Open XiaoFengbing opened 6 months ago
Hi @XiaoFengbing,
Thank you for your question. I've shared the parameters we used in LongLLMLingua in issues #7 and #12, which you can refer to.
Hi, @iofu728, thanks for your response very much!
I confused about context_budget
after reading #7 and #12.
context_budget
are different number in LongLLMLingua, so i am confused.
So I will set context_budget
to be +200
in LongLLMLingua, thanks!
I see dynamic_context_compression_ratio
seems to be 0.25
by default in LongLLMLingua paper B.2 OTHER IMPLEMENTATION DETAILS. I am confused why it is set to be 0.3
or 0.4
in #7 and #12.
Thanks for the LongLLMLingua parameters you shared in issues #7 and #12. I want to know the LLMLingua parameters you used—— prompt = compressor.compress_prompt( context=xxx, instruction=xxx, question=xxx, ratio=0.75, iterative_size=100, context_budget="*1.5", )
Is it right?
Hi @XiaoFengbing, sorry for the late reply.
I read the issue #7, #12 and #49, I guess the right papameters that LLMLingua uses are :
prompt = compressor.compress_prompt( context=xxx, instruction=xxx, question=xxx, ratio=0.75, iterative_size=100, context_budget="*2", )
, and LongLLMLingua isprompt = compressor.compress_prompt( context=xxx, instruction=xxx, question=xxx, ratio=0.75, iterative_size=200, condition_compare=True, condition_in_question='after_condition', rank_method='longllmlingua', reorder_context='sort', dynamic_context_compression_ratio=0.3, context_budget="*2", )
I have some questions:context_budget
should be*1.3
or+300
in LongLLMLingua, and #12 ,you saidcontext_budget
should be+200
in LongLLMLingua, so I am confused by the setting ofcontext_budget
, and meanwhile, in the LLMLingua and LongLLMLingua papers, thecontext_budget
seems to be*2
(in #49). So I want to know how to setcontext_budget
in LLMLingua and LongLLMLingua.context_budget
andtoken_budget_ratio
can be considered part of the control coefficient parameterk
. Can I think I just need to control thecontext_budget
? Because in #7 and #12, you do not change thetoken_budget_ratio
.dynamic_context_compression_ratio
parameter correspond to in LongLLMLingua paper? I do not find it in the implementation details.Please forgive my too long questions, because your LLMLingua and LongLLMLingua work are very interesting for me! Looking forward to your reply.