Instructions to use SkillScreen/SkillScreen-2B-Markdown with libraries, inference providers, notebooks, and local apps. Follow these links to get started.
- Libraries
- Transformers
How to use SkillScreen/SkillScreen-2B-Markdown with Transformers:
# Use a pipeline as a high-level helper from transformers import pipeline pipe = pipeline("text-classification", model="SkillScreen/SkillScreen-2B-Markdown")# pip install -U transformers accelerate # Load model directly from transformers import AutoProcessor, AutoModelForSequenceClassification processor = AutoProcessor.from_pretrained("SkillScreen/SkillScreen-2B-Markdown") model = AutoModelForSequenceClassification.from_pretrained("SkillScreen/SkillScreen-2B-Markdown", device_map="auto") - Notebooks
- Google Colab
- Kaggle
Download evaluation.json from SkillScreen/SkillScreen-2B-Markdown: direct link, hf CLI and curl.
- Browser
- Download file 3.5 kB
-
https://huggingface.co/SkillScreen/SkillScreen-2B-Markdown/resolve/main/evaluation.json
- Command line
-
hf download hf://SkillScreen/SkillScreen-2B-Markdown/evaluation.json
-
curl -L -o evaluation.json https://huggingface.co/SkillScreen/SkillScreen-2B-Markdown/resolve/main/evaluation.json
3.5 kB
| { | |
| "mode": "markdown8k", | |
| "seed": 42, | |
| "records": 13514, | |
| "malicious": 7500, | |
| "benign": 6014, | |
| "auprc_average_precision": 0.9966339274117415, | |
| "auroc": 0.9945986919410504, | |
| "folds_sha256": "c11c6f68dd23fddb70d7090512d93f5be54bb216596175d44fcaa7e7df80e202", | |
| "oof_sha256": "935e97b51798f6b553da87c4cc0738510a7575c14b0fe4bc6db2726145baa888", | |
| "folds": [ | |
| { | |
| "fold": 1, | |
| "predictions_sha256": "a43bd55e2f3256b7e5d959822ce449ecee79dfc2affbb1a3c6a6fab04bbc1ec0", | |
| "adapter_sha256": "79c2cfa9e230b5eb88dc26ae96a5f9b26ea86d6f7f5d187fbb9070b9518d7ec4" | |
| }, | |
| { | |
| "fold": 2, | |
| "predictions_sha256": "608d14ae22ea7579928f0722dc4ba9a4a7dfad923a57e8542d7ad7ae132ed52e", | |
| "adapter_sha256": "7683ceca69861929e1263dd46bf3befeb2534592b90686ad60b4abf21b2841dc" | |
| }, | |
| { | |
| "fold": 3, | |
| "predictions_sha256": "a759c2e629bf2e21382806eb332e0760de792683b337f2529f019e0b730286ba", | |
| "adapter_sha256": "3110418b5329db127917bad3135db46d97610cf126cc9ced6b08f048b7ba5100" | |
| } | |
| ], | |
| "created_at": "2026-11-10T16:05:55.944354+00:00", | |
| "default": { | |
| "raw_threshold": 0.0, | |
| "native_threshold": 0.5, | |
| "tp": 7272, | |
| "fp": 53, | |
| "fn": 228, | |
| "tn": 5961, | |
| "recall": 0.9696, | |
| "fpr": 0.008812770202859994, | |
| "precision": 0.9927645051194539, | |
| "f1": 0.9810455311973019, | |
| "flagged_fraction": 0.5420304869024715 | |
| }, | |
| "targets": { | |
| "0.95": { | |
| "raw_threshold": 5.374836521935736, | |
| "native_threshold": 0.9953896776695467, | |
| "tp": 7125, | |
| "fp": 9, | |
| "fn": 375, | |
| "tn": 6005, | |
| "recall": 0.95, | |
| "fpr": 0.0014965081476554707, | |
| "precision": 0.9987384356602187, | |
| "f1": 0.973759737597376, | |
| "flagged_fraction": 0.5278969957081545 | |
| }, | |
| "0.98": { | |
| "raw_threshold": -2.9843786119895848, | |
| "native_threshold": 0.0481366056169061, | |
| "tp": 7350, | |
| "fp": 243, | |
| "fn": 150, | |
| "tn": 5771, | |
| "recall": 0.98, | |
| "fpr": 0.0404057199866977, | |
| "precision": 0.9679968391939945, | |
| "f1": 0.9739614390777181, | |
| "flagged_fraction": 0.5618617729761729 | |
| }, | |
| "0.99": { | |
| "raw_threshold": -4.441408183360328, | |
| "native_threshold": 0.011642201918123018, | |
| "tp": 7425, | |
| "fp": 945, | |
| "fn": 75, | |
| "tn": 5069, | |
| "recall": 0.99, | |
| "fpr": 0.1571333555038244, | |
| "precision": 0.8870967741935484, | |
| "f1": 0.9357277882797732, | |
| "flagged_fraction": 0.6193577031226876 | |
| } | |
| }, | |
| "balanced": { | |
| "target": 0.95, | |
| "raw_threshold": 5.374836521935736, | |
| "native_threshold": 0.9953896776695467, | |
| "tp": 7125, | |
| "fp": 9, | |
| "fn": 375, | |
| "tn": 6005, | |
| "recall": 0.95, | |
| "fpr": 0.0014965081476554707, | |
| "precision": 0.9987384356602187, | |
| "f1": 0.973759737597376, | |
| "flagged_fraction": 0.5278969957081545 | |
| }, | |
| "high_recall": { | |
| "raw_threshold": -4.441408183360328, | |
| "native_threshold": 0.011642201918123018, | |
| "tp": 7425, | |
| "fp": 945, | |
| "fn": 75, | |
| "tn": 5069, | |
| "recall": 0.99, | |
| "fpr": 0.1571333555038244, | |
| "precision": 0.8870967741935484, | |
| "f1": 0.9357277882797732, | |
| "flagged_fraction": 0.6193577031226876 | |
| }, | |
| "selection": "Balanced maximizes recall minus FPR among predeclared 95/98/99% OOF recall gates; ties prefer higher recall. High-recall targets 99% OOF recall.", | |
| "scope": "Pooled three-fold out-of-fold scores. Every score is from a model that excluded that skill group. Thresholds transfer to a separate full-data refit; these are not held-out test results of that refit." | |
| } | |