from .h5 to keras

.gitattributes

@@ -33,3 +33,4 @@ saved_model/**/* filter=lfs diff=lfs merge=lfs -text
 *.zip filter=lfs diff=lfs merge=lfs -text
 *.zst filter=lfs diff=lfs merge=lfs -text
 *tfevents* filter=lfs diff=lfs merge=lfs -text
+*.keras filter=lfs diff=lfs merge=lfs -text

app.py

@@ -1,14 +1,20 @@
 import streamlit as st
-from tf_keras.models import load_model          # <-- switch to tf_keras
+import tensorflow as tf
+from tensorflow.keras.models import load_model
 from tensorflow.keras.preprocessing import image
 from tensorflow.keras.applications.resnet50 import preprocess_input
 import numpy as np
 from PIL import Image
 
+# --- Load Your Model and Class Names ---
+# Use st.cache_resource to load the model only once
 @st.cache_resource
 def load_my_model():
-    return load_model("resnet50_dryfruits.h5", compile=False) 
+    # Make sure this file name matches your model file
+    model = load_model('resnet50_dryfruits.keras') 
+    return model
 
+# --- This is the updated dictionary based on your list ---
 class_names = {
     0: 'AlmondGrade_A',
     1: 'CashewGrade_B',
@@ -20,25 +26,31 @@ class_names = {
     7: 'WalnutGrade_A',
     8: 'CashewGrade_C'
 }
+# --------------------------------------------------------
 
 model = load_my_model()
 
+# --- App Interface ---
 st.title("Dry Fruit Quality Grader")
 st.write("Upload an image of a dry fruit, and the model will predict its grade.")
 
 uploaded_file = st.file_uploader("Choose an image...", type=["jpg", "jpeg", "png"])
 
 if uploaded_file is not None:
-    img = Image.open(uploaded_file).convert("RGB")
+    # 1. Preprocess the image
+    img = Image.open(uploaded_file).convert('RGB') # Ensure 3 channels
     img = img.resize((224, 224))
     img_array = image.img_to_array(img)
     img_batch = np.expand_dims(img_array, axis=0)
     img_preprocessed = preprocess_input(img_batch)
-
-    preds = model.predict(img_preprocessed)
-    idx = int(np.argmax(preds[0]))
-    confidence = float(np.max(preds[0]))
-
+    
+    # 2. Make prediction
+    prediction = model.predict(img_preprocessed)
+    predicted_index = np.argmax(prediction[0])
+    predicted_class_name = class_names[predicted_index]
+    confidence = np.max(prediction[0])
+    
+    # 3. Display results
     st.image(img, caption="Uploaded Image", use_column_width=True)
-    st.markdown(f"## Prediction: **{class_names[idx]}**")
+    st.markdown(f"## Prediction: **{predicted_class_name}**")
     st.markdown(f"### Confidence: **{confidence * 100:.2f}%**")

resnet50_dryfruits.h5 → resnet50_dryfruits.keras

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:a4d1c65a486780cb6e1c33c9bda9eae34edb0c658ea450ca8cab12f064787d0e
-size 286585880
+oid sha256:bb526f58f4a6b027e5bb9cab55b7a51040b9e4dffdad8c64a0f71ba7f7234cef
+size 96054159