issues
search
konstmish
/
prodigy
The Prodigy optimizer and its variants for training neural networks.
MIT License
296
stars
17
forks
source link
issues
Newest
Newest
Most commented
Recently updated
Oldest
Least commented
Least recently updated
Text Encoder LR Scale for Stable Diffusion Training
#19
sangoi-exe
opened
2 months ago
2
Question on convergence
#18
ppbrown
opened
4 months ago
2
Can we combine this profitably with Schedule-Free?
#17
JDvorak
opened
5 months ago
1
Lowering TE or Unet average only
#16
trihardseven
opened
5 months ago
1
Growth_rate
#15
DarkAlchy
opened
6 months ago
0
Is Prodigy compatible with Pytorch's automatic mixed precision?
#14
crypdick
opened
6 months ago
0
Document incompatibility with gradient clipping
#13
crypdick
opened
6 months ago
2
Is there a way to monitor the estimated LR over time, if it has any meaning?
#12
ethansmith2000
closed
6 months ago
1
Possible to marry Prodigy and AdamW?
#11
askerlee
opened
8 months ago
8
T_MAX value (CosineAnnealingLR)
#10
josemerinom
closed
9 months ago
3
Layer-wise scaling
#9
adefazio
opened
9 months ago
2
Recommended Prodigy Settings (low steps per epoch)
#8
brandostrong
closed
10 months ago
8
Remove pdb as a dependency
#7
patrickvonplaten
closed
10 months ago
0
Question regarding t_max and d estimation
#6
DanPli
closed
10 months ago
1
Is the any rule of the thumb for tuning weight_decay of Prodigy when training transformers-based LLMs?
#5
DesperateExplorer
closed
12 months ago
3
Question
#4
Dentoty
closed
1 year ago
3
Question
#3
DarkAlchy
closed
10 months ago
38
module 'torch.optim' has no attribute 'Prodigy'
#2
DarkAlchy
closed
1 year ago
4
RuntimeError: Expected all tensors to be on the same device, but found at least two devices, cuda:0 and cpu!
#1
manyotherfunctions
opened
1 year ago
1