Instructions to use BillFisk/ai-text-detector with libraries, inference providers, notebooks, and local apps. Follow these links to get started.
- Libraries
- Transformers.js
How to use BillFisk/ai-text-detector with Transformers.js:
// npm i @huggingface/transformers import { pipeline } from '@huggingface/transformers'; // Allocate pipeline const pipe = await pipeline('text-classification', 'BillFisk/ai-text-detector');
Download serving_head.json from BillFisk/ai-text-detector: direct link, hf CLI and curl.
- Browser
- Download file 796 Bytes
-
https://huggingface.co/BillFisk/ai-text-detector/resolve/main/serving_head.json
- Command line
-
hf download hf://BillFisk/ai-text-detector/serving_head.json
-
curl -L -o serving_head.json https://huggingface.co/BillFisk/ai-text-detector/resolve/main/serving_head.json
796 Bytes
| { | |
| "backbone": "deberta", | |
| "architecture": "DesklibAIDetectionModel", | |
| "pooling": "mean", | |
| "num_labels": 4, | |
| "labels": [ | |
| "human", | |
| "ai", | |
| "ai_edited", | |
| "humanized" | |
| ], | |
| "trained_classes": [ | |
| "human", | |
| "ai", | |
| "ai_edited", | |
| "humanized" | |
| ], | |
| "human_index": 0, | |
| "ai_score": "1 - softmax(logits)[human]", | |
| "threshold_at_0.5fpr": 0.9974874649196864, | |
| "base_threshold_at_0.5fpr": 0.9755912190656111, | |
| "fine_tune": "@bandf/clueso LoRA v2", | |
| "calibration_protocol": "joint: false-positive budget split between document and window rules to maximize recall over AI and derived mixed documents with the union at target", | |
| "base_threshold_at_0_5fpr": 0.9974874649196864, | |
| "threshold_at_0_5fpr": 0.9990641518953358, | |
| "window_threshold_at_0_5fpr": 0.9999630607884616 | |
| } | |