Spaces:
Sleeping
Sleeping
import gradio as gr | |
from PIL import Image | |
from transformers import ViTImageProcessor, ViTForImageClassification | |
import torch | |
# Load the image processor and model | |
processor = ViTImageProcessor.from_pretrained('wambugu1738/crop_leaf_diseases_vit') | |
model = ViTForImageClassification.from_pretrained( | |
'wambugu1738/crop_leaf_diseases_vit', | |
ignore_mismatched_sizes=True | |
) | |
# Define a function to make predictions | |
def classify_image(image): | |
inputs = processor(images=image, return_tensors="pt") | |
outputs = model(**inputs) | |
logits = outputs.logits | |
predicted_class_idx = logits.argmax(-1).item() | |
return model.config.id2label[predicted_class_idx] | |
# Create the Gradio interface | |
app = gr.Interface( | |
fn=classify_image, | |
inputs=gr.Image(type="numpy"), # Corrected input type | |
outputs="text" | |
) | |
# Launch the Gradio app with a public link | |
app.launch(share=True) | |