issues
search
airoldilab
/
ai-sgd
Towards stability and optimality in stochastic gradient descent
5
stars
2
forks
source link
issues
Newest
Newest
Most commented
Recently updated
Oldest
Least commented
Least recently updated
Use publication quality plotting tools instead of ggplot2
#17
dustinvtran
opened
9 years ago
0
Run experiments on covertype data
#16
dustinvtran
opened
9 years ago
1
Benchmark using svrg.update() vs putting the code directly inside sgd()
#15
dustinvtran
opened
9 years ago
0
Modularity in sgd.R
#14
dustinvtran
closed
9 years ago
2
lambda parameter in sgd not working
#13
dustinvtran
opened
9 years ago
5
Add routine to search heuristically for optimal rates for AI-SGD
#12
ptoulis
closed
9 years ago
1
On the slope argument for the batch function
#11
dustinvtran
closed
9 years ago
0
Move the time benchmark function outside of exp_normal_correlated.R and make it generic
#10
dustinvtran
opened
9 years ago
2
Add signal-to-noise ratio for other GLMs in generate.data()
#9
dustinvtran
closed
9 years ago
0
Reorganize functions.R and disperse functions over multiple files
#8
dustinvtran
opened
9 years ago
1
Unit testing
#7
dustinvtran
opened
9 years ago
0
Remove need for defining sgd/batch in exp_normal_n5p2.R
#6
dustinvtran
opened
9 years ago
4
Need separate folder/better naming for experiments.
#5
ptoulis
closed
9 years ago
0
Averaged versions need to use larger learning rates.
#4
ptoulis
opened
9 years ago
5
Relevant datasets?
#3
ptoulis
closed
9 years ago
3
Implement averaged implicit (aiSGD) in sgd.R
#2
ptoulis
closed
9 years ago
0
batch learning needs update
#1
ptoulis
closed
9 years ago
0