---
license: mit
language: en
tags:
  - text-generation
  - chatbot
  - t5
pipeline_tag: text2text-generation
---

# T5 Chatbot

A fine-tuned T5 model for conversational FAQ-style responses. Given a user question, the model generates a natural-language answer, making it suitable for lightweight chatbot and Q&A applications.

## Model description

This model is a fine-tuned version of **[t5-small / t5-base]** (Google's T5 text-to-text transformer), adapted for conversational question-answering. It takes a user query as input and generates a relevant text response, framed as a text-to-text generation task.

## Intended uses & limitations

**Intended uses:**
- FAQ-style chatbots for websites, apps, or customer support
- Educational/demo projects exploring conversational AI with T5
- Quick prototyping of Q&A systems

**Limitations:**
- Trained on a limited dataset, so responses may be generic or repetitive outside the training domain
- Does not maintain multi-turn conversational context (treats each query independently)
- English only
- Not suitable for safety-critical or factual/medical/legal advice use cases
- May occasionally produce inaccurate or nonsensical answers (hallucination risk common to generative models)

## Training data

The model was fine-tuned on a **[custom FAQ dataset / dataset name, e.g. "a collection of customer support Q&A pairs"]**. **[Add: dataset size, source, and any preprocessing steps if known.]**

## How to use

```python
from transformers import pipeline

pipe = pipeline("text2text-generation", model="UMAR798/t5-chatbot")
response = pipe("What are your business hours?")
print(response)
```

## Training procedure

- Base model: **[t5-small / t5-base]**
- Epochs: **[e.g. 3]**
- Batch size: **[e.g. 8]**
- Learning rate: **[e.g. 5e-5]**

## Author

Developed by [Muhammad Umar Farooq](https://huggingface.co/UMAR798) — [LinkedIn](https://www.linkedin.com/in/muhammad-umar-farooq-6964a430b)