| # DistilBERT Classifier Checkpoint | |
| Copy the files of your fine-tuned `distilbert-base-uncased` sequence classification model (e.g. from Google Colab output or ShadowLens checkpoint) into this folder. | |
| The folder structure should contain: | |
| - `config.json` | |
| - `model.safetensors` or `pytorch_model.bin` | |
| - `tokenizer.json` | |
| - `tokenizer_config.json` | |
| - `vocab.txt` | |
| - `special_tokens_map.json` | |
| If the model files are not found or loaded in this directory at startup, the system will fall back to using **Layer 1 (Rule-Based)** and **Layer 2 (Heuristic Analysis)**, reporting degraded mode in health checks. | |