Update app.py

app.py

@@ -4,23 +4,28 @@ import re
 import gradio as gr
 from threading import Thread
 from transformers import TextIteratorStreamer, AutoTokenizer, AutoModelForCausalLM
-
 import subprocess
+
 subprocess.run('pip install flash-attn --no-build-isolation', env={'FLASH_ATTENTION_SKIP_CUDA_BUILD': "TRUE"}, shell=True)
 
 model_id = "vikhyatk/moondream2"
 revision = "2024-05-20"
 tokenizer = AutoTokenizer.from_pretrained(model_id, revision=revision)
 moondream = AutoModelForCausalLM.from_pretrained(
-    model_id, trust_remote_code=True, revision=revision,
-    torch_dtype=torch.bfloat16, device_map={"": "cuda"},
+    model_id,
+    trust_remote_code=True,
+    revision=revision,
+    torch_dtype=torch.bfloat16,
+    device_map={"": "cuda"},
     attn_implementation="flash_attention_2"
 )
 moondream.eval()
 
-
 @spaces.GPU(duration=10)
 def answer_question(img, prompt):
+    if img is None:
+        raise gr.Error("Please upload an image.")
+
     image_embeds = moondream.encode_image(img)
     streamer = TextIteratorStreamer(tokenizer, skip_special_tokens=True)
     thread = Thread(
@@ -33,28 +38,26 @@ def answer_question(img, prompt):
         },
     )
     thread.start()
-
     buffer = ""
     for new_text in streamer:
         buffer += new_text
         yield buffer.strip()
 
-
 with gr.Blocks() as demo:
     gr.Markdown(
         """
         # myAI - AMI Vision Module
-        A lightweight Computer Vision model by @vikhyat
-        - 🌔 [moondream2](https://github.com/vikhyat/moondream)
+        A lightweight Computer Vision model by @vikhyat - 🌔 [moondream2](https://github.com/vikhyat/moondream)
         """
     )
     with gr.Row():
-        prompt = gr.Textbox(label="Input", value="Identify people in this video", scale=4)
+        prompt = gr.Textbox(label="Input", value="Identify people in this image", scale=4)
         submit = gr.Button("Submit")
     with gr.Row():
         img = gr.Image(type="pil", label="Upload an Image")
         output = gr.TextArea(label="Response")
+
     submit.click(answer_question, [img, prompt], output)
     prompt.submit(answer_question, [img, prompt], output)
 
-demo.queue().launch()
+demo.queue().launch()