Create app.py

app.py

@@ -0,0 +1,72 @@
+import tensorflow as tf
+import streamlit as st
+from transformers import T5Tokenizer, T5ForConditionalGeneration, MarianMTModel, MarianTokenizer
+
+def load_summarization_model():
+    model = T5ForConditionalGeneration.from_pretrained('t5-small')
+    tokenizer = T5Tokenizer.from_pretrained('t5-small')
+    return model, tokenizer
+
+summarization_model, summarization_tokenizer = load_summarization_model()
+
+def load_translation_models():
+    models = {}
+    tokenizers = {}
+    language_pairs = {
+        'en-fr': 'Helsinki-NLP/opus-mt-en-fr',
+        'fr-en': 'Helsinki-NLP/opus-mt-fr-en',
+        'en-de': 'Helsinki-NLP/opus-mt-en-de',
+        'de-en': 'Helsinki-NLP/opus-mt-de-en',
+        'en-hi': 'Helsinki-NLP/opus-mt-en-hi',
+        'hi-en': 'Helsinki-NLP/opus-mt-hi-en',
+        'fr-de': 'Helsinki-NLP/opus-mt-fr-de',
+        'de-fr': 'Helsinki-NLP/opus-mt-de-fr',
+        'fr-hi': 'Helsinki-NLP/opus-mt-fr-hi',
+        'hi-fr': 'Helsinki-NLP/opus-mt-hi-fr',
+        'de-hi': 'Helsinki-NLP/opus-mt-de-hi',
+        'hi-de': 'Helsinki-NLP/opus-mt-hi-de'
+    }
+    for pair, model_name in language_pairs.items():
+        models[pair] = MarianMTModel.from_pretrained(model_name)
+        tokenizers[pair] = MarianTokenizer.from_pretrained(model_name)
+    return models, tokenizers
+
+translation_models, translation_tokenizers = load_translation_models()
+
+def summarize_text(article):
+    inputs = summarization_tokenizer.encode("summarize: " + article, return_tensors="pt", max_length=512, truncation=True)
+    summary_ids = summarization_model.generate(inputs, max_length=64, min_length=10, length_penalty=2.0, num_beams=4, early_stopping=True)
+    return summarization_tokenizer.decode(summary_ids[0], skip_special_tokens=True)
+
+def translate_text(text, source_lang, target_lang):
+    if source_lang == target_lang:
+        return text
+    language_pair = f'{source_lang}-{target_lang}'
+    model = translation_models[language_pair]
+    tokenizer = translation_tokenizers[language_pair]
+    inputs = tokenizer(text, return_tensors="pt", padding=True, truncation=True)
+    translated = model.generate(**inputs)
+    return tokenizer.decode(translated[0], skip_special_tokens=True)
+
+
+st.title("Multilingual Text Summarizer and Translator")
+
+option = st.selectbox('Choose a task', ('Summarize and Translate'))
+
+if option == 'Summarize and Translate':
+    article = st.text_area("Enter the article text here:")
+    source_lang = st.selectbox('Source Language', ('en', 'fr', 'de', 'hi'))
+
+    if st.button("Generate Summary"):
+        summary = summarize_text(article)
+        st.write("Summary:", summary)
+
+        target_lang = st.selectbox('Translate Summary to', ('en', 'fr', 'de', 'hi'))
+        if st.button("Translate Summary"):
+            translated_summary = translate_text(summary, source_lang, target_lang)
+            st.write("Translated Summary:", translated_summary)
+
+        if st.checkbox("Translate Original Article"):
+            target_lang_article = st.selectbox('Translate Article to', ('en', 'fr', 'de', 'hi'), key='article')
+            translated_article = translate_text(article, source_lang, target_lang_article)
+            st.write("Translated Article:", translated_article)