yolov12-comic-panel-detection"],"tags_count":0,"template":false,"template_full_name":null,"purl":"pkg:github/mosesab/YOLOV12-Comic-Panel-Detection","repository_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/mosesab%2FYOLOV12-Comic-Panel-Detection","tags_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/mosesab%2FYOLOV12-Comic-Panel-Detection/tags","releases_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/mosesab%2FYOLOV12-Comic-Panel-Detection/releases","manifests_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/mosesab%2FYOLOV12-Comic-Panel-Detection/manifests","owner_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/owners/mosesab","download_url":"https://codeload.github.com/mosesab/YOLOV12-Comic-Panel-Detection/tar.gz/refs/heads/main","sbom_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/mosesab%2FYOLOV12-Comic-Panel-Detection/sbom","host":{"name":"GitHub","url":"https://github.com","kind":"github","repositories_count":268326466,"owners_count":24232478,"icon_url":"https://github.com/github.png","version":null,"created_at":"2022-05-30T11:31:42.601Z","updated_at":"2022-07-04T15:15:14.044Z","status":"online","status_checked_at":"2025-08-01T02:00:08.611Z","response_time":67,"last_error":null,"robots_txt_status":"success","robots_txt_updated_at":"2025-07-24T06:49:26.215Z","robots_txt_url":"https://github.com/robots.txt","online":true,"can_crawl_api":true,"host_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub","repositories_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories","repository_names_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repository_names","owners_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/owners"}},"keywords":[],"created_at":"2025-08-02T01:43:41.385Z","updated_at":"2025-08-02T01:43:44.430Z","avatar_url":"https://github.com/mosesab.png","language":"Jupyter Notebook","funding_links":[],"categories":[],"sub_categories":[],"readme":"# YOLOv12 Comic Panel Detection\r\n\r\nThis repository contains a complete workflow for fine-tuning a **YOLOv12** model to detect panels in comic book pages. The project uses the powerful Ultralytics library and a custom dataset managed on Roboflow.\r\n\r\nThe primary goal is to demonstrate an end-to-end process: from setting up the environment and loading a pre-trained model to training, validating, and performing inference on new images.\r\n\r\n*Visit this space to try out the model right now: [`The_Best_Comic_Panel_Detection`](https://huggingface.co/spaces/mosesb/best-comic-panel-detection).*\r\n\r\n![Example Prediction](my_yolo_train/prediction/LTB001_247_jpg.rf.2b6fef98840c5669845b035be2179161.jpg)\r\n\r\n## Key Features\r\n\r\n*   **State-of-the-Art Model:** Fine-tunes the `yolo12x` model, the largest and most performant variant of the YOLOv12 series.\r\n*   **Custom Dataset:** Integrates with Roboflow for easy dataset download, versioning, and management.\r\n*   **End-to-End Workflow:** The entire process is detailed in a single Jupyter Notebook (`comic-boundary-detection.ipynb`).\r\n*   **Comprehensive Evaluation:** Generates and displays key performance metrics, including loss curves, confusion matrices, and precision-recall curves.\r\n*   **Reproducibility:** Aims for easy setup and execution, allowing others to replicate the results.\r\n\r\n## File Structure\r\n\r\nThe repository is organized as follows:\r\n\r\n```\r\n.\r\n├── comic-boundary-detection.ipynb  # The main notebook with all the code\r\n├── my_yolo_train/                  # Directory for all training outputs\r\n│   ├── demonstration/              # Output from the initial model test\r\n│   ├── finetuning/                 # All artifacts from the fine-tuning process\r\n│   │   ├── args.yaml               # Training configuration\r\n│   │   ├── results.csv             # Training metrics per epoch\r\n│   │   ├── results.png             # Plot of training/validation loss \u0026 metrics\r\n│   │   ├── confusion_matrix.png    # Performance visualization\r\n│   │   └── weights/\r\n│   │       ├── best.pt             # Best performing model checkpoint (recommended for inference)\r\n│   │       └── last.pt             # Model checkpoint from the final epoch\r\n│   ├── prediction/                 # Inference output on a sample test image\r\n│   └── validation/                 # Validation results on the test set\r\n└── ...                             # Other project files\r\n```\r\n\r\n## Training Results\r\n\r\nThe model was trained for 200 epochs, achieving high performance on the validation set. Below are the key training metrics over time.\r\n\r\n![Training Results](my_yolo_train/finetuning/results.png)\r\n\r\n### Val Pred Visualization\r\n\r\n![Val Visualization](my_yolo_train/finetuning/val_batch0_pred.jpg)\r\n\r\n## Model Checkpoints\r\n\r\nThe trained model weights are saved in the `my_yolo_train/finetuning/weights/` directory.\r\n\r\n*   `best.pt`: The model checkpoint with the best mAP50-95 score on the validation set. **This is the recommended model for inference.**\r\n*   `last.pt`: The model checkpoint saved at the end of the final training epoch.\r\n\r\n\r\n## Acknowledgements\r\n\r\n*   **Ultralytics** for the amazing [YOLOv12 model](https://github.com/ultralytics/ultralytics) and library.\r\n*   **Roboflow:** for their dataset hosting platform and **custom-workflow-3-object-detection-g24r5-fmfkb** for compiling and annotating this incredible dataset.\r\n\r\n## License\r\n\r\nThis project is licensed under the MIT License. See the `LICENSE` file for details. \r\n","project_url":"https://awesome.ecosyste.ms/api/v1/projects/github.com%2Fmosesab%2Fyolov12-comic-panel-detection","html_url":"https://awesome.ecosyste.ms/projects/github.com%2Fmosesab%2Fyolov12-comic-panel-detection","lists_url":"https://awesome.ecosyste.ms/api/v1/projects/github.com%2Fmosesab%2Fyolov12-comic-panel-detection/lists"}