Image Classification
PyTorch
timm
Transformers
text-rendering
text-to-image
chinese
cjk
glyph
clip
vit
deberta-v2
ids
quality-assurance
reward-model
Instructions to use ATH-MaaS/TextVerifier-CJK-CLIP with libraries, inference providers, notebooks, and local apps. Follow these links to get started.
- Libraries
- timm
How to use ATH-MaaS/TextVerifier-CJK-CLIP with timm:
import timm model = timm.create_model("hf-hub:ATH-MaaS/TextVerifier-CJK-CLIP", pretrained=True) - Transformers
How to use ATH-MaaS/TextVerifier-CJK-CLIP with Transformers:
# Use a pipeline as a high-level helper from transformers import pipeline pipe = pipeline("image-classification", model="ATH-MaaS/TextVerifier-CJK-CLIP") pipe("https://huggingface.co/datasets/huggingface/documentation-images/resolve/main/hub/parrots.png")# pip install -U transformers accelerate # Load model directly from transformers import AutoModel model = AutoModel.from_pretrained("ATH-MaaS/TextVerifier-CJK-CLIP", device_map="auto") - Notebooks
- Google Colab
- Kaggle
File size: 534 Bytes
adc233a | 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 | {
"schema_version": 1,
"architecture": "TextVerifier Stage-3, ViT-L/16 image encoder, IDS-aware dual encoder",
"image_size": 128,
"embedding_dim": 1024,
"prototype_shape": [
8189,
10,
1024
],
"num_characters": 8189,
"K": 10,
"threshold": 0.02152947708964348,
"error_score_rule": "1 - max_j cosine(image_embedding, prototypes[target_character, j])",
"prediction_rule": "render_error if error_score > threshold",
"prototype_norm_min": 0.9999998211860657,
"prototype_norm_max": 1.0000001192092896
}
|