Thanks for your contribution on accommondating qwen on self-extend. Qwen1.5 has al

<a target="_blank" rel="noopener noreferrer" href="https://private-user-images.githubu

What effect on qwen1.5 will be if i use self-extend trick? about longlm HOT 3 OPEN

datamllab commented on August 11, 2024 1

What effect on qwen1.5 will be if i use self-extend trick?

from longlm.

Comments (3)

Mooler0410 commented on August 11, 2024 4

We believe how good is self extend highly depends on how good is the extended model within its original pretraining context window. This means if Qwen1.5's 32k context window is not well trained, SelfExtend may not work. Otherwise, it will work well. [Currently, we have no plan for a serious test considering the massive computational resource requirement: 32k 8x -> 256k, 4x -> 128k. We may do serious benchmarking for Qwen1.5 when we have enough resources.]

from longlm.

WeixuanXiong commented on August 11, 2024 2

Results on 128k length in str (around 70k as for qwen tokenizer). It seems work!

from longlm.

WeixuanXiong commented on August 11, 2024 1

We believe how good is self extend highly depends on how good is the extended model within its original pretraining context window. This means if Qwen1.5's 32k context window is not well trained, SelfExtend may not work. Otherwise, it will work well. [Currently, we have no plan for a serious test considering the massive computational resource requirement: 32k 8x -> 256k, 4x -> 128k. We may do serious benchmarking for Qwen1.5 when we have enough resources.]

Ahhh, if I test it in the future work, i'll share it with you guys. Thanks for your reply~

from longlm.

Recommend Projects