Spaces:
Sleeping
Sleeping
File size: 2,104 Bytes
c414752 b55eee3 c414752 |
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 |
import streamlit as st
from transformers import AutoTokenizer, AutoModelForCausalLM
import torch
from huggingface_hub import login
# Authenticate with Hugging Face using the environment variable
huggingface_token = os.getenv("HUGGINGFACE_TOKEN")
login(token=huggingface_token)
# 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()
|