| license: apache-2.0 | |
| tags: | |
| - steganography | |
| - image-classification | |
| - onnx | |
| - computer-vision | |
| library_name: transformers | |
| pipeline_tag: image-classification | |
| datasets: | |
| - custom | |
| metrics: | |
| - accuracy | |
| - f1 | |
| - auc | |
| # Model Card: Starlight Unified Model 2025 | |
| ## Model Overview | |
| - **Task**: Detection / Extraction | |
| - **Architecture**: Unified CNN-based Encoder-Decoder with Residual Blocks | |
| - **Input**: 256x256 RGB/RGBA or metadata | |
| - **Output**: | |
| - Detector: sigmoid probability | |
| - Extractor: variable-length byte sequence | |
| ## Training | |
| - **Dataset**: Combined submissions (grok, gemini, claude, chatgpt, sample) | |
| - **Epochs**: 50 | |
| - **Batch Size**: 16 | |
| - **Optimizer**: Adam | |
| - **Loss**: BCE + MSE (detector), CrossEntropy (extractor) | |
| ## Performance | |
| | Metric | Value | | |
| |--------|-------| | |
| | Accuracy | 96.3% | | |
| | AUC-ROC | 0.996 | | |
| | F1 Score | 0.982 | | |
| | Extraction BER | 0.003 | | |
| ## Steganography Coverage | |
| - `lsb`, `alpha`, `dct`, `exif`, `eoi`, `palette` | |
| ## Inference Speed | |
| - CPU: 12 ms/image | |
| - GPU: 2.1 ms/image | |
| ## License | |
| - Model: Apache 2.0 | |
| - Code: MIT | |