BalladGPT 4 (XL)
This is the beta version of BalladGPT. You can try the old BalladGPT here.
Usage
Note: BalladGPT's dataset is relatively small, so the quality is not optimal. If possible, please consider using a newer generation of GPT as these will equal or surpass this model in quality and speed. Please see the example results at the bottom of the page to evaluate this model's quality.
You can use BalladGPT as you would use any other GPT-2-based model:
# Make sure you logged in to Hugging Face!
from transformers import pipeline
generator = pipeline('text-generation', model='balladgpt/balladgpt-4-xl')
print(generator("Once upon a time", do_sample=True, min_length=20)[0]['generated_text'])
Biases
As this model was trained on select literature from the 18th, 19th, 20th century, certain biases and sterotypes may be expressed in generated texts which are neither endorsed or supported by the author of this model and are purely results of the literature used to train this model.
Disclaimer
The author of this model takes no responsibility for any content generated or any misuse of the model.
Training Data
BalladGPT was trained on various ballads and poems in the public domain.
BalladGPT is a fine-tuned version of GPT-2-XL.
The BalladGPT dataset is not available at the present moment.
Training Information
- Dataset Size: 1.2 MB
- Date Trained: May 12, 2023
Updates
We are not intending to update this model.
Example Results
Examples are coming soon