blastwind
/

codegen-scala

codegen-scala / tokenizer.json

50k vocab size instead of 200k to match gpt-2

2a1f4da over 1 year ago

2.23 MB

File too large to display, you can check the raw version instead.