RRT-Foundation / tokenizer_pointer.txt
Tripstoph's picture
Open release: RRT-355M weights, CORE eval artifacts, README
4f983a9 verified
Raw
History Blame Contribute Delete
151 Bytes
Tokenizer: openai-community/gpt2 (canonical GPT-2 BPE, vocab 50257).
Load via `transformers.AutoTokenizer.from_pretrained('openai-community/gpt2')`.