initial commit: research summarizer with FLAN-T5

app.py

@@ -0,0 +1,40 @@
+import requests
+import gradio as gr
+from bs4 import BeautifulSoup
+from transformers import AutoTokenizer, AutoModelForSeq2SeqLM, pipeline
+
+# Lightweight model for instruction-tuned summarization
+tokenizer = AutoTokenizer.from_pretrained("google/flan-t5-base")
+model = AutoModelForSeq2SeqLM.from_pretrained("google/flan-t5-base")
+
+llm_pipeline = pipeline("text2text-generation", model=model, tokenizer=tokenizer)
+
+def extract_arxiv_abstract(url):
+    try:
+        headers = {"User-Agent": "Mozilla/5.0"}
+        response = requests.get(url, headers=headers)
+        soup = BeautifulSoup(response.text, "html.parser")
+        abstract = soup.find("blockquote", class_="abstract")
+        if abstract:
+            return abstract.get_text(strip=True).replace("Abstract:", "")
+        else:
+            return "Abstract not found on this page. Please check the URL."
+    except Exception as e:
+        return f"Failed to fetch abstract: {str(e)}"
+
+def summarize_research_paper(arxiv_url):
+    abstract_text = extract_arxiv_abstract(arxiv_url)
+    if abstract_text.startswith("Failed"):
+        return abstract_text
+
+    prompt = f"summarize: {abstract_text}"
+    summary = llm_pipeline(prompt, max_new_tokens=256)[0]['generated_text']
+    return f"📄 Original Abstract:\n{abstract_text}\n\n🧠 Summary:\n{summary}"
+
+gr.Interface(
+    fn=summarize_research_paper,
+    inputs=gr.Textbox(label="arXiv Paper URL", placeholder="https://arxiv.org/abs/2306.10001"),
+    outputs=gr.Textbox(label="Summary", lines=15),
+    title="🧠 Research Paper Summarizer",
+    description="Summarizes arXiv paper abstracts using FLAN-T5. Works fast on CPU Hugging Face Spaces."
+).launch()