issues
search
jayelm
/
gisting
Learning to Compress Prompts with Gist Tokens - https://arxiv.org/abs/2304.08467
Apache License 2.0
259
stars
23
forks
source link
issues
Newest
Newest
Most commented
Recently updated
Oldest
Least commented
Least recently updated
look me!
#19
Tshor
closed
4 months ago
1
inference detail
#18
Yxxxb
closed
5 months ago
2
does gist token support Question Answering use-case with static context, varying question, with 24K+ context length?
#17
IamExperimenting
closed
4 months ago
1
unable to run the demo
#16
Hannibal046
closed
10 months ago
1
About partial compression of instruction
#15
YoojuShin
closed
10 months ago
2
when I use model.generate (), the gist_mask metrix seems don't work
#14
pengfeiwu1999
closed
11 months ago
5
Conceptual Clarification
#13
FFFiend
closed
11 months ago
1
Any plans to create some sort of gisting framework for arbitrary models?
#12
FFFiend
closed
1 year ago
2
eval llama-7b from huggingface
#11
QeeKhan
closed
1 year ago
2
why there's a parameter "offset"?
#10
pengfeiwu1999
closed
1 year ago
9
Unable to reproduce LLaMA-7B results when training from scratch
#9
Xiuyu-Li
closed
1 year ago
23
AssertionError when recovering llama-7b-pos_control-1
#8
Xiuyu-Li
closed
1 year ago
4
Unable to reproduce results using LLaMA-7B gist checkpoint
#7
Xiuyu-Li
closed
1 year ago
2
ConfigValueError when Training smaller T5 Model according to READ.ME
#6
FFFiend
closed
1 year ago
26
gist for unseen prompt/task
#5
La-SilverLand
closed
1 year ago
3
Running out of memory when trying to compress
#4
FFFiend
closed
1 year ago
25
Is Gisting accessible as a Module?
#3
FFFiend
closed
1 year ago
0
Could you provide more training details?
#2
aopolin-lv
closed
1 year ago
2
Update README.md
#1
eltociear
closed
1 year ago
1