Image Classification
OpenCLIP
Safetensors
tryon-safety-classifier
content-moderation
safety
fashion
virtual-try-on
siglip2
Instructions to use Genlook/tryon-safety-classifier with libraries, inference providers, notebooks, and local apps. Follow these links to get started.
- Libraries
- OpenCLIP
How to use Genlook/tryon-safety-classifier with OpenCLIP:
import open_clip model, preprocess_train, preprocess_val = open_clip.create_model_and_transforms('hf-hub:Genlook/tryon-safety-classifier') tokenizer = open_clip.get_tokenizer('hf-hub:Genlook/tryon-safety-classifier') - Notebooks
- Google Colab
- Kaggle
Download config.json from Genlook/tryon-safety-classifier: direct link, hf CLI and curl.
- Browser
- Download file 1.05 kB
-
https://huggingface.co/Genlook/tryon-safety-classifier/resolve/main/config.json
- Command line
-
hf download hf://Genlook/tryon-safety-classifier/config.json
-
curl -L -o config.json https://huggingface.co/Genlook/tryon-safety-classifier/resolve/main/config.json
1.05 kB
| { | |
| "model_type": "tryon-safety-classifier", | |
| "version": "1.0.0", | |
| "base_model": "timm/ViT-SO400M-16-SigLIP2-384", | |
| "open_clip_name": "hf-hub:timm/ViT-SO400M-16-SigLIP2-384", | |
| "image_size": 384, | |
| "default_variant": "finetuned", | |
| "variants": { | |
| "finetuned": { | |
| "subfolder": null, | |
| "description": "SigLIP 2 vision tower fine-tuned (last 6 blocks) with four linear heads. Product score = mean of per-photo probabilities.", | |
| "embedding_scale": 10.0, | |
| "aggregation": "mean_of_photo_probabilities", | |
| "adult_threshold": 0.3 | |
| }, | |
| "frozen": { | |
| "subfolder": "frozen", | |
| "description": "Original SigLIP 2 vision tower with four linear heads on concat(mean, max) of L2-normalised photo embeddings.", | |
| "aggregation": "concat_mean_max_embeddings", | |
| "adult_threshold": 0.3 | |
| } | |
| }, | |
| "heads": { | |
| "safety": ["ok", "revealing", "lingerie", "adult"], | |
| "subject": ["human_clothing", "pet", "not_clothing"], | |
| "swim": ["not_swimwear", "regular", "near_micro", "micro"], | |
| "rave": ["no", "yes"] | |
| } | |
| } | |