Update src/streamlit_app.py

src/streamlit_app.py

@@ -8,21 +8,12 @@ from PIL import Image
 import torch
 from transformers import DetrImageProcessor, DetrForObjectDetection
 
-# === Ensure cache directory for model download ===
-os.makedirs("./model_cache", exist_ok=True)
-
-# === Load Hugging Face Model ===
-processor = DetrImageProcessor.from_pretrained(
-    "NaveenKumar5/Solar_panel_fault_detection",
-    cache_dir="./model_cache"
-)
-model = DetrForObjectDetection.from_pretrained(
-    "NaveenKumar5/Solar_panel_fault_detection",
-    cache_dir="./model_cache"
-)
+# === Load Hugging Face Model using default cache (handles permissions safely) ===
+processor = DetrImageProcessor.from_pretrained("NaveenKumar5/Solar_panel_fault_detection")
+model = DetrForObjectDetection.from_pretrained("NaveenKumar5/Solar_panel_fault_detection")
 model.eval()
 
-# === Streamlit App ===
+# === Streamlit App Setup ===
 st.set_page_config(page_title="Solar Panel Fault Detection", layout="wide")
 st.title("🔍 Solar Panel Fault Detection (DETR - Hugging Face)")
 st.write("Upload a thermal video (MP4). Faults will be detected using a DETR model from Hugging Face.")
@@ -60,7 +51,7 @@ def detect_faults(frame, frame_idx, fps):
         })
     return frame, faults
 
-# === Video Processing ===
+# === Video Processing Function ===
 def process_video(video_path):
     cap = cv2.VideoCapture(video_path)
     fps = int(cap.get(cv2.CAP_PROP_FPS))
@@ -90,7 +81,7 @@ def process_video(video_path):
     writer.release()
     return output_path, fault_log
 
-# === CSV Download Helper ===
+# === CSV Helper ===
 def convert_df(df):
     return df.to_csv(index=False).encode('utf-8')
 
@@ -120,4 +111,4 @@ if uploaded_file:
     os.unlink(output_path)
 
 st.markdown("---")
-st.caption("Built with Streamlit + Hugging Face DETR + OpenCV") 
+st.caption("Built with Streamlit + Hugging Face DETR + OpenCV")