Create your own GPT-2 model - train it with your own dataset or load the pre-trained weights.
torch==1.12.1
transformers==4.25.1
einops==0.6.0
fancy_einsum==0.0.3
tqdm==4.64.1
wandb==0.13.5
It is the run.ipnyb
file uploaded and ready to run in Google Colab.