Update app.py

app.py

@@ -1,38 +1,46 @@
 import gradio as gr
+import torch
+from transformers import AutoModel, AutoTokenizer
 from PIL import Image
-# Assuming 'model' and 'tokenizer' are defined elsewhere in your code
-# from your_model_file import model, tokenizer
-
-def load_image(image_file):
-    """Load and preprocess the image."""
-    if isinstance(image_file, Image.Image):  # Check if the input is an Image object
-        return image_file.convert("RGB")  # Convert to RGB if necessary
-    elif isinstance(image_file, str) and (image_file.startswith('http') or image_file.startswith('https')):
-        # Handle URL case (you can use an external library to fetch the image if needed)
-        return Image.open(requests.get(image_file, stream=True).raw).convert("RGB")
-    else:
-        # Handle file path case
-        return Image.open(image_file).convert("RGB")
 
+# Check GPU availability
+if torch.cuda.is_available():
+    print("CUDA is available! GPU is present.")
+    print(f"Number of GPUs: {torch.cuda.device_count()}")
+    print(f"GPU Name: {torch.cuda.get_device_name(0)}")
+else:
+    print("CUDA is not available. Running on CPU.")
+
+# Load the tokenizer and model
+tokenizer = AutoTokenizer.from_pretrained('ucaslcl/GOT-OCR2_0', trust_remote_code=True)
+
+# Initialize the model
+if torch.cuda.is_available():
+    model = AutoModel.from_pretrained('ucaslcl/GOT-OCR2_0', trust_remote_code=True, low_cpu_mem_usage=True, device_map='cuda', use_safetensors=True, pad_token_id=tokenizer.eos_token_id)
+    model = model.eval().cuda()
+else:
+    model = AutoModel.from_pretrained('ucaslcl/GOT-OCR2_0', trust_remote_code=True, low_cpu_mem_usage=True, pad_token_id=tokenizer.eos_token_id)
+    model = model.eval()  # Keep model on CPU
+
+# Define the OCR function
 def perform_ocr(image):
-    """Perform OCR on the uploaded image."""
-    try:
-        # Load and preprocess the image
-        processed_image = load_image(image)
-        # Use the model for OCR
-        res = model.chat(tokenizer, processed_image, ocr_type='ocr')
-        return res
-    except Exception as e:
-        return str(e)  # Return the error message
-
-# Gradio interface setup
-iface = gr.Interface(
+    # Convert PIL image to RGB format (if necessary)
+    if image.mode != "RGB":
+        image = image.convert("RGB")
+
+    # Perform OCR using the model
+    res = model.chat(tokenizer, image, ocr_type='ocr')
+
+    return res
+
+# Define the Gradio interface
+interface = gr.Interface(
     fn=perform_ocr,
-    inputs=gr.Image(type="pil"),  # Ensure Gradio accepts images as PIL images
-    outputs="text",
-    title="OCR Application",
-    description="Upload an image to perform Optical Character Recognition (OCR)."
+    inputs=gr.Image(type="pil", label="Upload Image"),
+    outputs=gr.Textbox(label="Extracted Text"),
+    title="OCR and Document Search Web Application",
+    description="Upload an image to extract text using the GOT-OCR2_0 model."
 )
 
-if _name_ == "_main_":
-    iface.launch()
+# Launch the Gradio app
+interface.launch()