DA-MORPH-CEREBRAS / README.md
MikkelWK's picture
Update README.md
7ea989e verified
|
raw
history blame
399 Bytes
metadata
library_name: transformers
tags: []

Mixed CerebrasGPT Danish Model

This is an experimental Danish language model fine-tuned on a combination of tokenizers, including both morphological and Byte-Pair-Encoding (BPE) approaches. Built on the CerebrasGPT-111M architecture, this model explores the impact of different tokenization strategies on Danish language understanding and generation.