Add: [AI-ML] AI-Text-Rewriter

Domains/AI-ML/MiniProjects/AI_Rewriter/main.py

@@ -0,0 +1,39 @@
+import streamlit as st
+# autotokenizer for converting words into token
+# automedelforseq2seq loads a sequnenvce to sequence model for text generation and rewrting
+from transformers import AutoTokenizer, AutoModelForSeq2SeqLM
+
+@st.cache_resource #load model from cache if exists
+def load_model():
+    model_name = "google/flan-t5-small"
+    tokenizer = AutoTokenizer.from_pretrained(model_name) # laods tokenizer
+    model = AutoModelForSeq2SeqLM.from_pretrained(model_name) # loads actual neural network
+    return tokenizer, model
+
+tokenizer , model = load_model()
+
+st.title("🧠 AI Text Rewriter")
+st.write("Rewrite your text into different tones using AI (FLAN-T5).")
+
+user_text = st.text_area("Enter text to rewrite", height=150)
+style = st.selectbox("Choose rewriting style", ["Formal", "Casual", "Persuasive"])
+
+if st.button("Rewrite Text"):
+    if user_text.strip() == "":
+        st.warning("Please enter some text first!") 
+    else:
+        prompt = f"Rewrite the following text in a {style.lower()} tone:\n\n{user_text}"
+        # return sensors as torches
+        # max_length to limit input size
+        # truncation to cut off extra long inputs 
+        inputs = tokenizer(prompt, return_tensors="pt", max_length=512, truncation=True)
+
+        outputs = model.generate(**inputs, max_length=256, temperature=0.7)
+        # decode tokens and skip special tokens
+        rewritten = tokenizer.decode(outputs[0], skip_special_tokens=True)
+
+        st.subheader("✨ Rewritten Text:")
+        st.write(rewritten)
+
+
+