BPE Tokenizer
BabyLM-130M-ilm_it-ilm_pl-tokenizer_ilm_bilingual_it_pl_bpe_trailing_ws_32000BabyLM-130M-ilm_eu-tokenizer_ilm_mono_eu_bpe_32000BabyLM-130M-ilm_es-tokenizer_ilm_mono_es_bpe_32000BabyLM-130M-ilm_it-tokenizer_ilm_mono_it_bpe_32000BabyLM-130M-ilm_pl-tokenizer_ilm_mono_pl_bpe_32000BabyLM-130M-ilm_sl-tokenizer_ilm_mono_sl_bpe_32000BabyLM-130M-ilm_eu-ilm_it-tokenizer_ilm_bilingual_eu_it_bpe_32000BabyLM-130M-ilm_pl-ilm_sl-tokenizer_ilm_bilingual_pl_sl_bpe_32000
allenai_olmo2_superbpe_10G_1M_tokenizer_tulu-3-sft-olmo-2-mixture-0225-bpe-to-superbpesamples processed: 866138
Total user BPE tokens: 252198220
Total assistant BPE tokens: 402038966
Total assistant SuperBPE tokens: 198274197
Assistant reduction vs BPE = 50.68284077717979%
bigcode-bpe-tokenizerbpe-tokenizeruniversal-bpe
