128K sequence length means 131072 or 128000
syp1997 opened this issue · 1 comments
syp1997 commented
128K sequence length means 131072 or 128000
hsiehjackson commented
Normally, it means 131072 but some models use 128000 (like GPT4).
syp1997 opened this issue · 1 comments
Normally, it means 131072 but some models use 128000 (like GPT4).