Created app.py

app.py

@@ -0,0 +1,52 @@
+import streamlit as st
+from transformers import AutoTokenizer, AutoModelForCausalLM
+import torch
+
+# Load the tokenizer and model from Hugging Face
+@st.cache_resource
+def load_model():
+    model_name = "meta-llama/Meta-Llama-3.1-70B-Instruct"
+    tokenizer = AutoTokenizer.from_pretrained(model_name)
+    model = AutoModelForCausalLM.from_pretrained(model_name, device_map="auto", torch_dtype=torch.float16)
+    return tokenizer, model
+
+tokenizer, model = load_model()
+
+# Supported languages
+languages = ['English', 'French', 'Spanish', 'Hindi', 'Punjabi']
+
+# Streamlit app
+def main():
+    st.title("Language Translator")
+
+    # User input for input language
+    input_language = st.selectbox("Select Input Language", languages)
+
+    # User input for output language
+    output_language = st.selectbox("Select Output Language", languages)
+
+    # Text input box for user to input text
+    input_text = st.text_area("Enter the text to translate")
+
+    if st.button("Translate"):
+        if input_text.strip() == "":
+            st.error("Please enter some text to translate.")
+        elif input_language == output_language:
+            st.warning("Input and output languages are the same. Please select different languages.")
+        else:
+            # Perform translation
+            translation = translate_text(input_text, input_language, output_language)
+            st.success("Translation:")
+            st.write(translation)
+
+# Function to translate text using the LLaMA model
+def translate_text(text, input_language, output_language):
+    prompt = f"Translate the following from {input_language} to {output_language}:\n\n{text}"
+    inputs = tokenizer(prompt, return_tensors="pt").to("cuda")
+    with torch.no_grad():
+        outputs = model.generate(**inputs, max_new_tokens=200)
+    translation = tokenizer.decode(outputs[0], skip_special_tokens=True)
+    return translation
+
+if __name__ == "__main__":
+    main()