Update app.py

app.py

@@ -20,22 +20,24 @@ moondream.eval()
 
 @spaces.GPU(duration=10)
 def answer_questions(image_tuples, prompt_text):
-    print(f"prompt_text:\n{prompt_text}\n")
-    print(f"image_tuples:\n{image_tuples}\n")
+    result = ""
     
+    print(f"prompt_text: {prompt_text}\n")
     prompts = [p.strip() for p in prompt_text.split(',')]  # Splitting and cleaning prompts
-    image_embeds = [img[0] for img in image_tuples if img[0] is not None]  # Extracting images from tuples, ignoring None
+    print(f"prompts: {prompts}\n")
     
-    print(f"image_embeds:\n{image_embeds}\n")
-    print(f"split prompts:\n{prompts}\n")
+    image_embeds = [img[0] for img in image_tuples if img[0] is not None]  # Extracting images from tuples, ignoring None
     
+    # Check if the lengths of image_embeds and prompts are equal
+    if len(image_embeds) != len(prompts):
+        return ("Error: The number of images input and prompts input (seperate by commas in input text field) must be the same.")
+        
     answers = moondream.batch_answer(
         images=image_embeds,
         prompts=prompts,
         tokenizer=tokenizer,
     )
     
-    result = ""
     for question, answer in zip(prompts, answers):
         print(f"Q: {question}")
         print(f"A: {answer}")
@@ -45,10 +47,11 @@ def answer_questions(image_tuples, prompt_text):
     return result
 
 with gr.Blocks() as demo:
+    gr.Markdown("# moondream2 unofficial batch processing demo")
     gr.Markdown("# 🌔 moondream2\nA tiny vision language model. [GitHub](https://github.com/vikhyatk/moondream)")
     with gr.Row():
         img = gr.Gallery(label="Upload Images", type="pil")
-        prompt = gr.Textbox(label="Input Prompts", placeholder="Enter prompts separated by commas. Ex: Describe this image, What is in this image?", lines=2)
+        prompt = gr.Textbox(label="Input Prompts", placeholder="Enter prompts (one prompt for each image provided) separated by commas. Ex: Describe this image, What is in this image?", lines=2)
         submit = gr.Button("Submit")
     output = gr.TextArea(label="Responses", lines=4)
     submit.click(answer_questions, [img, prompt], output)