Update app.py

app.py

@@ -1,7 +1,33 @@
 import gradio as gr
+from transformers import VisionEncoderDecoderModel, TrOCRProcessor
+import torch
 
-def greet(name):
-    return "Hello " + name + "!!"
+def recognize_captcha(input, mdl):
+    input = transform(input.convert('RGB'))
 
-demo = gr.Interface(fn=greet, inputs="text", outputs="text")
-demo.launch()
+    # Load model and processor
+    processor = TrOCRProcessor.from_pretrained(mdl)
+    model = VisionEncoderDecoderModel.from_pretrained(mdl)
+    
+    # Prepare image
+    pixel_values = processor(input, return_tensors="pt").pixel_values
+
+    # Generate text
+    generated_ids = model.generate(pixel_values)
+    generated_text = processor.batch_decode(generated_ids, skip_special_tokens=True)[0]
+    return generated_text
+
+iface = gr.Interface(
+    recognize_captcha[
+        gr.Image,
+        gr.Dropdown(
+            ['anuashok/ocr-captcha-v3','anuashok/ocr-captcha-v2','anuashok/ocr-captcha-v1'], label='Model to use'
+        )
+    ],
+    'text',
+    title  = "character sequence recognition from scene-image (captcha)",
+    description = "Using some TrOCR models found on the HF Hub. Will you have to train your own?",
+    examples = ['','']
+)
+
+iface.launch()