Simple gpt-2 Flask API

Usage:

1. Download initial pretrained model.

gpt2.download_gpt2(model_name="355M")

2. Setup training data (optional).

file_name = "yourdata.txt"
steps = 1000 #change this to your preference

3. Run main.py.
4. Generating content based on trained data.

#Visit localhost on the provided port using your finetuned model from ./checkpoints to generate.
/request/<model_name>/<generationCount>

Known issues:

• Rate limitation is not yet present, and as a result can allow multiple requests to saturate machine resources.