metadata
license: other
widget:
- text: There was a ghost
example_title: First Prompt used in video
- text: I was playing Terraria but then
example_title: Second prompt used in video
inference:
parameters:
temperature: 0.6
repetition_penalty: 1.15
min_length: 128
max_length: 468
This is the model trained for this video:
https://www.youtube.com/watch?v=OEPL5Tm3mmQ
Due to hardware limitations, I trained this model with only a batch size of 2. (I know this isn't ideal).
The quality of the model may be affected.
After training was complete, the best model according to a hold-out set was used.
This model was trained using a filtered version of this dataset:
https://www.kaggle.com/datasets/thomaskonstantin/3500-popular-creepypastas
This dataset had a lot of blank entries and missing text.
Please subscribe to my YouTube Channel for bad quality videos and poorly trained models.