Freedream-1B-Instruct-V2-preview

A preview version of Freedream-1B-Instruct-V2 has been released on Hugging Face under the Digilidaahz account, trained from the Freedream-1B base model that saw 24 billion tokens. The update expands the tokenizer vocabulary to 32,001 tokens, fixing the problem of '{' being tokenized as an unknown token, with credit given to @John6666. The poster is soliciting technical suggestions on the preview.

Hello guys, This is an preview version for Freedream-1B-Instruct-V2, it is trained based on the Freedream-1B 24B tokens seen base model where the tokenizer vocabulary became 32001 that fixes the problem of ‘{’ being tokenized as unk token, special thanks to @John6666 https://discuss.huggingface.co/u/john6666 , and the model is now available on: Digilidaahz/Freedream-1B-Instruct-V2 · Hugging Face https://huggingface.co/Digilidaahz/Freedream-1B-Instruct-V2 , any techinical suggestions are appreciated. Thanks for reading, have a good day.