issues
search
LTH14
/
mar
PyTorch implementation of MAR+DiffLoss https://arxiv.org/abs/2406.11838
MIT License
745
stars
39
forks
source link
issues
Newest
Newest
Most commented
Recently updated
Oldest
Least commented
Least recently updated
Request for Causal AR Version Release
#39
aengusng8
opened
4 hours ago
0
CFG for cross-entropy
#38
shaochenze
closed
6 hours ago
2
About Training Loss
#37
Ferry1231
opened
23 hours ago
4
About Train
#36
pokameng
opened
1 day ago
4
Doesn't work well in speech generation task.
#35
FacePoluke
opened
2 days ago
7
model and training code for the AR variant
#34
MikeWangWZHL
opened
2 days ago
1
Is Autoencoder ok?
#33
Ferry1231
opened
3 days ago
1
Add HF integration to MAR
#32
jadechoghari
opened
4 days ago
10
The CFG strategy - linear. vs constant
#31
yuhuUSTC
opened
1 week ago
9
Difference Between MAR and MAGE
#30
JeremyCJM
opened
1 week ago
5
Faster training with fp16 or bf16
#29
shaochenze
closed
1 week ago
2
generate images with arbitrary resolutions,
#28
Leiii-Cao
opened
1 week ago
4
Questions about causal methods
#27
Tom-zgt
closed
2 weeks ago
5
Inference details of an ablation experiment.
#26
tgxs002
closed
2 weeks ago
1
VAE decoded as NaN in early stages of training
#25
xiazhi1
closed
2 weeks ago
2
Why main_cache do not use flip augment?
#24
xiazhi1
closed
3 weeks ago
2
How should inference be performed when using VQ-16 (discrete)? During decoding, should we use the AR output for VQ and then decode?
#23
Tom-zgt
closed
3 weeks ago
2
About the mask schedule during training
#22
zythenoob
closed
3 weeks ago
4
Buffer Size for Class Condition
#21
zhuole1025
closed
3 weeks ago
9
Question on the Value of Training Loss for DiffuLoss with MAR and Causal Methods
#20
bugWholesaler
opened
4 weeks ago
18
Train Code for VAE Used in Paper
#19
Ferry1231
opened
1 month ago
16
MAR for Image-to-Image Generation
#18
Bili-Sakura
closed
4 weeks ago
3
Develop
#17
YecanLee
closed
1 month ago
0
Training settings for MAR series
#16
HuangOwen
closed
1 month ago
2
Latent Dimensions of VAE
#15
Vinnieassaulter
closed
1 month ago
4
why the inputs of diffusion are multiplied by diffusion_batch_mul?
#14
Erisura
opened
1 month ago
3
Loss for training KL-VAE
#13
Vinnieassaulter
closed
1 month ago
1
Training epochs
#12
sihyun-yu
closed
1 month ago
2
About the "per-token" distribution modeled by diffusion model
#11
chrisway613
opened
1 month ago
4
Why not whole DIT block?
#10
WeitaoLu
opened
1 month ago
2
Is diffusion position embedding necessary?
#9
chrisway613
opened
1 month ago
2
FID evaluation reference data
#8
MArSha1147
opened
1 month ago
18
add grad checkpointing
#7
Jiawei-Yang
closed
1 month ago
1
Generation FID is much lower than Reconstruction FID for models using VQ-16 (discrete) provided by LDM codebase
#6
ShiFengyuan1999
closed
1 month ago
4
The Impact of MLP Depth
#5
Robertwyq
closed
1 month ago
2
Class-Conditional Free Text-Guided Generation
#4
Jake-wei
closed
1 month ago
4
About the VAE
#3
Haochen-Wang409
closed
1 month ago
6
Where is the 0.2325 from ?
#2
LZY-the-boys
closed
1 month ago
1
Fix online evaluation
#1
byronyi
closed
1 month ago
1