app uploaded

Insect_HFspace_Streamlit_App.py

@@ -0,0 +1,169 @@
+# -*- coding: utf-8 -*-
+"""
+Created on Tue Nov 18 09:07:10 2025
+
+@author: THYAGHARAJAN
+"""
+
+import streamlit as st
+import tensorflow as tf
+import numpy as np
+from PIL import Image
+from huggingface_hub import hf_hub_download, list_repo_files
+
+# ------------------------------
+# CONFIGURATION
+# ------------------------------
+REPO_ID = "kkthyagharajan/KKT-HF-TransferLearning-Models"     # <<< CHANGE THIS
+IMG_SIZE = (300, 300)
+
+st.set_page_config(page_title="Insect Classifier", layout="wide")
+
+# Cache dictionaries
+@st.cache_resource
+def load_tf_model(model_path):
+    return tf.keras.models.load_model(model_path, compile=False)
+
+@st.cache_resource
+def load_class_names(model_dir):
+    class_file = hf_hub_download(repo_id=REPO_ID, filename=f"{model_dir}/class_names.txt")
+    with open(class_file, "r") as f:
+        return [x.strip() for x in f.read().split(",")]
+
+# ----------------------------------
+# Helper Functions
+# ----------------------------------
+def get_available_models():
+    """Return mapping: model_dir → model file (.h5 preferred over .keras)."""
+    files = list_repo_files(REPO_ID)
+    models = {}
+
+    # Prefer .h5
+    for file in files:
+        if file.endswith(".h5"):
+            dir = file.split("/")[0]
+            models[dir] = file
+
+    # Use .keras only if .h5 missing
+    for file in files:
+        if file.endswith(".keras"):
+            dir = file.split("/")[0]
+            if dir not in models:
+                models[dir] = file
+
+    return models
+
+def get_sample_images(model_dir):
+    """List sample images inside model_dir/sample_images/"""
+    files = list_repo_files(REPO_ID)
+    sample_imgs = []
+    prefix = f"{model_dir}/sample_images/"
+
+    for f in files:
+        if f.startswith(prefix) and f.lower().endswith((".jpg", ".jpeg", ".png")):
+            sample_imgs.append(f.replace(prefix, ""))  
+
+    return sample_imgs
+
+def load_sample_image(model_dir, image_name):
+    """Download sample image."""
+    path = hf_hub_download(repo_id=REPO_ID, filename=f"{model_dir}/sample_images/{image_name}")
+    return Image.open(path)
+
+def preprocess(img):
+    img = img.resize(IMG_SIZE)
+    arr = np.array(img) / 255.0
+    arr = arr.reshape(1, IMG_SIZE[0], IMG_SIZE[1], 3)
+    return arr
+
+# ----------------------------------
+# UI Layout
+# ----------------------------------
+st.title("🦋 Insect Classification System")
+st.markdown("""
+### A Multi-Model Deep Learning Web App  
+Developed by **Dr. Thyagharajan K K, Professor & Dean (Research)**  
+RMD Engineering College  
+""")
+
+col1, col2 = st.columns([1, 1])
+
+# ----------------------------------
+# LEFT PANEL
+# ----------------------------------
+with col1:
+    st.subheader("1️⃣ Select Model")
+    models = get_available_models()
+
+    if not models:
+        st.error("No models found in HuggingFace repo.")
+        st.stop()
+
+    model_choice = st.selectbox("Choose a model", list(models.keys()))
+
+    st.subheader("2️⃣ Choose Image Source")
+    input_mode = st.radio("Select input method:", ["Upload Image", "Use Sample Image"])
+
+    input_image = None
+
+    # Upload
+    if input_mode == "Upload Image":
+        uploaded = st.file_uploader("Upload image", type=["jpg", "jpeg", "png"])
+        if uploaded:
+            input_image = Image.open(uploaded)
+
+    # Sample Images
+    else:
+        sample_images = get_sample_images(model_choice)
+        if sample_images:
+            selected_sample = st.selectbox("Choose sample image", sample_images)
+            if selected_sample:
+                input_image = load_sample_image(model_choice, selected_sample)
+                st.image(input_image, caption="Sample Image", width=250)
+        else:
+            st.warning("No sample images found for this model.")
+
+    st.markdown("---")
+    predict_btn = st.button("🔍 Predict", use_container_width=True)
+
+# ----------------------------------
+# RIGHT PANEL
+# ----------------------------------
+with col2:
+    st.subheader("📊 Prediction Results")
+
+    if predict_btn:
+        if input_image is None:
+            st.error("Please upload or select an image.")
+        else:
+            # Show image
+            st.image(input_image, caption="Input Image", width=300)
+
+            # Load model
+            model_path = hf_hub_download(repo_id=REPO_ID, filename=models[model_choice])
+            model = load_tf_model(model_path)
+            class_names = load_class_names(model_choice)
+
+            # Predict
+            arr = preprocess(input_image)
+            preds = model.predict(arr, verbose=0)[0]
+
+            idx = np.argmax(preds)
+            predicted = class_names[idx]
+
+            st.success(f"### 🟩 Predicted: **{predicted}** ({preds[idx]*100:.2f}%)")
+
+            # Top-3 Predictions
+            st.subheader("Top 3 Predictions")
+            top3 = preds.argsort()[-3:][::-1]
+            for i in top3:
+                st.write(f"**{class_names[i]}** — {preds[i]*100:.2f}%")
+
+# Footer
+st.markdown("---")
+st.markdown("""
+**Developed by:** Dr. Thyagharajan K K  
+**Professor & Dean (Research)**  
+RMD Engineering College  
+📧 **kkthyagharajan@yahoo.com**
+""")

README.md

@@ -1,20 +1,68 @@
----
-title: Insect Streamlit App
-emoji: 🚀
-colorFrom: red
-colorTo: red
-sdk: docker
-app_port: 8501
-tags:
-- streamlit
-pinned: false
-short_description: Insect detection Pretrained DL Models
-license: mit
----
-
-# Welcome to Streamlit!
-
-Edit `/src/streamlit_app.py` to customize this app to your heart's desire. :heart:
-
-If you have any questions, checkout our [documentation](https://docs.streamlit.io) and [community
-forums](https://discuss.streamlit.io).
+---
+title: Insect Detection
+emoji: 🐝
+colorFrom: yellow
+colorTo: green
+sdk: docker
+app_file: Insect_HFspace_Streamlit_App.py
+pinned: false
+license: mit
+tags:
+  - computer-vision
+  - image-classification
+  - insect-classification
+  - deep-learning
+  - tensorflow
+  - mobilenet
+  - efficientnet
+  - resnet
+  - inception
+---
+
+# 🦋 Multi-Model Insect Classification System - A Web/Mobile App
+### Developed by Dr. Thyagharajan K K
+
+## Description
+
+AI-powered insect classification application featuring multiple state-of-the-art deep learning models. Upload images to identify insect species with confidence scores and top-3 predictions.
+
+## Features
+
+- 🎯 Multiple pre-trained models (Inception V3, EfficientNet, ResNet50)
+- 📸 Upload custom images or use sample test images
+- 📊 Confidence scores with top-3 predictions
+- 🚀 Fast inference with model caching
+- 📱 Responsive design for web and mobile
+
+## Models Available
+
+- **Inception V3** - High accuracy, balanced performance
+- **EfficientNet B0** - Efficient and lightweight
+- **ResNet50** - Deep residual learning
+- (More models coming soon...)
+
+## How to Use
+
+1. Select a model from the dropdown
+2. Upload an insect image or choose from sample images
+3. Click "Predict" to get classification results
+4. View predicted class with confidence score
+
+## Technical Details
+
+- **Framework:** TensorFlow/Keras
+- **Input Size:** 300×300 pixels
+- **Interface:** Streamlit
+- **Hosted on:** Hugging Face Spaces
+
+## License
+
+This project is licensed under the MIT License - see the LICENSE file for details.
+
+## Citation
+
+If you use this application in your research or educational projects, please provide appropriate attribution.
+
+## Contact
+
+For questions or collaboration opportunities, please open a discussion in this Space.

requirements.txt

@@ -1,3 +1,5 @@
-altair
-pandas
-streamlit
+streamlit
+tensorflow-cpu==2.14.0
+huggingface_hub
+numpy
+Pillow