Update app.py

app.py

@@ -1,36 +1,36 @@
-# app.py 
+# app.py - Final Version with English Comments and UI Text
 import gradio as gr
 import torch
 from transformers import AutoTokenizer, AutoModelForSeq2SeqLM
 
-# --- 1. fetching model from huggingface ---
+# --- 1. Define model and tokenizer paths ---
 model_name = "doganbilir/mt5-Turkish-English-Summarizer" 
-
 device = torch.device("cuda" if torch.cuda.is_available() else "cpu")
 
-# --- 2. download model from hub ---
+# --- 2. Load model and tokenizer from the Hub ---
 try:
     tokenizer = AutoTokenizer.from_pretrained(model_name)
     model = AutoModelForSeq2SeqLM.from_pretrained(model_name).to(device)
-    print("Model ve tokenizer başarıyla yüklendi!")
+    print("Model and tokenizer loaded successfully!")
 except Exception as e:
-    print(f"Model yüklenirken hata oluştu: {e}")
+    print(f"An error occurred while loading the model: {e}")
     model = None
 
-# --- 3. Define summarize function ---
+# --- 3. Define the summarization function ---
 def summarize(text, language_choice, min_length):
     if model is None:
-        return "Model yüklenemediği için özetleme yapılamıyor. Lütfen Space loglarını kontrol edin."
+        return "Model could not be loaded. Please check the Space logs."
     if not text or not text.strip():
-        return "Lütfen özetlemek için bir metin girin."
-
-    if language_choice == "Türkçe Metin -> Türkçe Özet":
+        return "Please enter text to summarize."
+    
+    # Set the prompt based on the language choice
+    if language_choice == "Turkish Text -> Turkish Summary":
         prompt = f"summarize: {text}"
-    elif language_choice == "İngilizce Metin -> İngilizce Özet":
+    elif language_choice == "English Text -> English Summary":
         prompt = f"summarize: {text}"
-    else: # TR -> EN (Deneysel)
+    else: # TR -> EN (Experimental)
         prompt = f"summarize Turkish to English: {text}"
-
+        
     try:
         inputs = tokenizer(prompt, return_tensors="pt", max_length=1024, truncation=True).to(device)
         
@@ -39,7 +39,7 @@ def summarize(text, language_choice, min_length):
             attention_mask=inputs.attention_mask,
             max_new_tokens=150,
             num_beams=5,
-            min_new_tokens=int(min_length), # Slider'dan gelen değeri kullan
+            min_new_tokens=int(min_length), # Use the value from the slider
             early_stopping=True,
             no_repeat_ngram_size=3
         )
@@ -47,32 +47,34 @@ def summarize(text, language_choice, min_length):
         summary = tokenizer.batch_decode(summary_ids, skip_special_tokens=True, clean_up_tokenization_spaces=True)[0]
         return summary.strip()
     except Exception as e:
-        return f"Özetleme sırasında bir hata oluştu: {str(e)}"
+        return f"An error occurred during summarization: {str(e)}"
 
-# --- 4. Create gradio Interface ---
+# --- 4. Create the Gradio Interface ---
 iface = gr.Interface(
     fn=summarize,
     inputs=[
         gr.Textbox(lines=15, placeholder="Paste the text you want to summarize here...", label="Text Input"),
         gr.Radio(
             ["Turkish Text -> Turkish Summary", "English Text -> English Summary", "Turkish Text -> English Summary (Experimental)"],
-            label="Summary Type",
-            value="Turkish Text -> Turkish Summary" 
+            label="Summarization Type",
+            value="Turkish Text -> Turkish Summary"
         ),
         gr.Slider(
             minimum=10, 
             maximum=100, 
-            value=10, #default text size to create short summaries
+            value=10, # Default value for natural, short summaries
             step=5, 
             label="Minimum Summary Length (Tokens)",
             info="Increase this value to force the model to generate longer summaries."
         )
     ],
-    outputs=gr.Textbox(lines=5, label="Oluşturulan Özet"),
+    outputs=gr.Textbox(lines=5, label="Generated Summary"),
     title="Multilingual Text Summarization Model",
     description="""This demo generates headline-style summaries for Turkish or English texts.
-The model is fine-tuned on 30,000 Turkish-English summary pairs using LoRA, 
-based on the `google/mt5-small` model.""",
+    The model was fine-tuned on 30,000 Turkish-English summary pairs using the LoRA technique, 
+    based on the `google/mt5-small` model.""",
+    
+    # The values in the examples list are now 100% consistent with the Radio button choices.
     examples=[
         ["İstanbul’da açılan yeni bilim merkezi, hem çocuklar hem de yetişkinler için interaktif deneyimler sunuyor. Ziyaretçiler, sanal gerçeklik ve artırılmış gerçeklik teknolojilerini kullanarak uzay, biyoloji ve fizik konularını deneyimleyebiliyor. Merkez, özellikle eğitim kurumlarıyla yaptığı iş birliği sayesinde öğrencilerin bilimsel merakını artırmayı hedefliyor. Açılışından bu yana, binlerce kişi merkezi ziyaret etti ve sosyal medyada paylaşılan videolar büyük ilgi gördü.", "Turkish Text -> Turkish Summary", 10],
         ["NASA recently announced that the James Webb Space Telescope has captured highly detailed images of distant exoplanets, allowing scientists to analyze their atmospheres in unprecedented detail. The observations reveal clues about planetary composition, weather patterns, and potential habitability. This milestone represents a significant step forward in our understanding of planets outside our solar system, and researchers plan to continue studying these celestial bodies to uncover more about the diversity and nature of exoplanets.", "English Text -> English Summary", 10]