Update app.py

app.py

@@ -1,5 +1,6 @@
 import os
 import subprocess
+import spaces
 
 # Install necessary packages if not already installed
 def install_packages():
@@ -10,56 +11,95 @@ def install_packages():
         except ImportError:
             subprocess.call(["pip", "install", package])
 
-install_packages()  # Execute installation before running the app
+install_packages()  # Install dependencies before running the app
 
 # Import required libraries
-from transformers import pipeline
+from transformers import pipeline, AutoModelForCausalLM, AutoTokenizer
 import gradio as gr
 import requests
+import torch
 
-# Load the image classification model from Hugging Face
-classifier = pipeline("zero-shot-image-classification", model="openai/clip-vit-base-patch32")
+# Load the image classification model
+classifier = pipeline("image-classification", model="umutbozdag/plant-identity")
 
-# API URL for retrieving plant information (Replace with the correct API endpoint)
-API_URL = "https://your-new-space-url.hf.space/api"
+
+# Function to get the appropriate text-generation model
+def get_model_name(language):
+    model_mapping = {
+        "English": "microsoft/Phi-3-mini-4k-instruct",
+        "Arabic": "ALLaM-AI/ALLaM-7B-Instruct-preview"
+    }
+    return model_mapping.get(language, "ALLaM-AI/ALLaM-7B-Instruct-preview")
+
+# Function to load the text-generation model
+def load_text_model(model_name):
+    device = "cuda" if torch.cuda.is_available() else "cpu"
+    model = AutoModelForCausalLM.from_pretrained(
+        model_name,
+        device_map=device,
+        torch_dtype="auto",
+        trust_remote_code=True,
+    )
+    tokenizer = AutoTokenizer.from_pretrained(model_name)
+    generator = pipeline(
+        "text-generation",
+        model=model,
+        tokenizer=tokenizer,
+        return_full_text=False,
+        max_new_tokens=500,
+        do_sample=False
+    )
+    return generator
+    
+@spaces.GPU
 
 # Function to classify plant images and fetch plant information
-def classify_plant(image):
-    labels = ["Rose", "Sunflower", "Tulip", "Daisy", "Dandelion"]  # List of plant labels
-    result = classifier(image, candidate_labels=labels)
-    plant_name = result["labels"][0]  # Get the highest predicted label
-
-    # API request to fetch plant information based on the predicted plant name
-    try:
-        response = requests.post(API_URL, json={"data": [plant_name]})
-        if response.status_code == 200:
-            plant_info = response.json()["data"][0]  # Extract information from API response
-        else:
-            plant_info = "⚠ No information available for this plant."
-    except:
-        plant_info = "❌ Error connecting to the Plant Info API."
+def classify_and_get_info(image, language):
+    result = classifier(image)
+
+    if result:  # Ensure result is not empty
+        plant_name = result[0]["label"]  # Extract the top predicted class
+    else:
+        return "Unknown", "Could not classify the plant."
+
+
+    # Load the appropriate text-generation model based on language
+    model_name = get_model_name(language)
+    text_generator = load_text_model(model_name)
+
+    # Define the prompt for plant information
+    prompt = (
+        f"Provide detailed information about {plant_name}. Include its scientific name, growing conditions, common uses, and care tips."
+        if language == "English"
+        else f"قدم معلومات مفصلة عن {plant_name}. اذكر اسمه العلمي، وظروف نموه، واستخداماته الشائعة، ونصائح العناية به."
+    )
+
+    messages = [{"role": "user", "content": prompt}]
+    output = text_generator(messages)
+
+    plant_info = output[0]["generated_text"] if output else "No detailed information available."
 
     return plant_name, plant_info
 
-# Gradio interface with a navy blue background
+# Gradio interface with a styled theme
 with gr.Blocks(css="""
     .gradio-container { 
-        background-color: #001f3f;  /* Navy blue background */
+        background-color: #d9ccdf;
         font-family: 'Arial', sans-serif; 
         color: white;
         text-align: center;
     }
     h1 {
-        color: #ffffff;
+        color: #333333;
         font-size: 32px;
         margin-bottom: 10px;
     }
     p {
-        color: #d3d3d3;
+        color: #181817;
         font-size: 18px;
     }
     .gradio-container .btn {
-        background-color: #00A86B !important;  /* Deep green button */
+        background-color: #00A86B !important;
         color: white !important;
         font-size: 18px;
         padding: 10px 20px;
@@ -74,7 +114,7 @@ with gr.Blocks(css="""
         font-size: 16px;
         font-weight: bold;
         color: #ffffff;
-        background-color: #003366;
+        background-color: #b69dc2;
         border: 2px solid #00A86B;
         padding: 10px;
         border-radius: 8px;
@@ -82,23 +122,24 @@ with gr.Blocks(css="""
 """) as interface:
     
     # App title and description
-    gr.Markdown("<h1>🌿 Plant Classifier AI</h1>")
-    gr.Markdown("<p>Upload a plant image to classify it and retrieve detailed information.</p>")
+    gr.Markdown("<h1>🌿 AI Plant Visión </h1>")
+    gr.Markdown("<p>Upload an image to identify a plant and retrieve detailed information in English or Arabic.</p>")
 
-    # Layout with columns for image input and classification results
+    # Layout for user input and results
     with gr.Row():
         with gr.Column(scale=1):
             image_input = gr.Image(type="pil", label="📸 Upload a Plant Image")
-            classify_button = gr.Button("🔍 Identify Plant & Get Info")
+            language_selector = gr.Radio(["English", "Arabic"], label="🌍 Choose Language", value="English")
+            classify_button = gr.Button("🔍 Identify & Get Info")
         with gr.Column(scale=1):
             plant_name_output = gr.Textbox(label="🌱 Identified Plant Name", interactive=False, elem_classes="textbox")
             plant_info_output = gr.Textbox(label="📖 Plant Information", interactive=False, lines=4, elem_classes="textbox")
 
-    # Connect the classify button to the function
-    classify_button.click(classify_plant, inputs=[image_input], outputs=[plant_name_output, plant_info_output])
+    # Connect button click to function
+    classify_button.click(classify_and_get_info, inputs=[image_input, language_selector], outputs=[plant_name_output, plant_info_output])
 
-    # Footer text
+    # Footer
     gr.Markdown("<p style='font-size: 14px; text-align: center;'>Developed with ❤️ using Hugging Face & Gradio</p>")
 
-# Launch the application
-interface.launch()
+# Launch the application with public link
+interface.launch(share=True)