Update app.py

app.py

@@ -1,12 +1,17 @@
-import torch
 import streamlit as st
 import numpy as np
 from PIL import Image
 from transformers import CLIPProcessor, CLIPModel, YolosImageProcessor, YolosForObjectDetection
+import torch
 
 st.title("CLIP & CROP")
 st.markdown("**Extract sections of images from your image by using OpenAI's CLIP and Facebooks Detr implemented on HuggingFace Transformers, if the similarity score is not so much, then please consider the prediction to be void.**")
 
+IMAGE_INPUT = st.file_uploader(type=["jpg", "png"], label="Input image")
+TEXT_INPUT = st.text_input(label="Description for section to extracted")
+NUMBER_INPUT = st.number_input(value=0.96, label="Threshold percentage score")
+
+
 with st.spinner("Models are loading"):
     feature_extractor = YolosImageProcessor.from_pretrained("hustvl/yolos-tiny")
     dmodel = YolosForObjectDetection.from_pretrained('hustvl/yolos-tiny')
@@ -14,10 +19,6 @@ with st.spinner("Models are loading"):
     model = CLIPModel.from_pretrained("openai/clip-vit-base-patch16")
     processor = CLIPProcessor.from_pretrained("openai/clip-vit-base-patch16")
  
-IMAGE_INPUT = st.file_uploader(type="pil", label="Input image")
-TEXT_INPUT = st.text_input(label="Description for section to extracted")
-NUMBER_INPUT = st.number_input(value=0.96, label="Threshold percentage score")
-
 SUBMIT_BUTTON = st.button("SUBMIT")
 
 def extract_image(image, text, prob, num=1):