BPE Tokenizer Model trained on the BabyLM dataset with a vocab size of 8192.