metadata
license: cc-by-sa-4.0
datasets:
- facebook/belebele
Pretrained toy model. Made with Andrej Karpathy's NanoGPT, ~2023. Trained on part of Tagalog portion of Belebele.
Parameters:
- batch_size = 64
- block_size = 256
- n_layer = 8
- n_head = 8
- n_embd = 768
Everything else is left as is.