do it normally

app.py

@@ -41,37 +41,43 @@ GOT-OCR-2.0 can handle:
 
 model_name = 'ucaslcl/GOT-OCR2_0'
 
+
 tokenizer = AutoTokenizer.from_pretrained('ucaslcl/GOT-OCR2_0', trust_remote_code=True)
-config = AutoConfig.from_pretrained(model_name, trust_remote_code=True)
 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()
 model.config.pad_token_id = tokenizer.eos_token_id
 
-def image_to_base64(image):
-    buffered = io.BytesIO()
-    image.save(buffered, format="PNG")
-    return base64.b64encode(buffered.getvalue()).decode()
+# def image_to_base64(image):
+#     buffered = io.BytesIO()
+#     image.save(buffered, format="PNG")
+#     return base64.b64encode(buffered.getvalue()).decode()
 
 @spaces.GPU
-def process_image(image, task, ocr_type=None, ocr_box=None, ocr_color=None, render=False):
-    if task == "Plain Text OCR":
-        res = model.chat(tokenizer, image, ocr_type='ocr')
-    elif task == "Format Text OCR":
-        res = model.chat(tokenizer, image, ocr_type='format')
-    elif task == "Fine-grained OCR (Box)":
-        res = model.chat(tokenizer, image, ocr_type=ocr_type, ocr_box=ocr_box)
-    elif task == "Fine-grained OCR (Color)":
-        res = model.chat(tokenizer, image, ocr_type=ocr_type, ocr_color=ocr_color)
-    elif task == "Multi-crop OCR":
-        res = model.chat_crop(tokenizer, image_file=image)
-    elif task == "Render Formatted OCR":
-        res = model.chat(tokenizer, image, ocr_type='format', render=True, save_render_file='./demo.html')
-        with open('./demo.html', 'r') as f:
-            html_content = f.read()
-        return res, html_content
-    
-    return res, None
+def process_image(image, task, ocr_type=None, ocr_box=None, ocr_color=None):
+    try:
+        if image is None:
+            return "No image provided", None
 
+        if task == "Plain Text OCR":
+            res = model.chat(tokenizer, image, ocr_type='ocr')
+        elif task == "Format Text OCR":
+            res = model.chat(tokenizer, image, ocr_type='format')
+        elif task == "Fine-grained OCR (Box)":
+            res = model.chat(tokenizer, image, ocr_type=ocr_type, ocr_box=ocr_box)
+        elif task == "Fine-grained OCR (Color)":
+            res = model.chat(tokenizer, image, ocr_type=ocr_type, ocr_color=ocr_color)
+        elif task == "Multi-crop OCR":
+            res = model.chat_crop(tokenizer, image_file=image)
+        elif task == "Render Formatted OCR":
+            res = model.chat(tokenizer, image, ocr_type='format', render=True, save_render_file='./results/demo.html')
+            with open('./results/demo.html', 'r') as f:
+                html_content = f.read()
+            return res, html_content
+        
+        return res, None
+    except Exception as e:
+        return str(e), None
+    
 def update_inputs(task):
     if task == "Plain Text OCR" or task == "Format Text OCR" or task == "Multi-crop OCR":
         return [gr.update(visible=False)] * 4