README.md

metadata

title: T5 Email Summarizer Demo v3
emoji: 📧
colorFrom: blue
colorTo: green
sdk: gradio
sdk_version: 4.44.1
app_file: app.py
pinned: true
license: apache-2.0
models:
  - wordcab/t5-small-email-summarizer
datasets:
  - argilla/FinePersonas-Conversations-Email-Summaries
space_hardware: cpu-basic

T5 Email Summarizer - Interactive Demo v3

This Space provides an interactive demo of the wordcab/t5-small-email-summarizer model.

🔧 v3 Major Updates

• Separate Subject/Body fields for better email structure
• General title normalization (Mr. → Mr, Dr. → Dr, Prof. → Prof)
• Improved unicode handling for special characters
• Robust preprocessing that handles all edge cases

Features

• 🎯 Dual-mode summarization: Brief (1-2 sentences) or Full (detailed)
• 🚀 Fast inference: Quick processing even on CPU
• 💪 Robust: Handles typos, abbreviations, and informal language
• 🔄 Auto-detect: Automatically chooses brief or full based on email length
• 🔧 Smart preprocessing: General solution for title and unicode issues

Model Details

• Architecture: T5-small (60M parameters)
• Training Data: argilla/FinePersonas-Conversations-Email-Summaries (364K examples)
• Max Input: 512 tokens (~2500 characters)
• License: Apache 2.0

Try It Out

1. Enter subject line (optional) and email body separately
2. Select summary type or use auto-detect
3. Click "Generate Summary"

The model will produce a concise, accurate summary with automatic normalization of titles and special characters!