/GPT2-FairyTales

This project generate Fairytale using GPT-2 model.

Primary LanguageHTML

GPT2 Fairytale

Run on Ainize

This project generate Fairytale using GPT-2 model.

You can download the model here: Hugging face

Fine tuning data: Kaggle

Model information

Base model: gpt-2 large
Epoch: 30
Train runtime: 17861.6048 secs
Loss: 0.0412

How to use

* First, Fill what the base text. This will be base of Fairytale.
* And then, Fill number in length. Text is created as long as "length". I recommend between 100 and 300.
* If length is so big, generate time will be long.

Post parameter

text: The base of Fairytale.
length: The size of generated text.

Output format

{"0": generated text}

* With CLI *

Input example

curl -X 'POST' \
'https://feature-add-torch-serve-gpt-2-server-gkswjdzz.endpoint.ainize.ai/infer/GPT2-large_Fairytale' \
-H 'accept: */*' \
-H 'Content-Type: application/json' \
-d '{
"text": "One day,",
"num_samples": 1,
"length": 300
}'

Output example

{
  "0": "One day, when the King and his seven sons-in-law were in his court-house, and it was full of people, the young prince said to him, \"There are six thieves here in your court-house.\" \"Six thieves!\" said the"
}

* With swagger *

API page: Ainize

* With a Demo *

Demo page: End-point