File size: 1,481 Bytes
36c09a4
3dfb09d
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
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
import streamlit as st
from transformers import AutoTokenizer, AutoModelForSeq2SeqLM
import torch

# Load model and tokenizer
def load_model():
    model = AutoModelForSeq2SeqLM.from_pretrained("majorSeaweed/T5-Conv_summarisation_small")
    tokenizer = AutoTokenizer.from_pretrained("majorSeaweed/T5-Conv_summarisation_small")
    return tokenizer, model

tokenizer, model = load_model()

# Title
st.title("🧠 Dialogue Summarisation (T5)")
st.markdown(
    "This app generates **non-toxic summaries** of multi-turn conversations using a SFT Tuned `t5-small` model."
)

# Input box
dialogue = st.text_area(
    "Paste a conversation below (multi-turn dialogue with speaker turns):",
    height=200,
    placeholder="User: Hey, you there?\nBot: Yeah, what's up?\nUser: I just wanted to say you're terrible at this.",
)

# Generate summary
if st.button("Generate Summary"):
    if not dialogue.strip():
        st.warning("Please enter a conversation.")
    else:
        with st.spinner("Generating summary..."):
            inputs = tokenizer(dialogue, return_tensors="pt", truncation=True, max_length=512)
            summary_ids = model.generate(
                **inputs,
                max_length=60,
                num_beams=4,
                early_stopping=True,
                no_repeat_ngram_size=2
            )
            summary = tokenizer.decode(summary_ids[0], skip_special_tokens=True)
            st.subheader("📝 Summary")
            st.success(summary)