Create app.py

app.py

@@ -0,0 +1,63 @@
+from transformers import M2M100ForConditionalGeneration, M2M100Tokenizer
+from transformers import MBartForConditionalGeneration, MBart50Tokenizer
+from transformers import pipeline
+import gradio as gr
+import requests
+import io
+from PIL import Image
+import os
+import torch  # For LLaMA text generation
+
+# Load the translation model and tokenizer
+model_name = "facebook/mbart-large-50-many-to-one-mmt"
+tokenizer = MBart50Tokenizer.from_pretrained(model_name)
+model = MBartForConditionalGeneration.from_pretrained(model_name)
+
+# Load the LLaMA model for text generation
+model_id = "meta-llama/Llama-3.2-1B"  # Use LLaMA model for text generation
+pipe = pipeline(
+    "text-generation", 
+    model=model_id, 
+    torch_dtype=torch.bfloat16,  # Using bfloat16 for reduced memory footprint
+    device_map="auto"  # Automatically assign devices for multi-GPU or CPU fallback
+)
+
+# Use the Hugging Face API key from environment variables for text-to-image model
+API_URL = "https://api-inference.huggingface.co/models/ZB-Tech/Text-to-Image"
+headers = {"Authorization": f"Bearer {os.getenv('hf_tokens')}"}
+
+# Define the translation, text generation, and image generation function
+def translate_and_generate_image(tamil_text):
+    # Step 1: Translate Tamil text to English using mbart-large-50
+    tokenizer.src_lang = "ta_IN"
+    inputs = tokenizer(tamil_text, return_tensors="pt")
+    translated_tokens = model.generate(**inputs, forced_bos_token_id=tokenizer.lang_code_to_id["en_XX"])
+    translated_text = tokenizer.batch_decode(translated_tokens, skip_special_tokens=True)[0]
+
+    # Step 2: Generate descriptive English text using LLaMA model
+    generated_text = pipe(translated_text, max_length=100, num_return_sequences=1)[0]['generated_text']
+    
+    # Step 3: Use the generated English text to create an image
+    def query(payload):
+        response = requests.post(API_URL, headers=headers, json=payload)
+        return response.content
+
+    # Generate image using the generated text
+    image_bytes = query({"inputs": generated_text})
+    image = Image.open(io.BytesIO(image_bytes))
+
+    return translated_text, generated_text, image
+
+# Gradio interface setup
+iface = gr.Interface(
+    fn=translate_and_generate_image,
+    inputs=gr.Textbox(lines=2, placeholder="Enter Tamil text here..."),
+    outputs=[gr.Textbox(label="Translated English Text"), 
+             gr.Textbox(label="Generated Descriptive Text"),
+             gr.Image(label="Generated Image")],
+    title="Tamil to English Translation, Text Generation with LLaMA, and Image Creation",
+    description="Translate Tamil text to English using Facebook's mbart-large-50 model, generate descriptive text using Meta's LLaMA model, and create an image using the generated text.",
+)
+
+# Launch Gradio app with a shareable link
+iface.launch(share=True)