Instructions to use Carlosian/hermes-katana-v15-large with libraries, inference providers, notebooks, and local apps. Follow these links to get started.
- Libraries
- Transformers
How to use Carlosian/hermes-katana-v15-large with Transformers:
# Use a pipeline as a high-level helper from transformers import pipeline pipe = pipeline("text-classification", model="Carlosian/hermes-katana-v15-large")# Load model directly from transformers import AutoTokenizer, AutoModelForSequenceClassification tokenizer = AutoTokenizer.from_pretrained("Carlosian/hermes-katana-v15-large") model = AutoModelForSequenceClassification.from_pretrained("Carlosian/hermes-katana-v15-large") - Notebooks
- Google Colab
- Kaggle
Hermes Katana v15 Large
Optional high-accuracy local Scabbard model artifact for Hermes Katana V3. This is not selected by default and should be distributed carefully.
Distribution Status
This repository is staged as private for the V3 release workflow. Do not upload raw training datasets, red-team corpora, attack prompts, or run traces here.
Intended Use
This artifact is used by Hermes Katana's optional Scabbard local classifier path. The base Hermes Katana package works without downloading this artifact.
Files
See artifact_manifest.json for SHA-256 hashes and byte sizes. Hermes Katana verifies the manifest before marking the artifact ready.
Safety Notes
This model is a defensive prompt-injection/security classifier. It can still be dual-use because detector artifacts may reveal signal about the defense surface. Keep access private or gated until the release decision is final.
Usage
katana artifacts download large --revision v3.0.0
katana artifacts status large
- Downloads last month
- 19