Update app.py

app.py

@@ -17,10 +17,10 @@ if hf_api_key is None:
 else:
     headers = {"Authorization": f"Bearer {hf_api_key}"}
 
-# Text-to-image model URL
-API_URL = "https://api-inference.huggingface.co/models/CompVis/stable-diffusion-v1-4"
+# Use a faster text-to-image model for quicker generation
+API_URL = "https://api-inference.huggingface.co/models/runwayml/stable-diffusion-v1-5"
 
-# Load the text generation model to maintain the same functionality
+# Load the text generation model
 text_generation_model_name = "EleutherAI/gpt-neo-1.3B"
 text_tokenizer = AutoTokenizer.from_pretrained(text_generation_model_name)
 text_model = AutoModelForCausalLM.from_pretrained(text_generation_model_name)
@@ -32,7 +32,7 @@ text_generator = pipeline("text-generation", model=text_model, tokenizer=text_to
 def generate_image_from_text(translated_text):
     try:
         print(f"Generating image from translated text: {translated_text}")
-        response = requests.post(API_URL, headers=headers, json={"inputs": translated_text}, timeout=15)  # Reduced timeout to 15 seconds
+        response = requests.post(API_URL, headers=headers, json={"inputs": translated_text}, timeout=20)  # Increased timeout to 20 seconds
 
         if response.status_code != 200:
             print(f"Error generating image: {response.text}")
@@ -51,8 +51,14 @@ def generate_image_from_text(translated_text):
 def generate_short_paragraph_from_text(translated_text):
     try:
         print(f"Generating a short paragraph from translated text: {translated_text}")
-        # Generate a shorter paragraph using a lower max_length for faster processing
-        paragraph = text_generator(translated_text, max_length=100, num_return_sequences=1, temperature=0.6, top_p=0.9)[0]['generated_text']
+        paragraph = text_generator(
+            translated_text, 
+            max_length=80,  # Reduced to 80 tokens
+            num_return_sequences=1, 
+            temperature=0.7, 
+            top_p=0.9,
+            no_repeat_ngram_size=3  # Prevent repetitive text
+        )[0]['generated_text']
         print(f"Paragraph generation completed: {paragraph}")
         return paragraph
     except Exception as e:
@@ -65,8 +71,8 @@ def translate_generate_paragraph_and_image(tamil_text):
     try:
         print("Translating Tamil text to English...")
         tokenizer.src_lang = "ta_IN"
-        inputs = tokenizer(tamil_text, return_tensors="pt", max_length=50, truncation=True)  # Limited max_length to 50 for faster translation
-        translated_tokens = model.generate(**inputs, forced_bos_token_id=tokenizer.lang_code_to_id["en_XX"])
+        inputs = tokenizer(tamil_text, return_tensors="pt", max_length=40, truncation=True)  # Limit length to 40 tokens
+        translated_tokens = model.generate(**inputs, max_length=50, num_beams=4, early_stopping=True)  # Use beam search for faster output
         translated_text = tokenizer.batch_decode(translated_tokens, skip_special_tokens=True)[0]
         print(f"Translation completed: {translated_text}")
     except Exception as e:
@@ -84,7 +90,7 @@ def translate_generate_paragraph_and_image(tamil_text):
 
     return translated_text, paragraph, image, None
 
-# Gradio interface setup without `enable_queue`
+# Gradio interface setup
 iface = gr.Interface(
     fn=translate_generate_paragraph_and_image,
     inputs=gr.Textbox(lines=2, placeholder="Enter Tamil text here..."),
@@ -94,7 +100,7 @@ iface = gr.Interface(
     title="Tamil to English Translation, Short Paragraph Generation, and Image Creation",
     description="Translate Tamil text to English using Facebook's mbart-large-50 model, generate a short paragraph, and create an image using the translated text.",
     theme="default",
-    allow_flagging="never"  # Disable flagging to reduce UI load
+    allow_flagging="never"
 )
 
 # Launch the interface