Difference between bilingual-embedding-large and bilingual-embedding-large-8k
#1
by thomlevy - opened
Hello,
Thanks for this model.
This doesn't seem obvious from the model card.
What is the difference between bilingual-embedding-large and bilingual-embedding-large-8k?
What does the "8k" suffix means?
Thanks
Hi @thomlevy
Token input of bilingual-embedding-large is 512 tokens, and bilingual-embedding-large-8k is 8096 tokens.
Tuan
Hi @dangvantuan
Thanks. Very clear.
May be it could help if you could mention this updated max_seq_length in the model card:
https://huggingface.co/Lajavaness/bilingual-embedding-large-8k#full-model-architecture
Thomas
thomlevy changed discussion status to closed