Regarding the results in Table 8 and Table 14
Statisticss opened this issue · 0 comments
Statisticss commented
Hi,
I have a question regarding the results in Table 8 and Table 14 (05 Dec., 2023 version).
In Table 8, for 7B context length 8192, the ppl for full FT is 8.02 and the ppl for LongLoRA is 8.04.
In Table 14, for 7B context length 8192, the ppl for full FT is 6.98 and the ppl for LongLoRA is 7.14.
According to the description, I think the only difference between Table 8 and 14 is, for Table 8 you used PG19 validation set and for Table 14 you used PG19 test set. Am I understanding correctly? If so, why is there such a big difference between validation and test set?