issues
search
alycialee
/
beyond-scale-language-data-diversity
Apache License 2.0
10
stars
12
forks
source link
issues
Newest
Newest
Most commented
Recently updated
Oldest
Least commented
Least recently updated
How are we guaranteeing that our model is right, shifting correctly?
#22
brando90
closed
1 month ago
3
make sure same token homogenous token code works to compute diversity coefficient
#20
brando90
opened
3 months ago
0
Fairseq Installation Fails: Missing version.txt Error During Build Pyhton 3.11 how to fix?
#19
brando90
opened
3 months ago
1
training gpt2 xl from stratch?
#18
brando90
closed
3 months ago
15
Larger LLM div computations and efficiently multi GPU multi proc plus running average wand.log
#17
brando90
opened
9 months ago
0
Coding div experiments
#16
brando90
opened
9 months ago
0
diverse data sets helps most when our target is a general data set, so when it's not diverse, the loss should be higher
#15
brando90
opened
9 months ago
0
better fitting as a cofounder of lower test loss
#14
brando90
opened
9 months ago
0
do data contamination experiemts to know it as a cofounder
#12
brando90
opened
9 months ago
0
running diversity/runner.sh scripts takes longer time in GPU server
#11
narayanal
closed
9 months ago
1
(v1) GPT-2 training and evaluation code
#10
SudharsanSundar
closed
9 months ago
4
notes for projects in
#9
brando90
closed
1 year ago
0
adding new api for div coeff
#8
brando90
closed
1 year ago
0
div coeff works
#7
brando90
closed
1 year ago
0
added alginemnt and cross diversity coeff
#6
brando90
closed
1 year ago
0
added alginemnt and cross diversity coeff
#5
brando90
closed
1 year ago
0
restructured to be a normal python project
#4
brando90
closed
1 year ago
0
How do the LLM DIVs compare to the original task2vec?
#3
brando90
closed
9 months ago
2
setup.py added and tutorial improved
#2
brando90
closed
1 year ago
0
link to tutorial quick start on main readme of repo needed
#1
brando90
closed
1 year ago
0