🧠 GPT-2 Inference API Development

In this project, I used the HuggingFace GPT-2 model for inference using FastAPI and tested on Google Colab.

---

🚀 1. GPT-2 Inference Server-side (FastAPI + ngrok)

Run this notebook to:

• Load a GPT-2 model using the HuggingFace transformers library
• Test the FastAPI using ngrok

👉 GPT-2 Inference Server Google Colab

🛠️ Features

• /generate endpoint for single-prompt inference
• /batch_generate endpoint for multi-prompt batch inference
• Build your own prototype or demos to further test it
• Experimenting with quantization, batching, and latency

---

🧪 2. API Client-end Testing

Use this notebook to:

• Send test requests to your public FastAPI endpoint
• Measure tokens generated and latency
• Support both single and batch inference calls

👉 Client-end Testing Google Colab

---

🙌 Credits

Built with:

• ⚡ FastAPI – High-performance Python web framework for building APIs
• 🤗 Hugging Face Transformers – State-of-the-art NLP models
• 🧠 Google Colab – Free cloud notebooks with GPU support
• 🌐 ngrok – Public URLs for localhost APIs