Upload README.md with huggingface_hub

README.md

@@ -6,31 +6,24 @@ tags:
 - pytorch
 datasets:
 - wendys-llc/chkbx
-widget:
-- src: https://i.imgur.com/ExampleChecked.jpg
-  candidate_labels: unchecked, checked
-  example_title: Checked Box
-- src: https://i.imgur.com/ExampleUnchecked.jpg  
-  candidate_labels: unchecked, checked
-  example_title: Unchecked Box
 ---
 
 # Checkbox Classifier
 
-Binary classifier for checkbox states (checked/unchecked). Trained on the [wendys-llc/chkbx](https://huggingface.co/datasets/wendys-llc/chkbx) dataset.
+Binary classifier for checkbox states (checked/unchecked).
 
-## Usage
+## Usage with Transformers
 
 ```python
 from transformers import pipeline
-from PIL import Image
 
 # Load pipeline
 classifier = pipeline("image-classification", 
                      model="wendys-llc/checkbox-classifier", 
                      trust_remote_code=True)
 
-# Classify an image
+# Predict
+from PIL import Image
 image = Image.open("checkbox.jpg")
 result = classifier(image)
 print(result)
@@ -38,29 +31,30 @@ print(result)
 #   {'label': 'checked', 'score': 0.99},
 #   {'label': 'unchecked', 'score': 0.01}
 # ]
-
-# Get just the top prediction
-top_result = classifier(image, top_k=1)
-print(f"State: {top_result[0]['label']}")
 ```
 
-## Model Details
-
-- **Architecture**: EfficientNetV2-S
-- **Input Size**: 128x128 RGB images  
-- **Training**: Mixed precision on A100 GPU
-- **Validation Accuracy**: 97.1%
-- **License**: Apache 2.0
+## Direct Usage
 
-## Intended Use
+```python
+from transformers import AutoModelForImageClassification, AutoImageProcessor
+import torch
+from PIL import Image
 
-This model is designed to classify UI checkboxes in screenshots or interface images. It works best on:
-- Clear, high-contrast checkbox images
-- Standard UI checkboxes (not hand-drawn)
-- Images where the checkbox is the main focus
+model = AutoModelForImageClassification.from_pretrained(
+    "wendys-llc/checkbox-classifier", 
+    trust_remote_code=True
+)
+processor = AutoImageProcessor.from_pretrained("wendys-llc/checkbox-classifier")
 
-## Limitations
+image = Image.open("checkbox.jpg")
+inputs = processor(images=image, return_tensors="pt")
+
+with torch.no_grad():
+    outputs = model(**inputs)
+    logits = outputs.logits
+    predicted_class = logits.argmax(-1).item()
+    
+print(model.config.id2label[predicted_class])
+```
 
-- May not work well on hand-drawn checkmarks
-- Trained on UI checkboxes, not paper forms
-- Best performance when checkbox is clearly visible and not too small
+## Accuracy: 97.1%