TencentARC Open-MAGVIT2 issues

TencentARC / Open-MAGVIT2

Open-MAGVIT2: Democratizing Autoregressive Visual Generation

Apache License 2.0

711 stars 29 forks source link

issues

Newest

Newest Most commented Recently updated Oldest Least commented Least recently updated

questions about the autoregressive model vocab size

#41 sysuyy closed 3 weeks ago
2
数据集格式怎么放

#40 Amber530 opened 4 weeks ago
0
gpu训练很慢

#39 syyxsxx closed 1 month ago
1
There is a problem with the inference result

#38 wen020 closed 1 month ago
1
Any plan for 3D magvit-v2?

#37 txytju closed 2 months ago
1
Could you improve the mushroom shape in the sample?

#36 TTPlanetPig closed 1 month ago
1
Curious About Subspace Size

#35 Lqf-HFNJU closed 2 months ago
2
Make it a package

#34 vinyesm closed 2 months ago
1
Could we make it a package?

#33 vinyesm closed 2 months ago
3
About the results without cfg

#32 youngsheen closed 2 months ago
1
Face loss for face reconstruction

#31 kl2004 opened 3 months ago
1
UNet Model

#30 JeremyCJM closed 2 months ago
1
Which version of ImageNet?

#29 JeremyCJM closed 3 months ago
2
Questions about the update timeline

#28 Liam-lyr closed 3 months ago
1
Better to put `self.model_ema = LitEma(self)` before `self.init_from_ckpt(...)`?

#27 AIWanderer-X closed 2 months ago
1
Why not use `Adaptive GroupNorm` On Decoder & Sampler asymmetrical?

#26 xesdiny closed 3 months ago
6
Why not use `per_sample_entropy` loss for backward?

#25 xesdiny closed 4 months ago
1
training bash script bug for 128 x 128

#24 deuterium1729 closed 2 months ago
0
No Checkpoint found after running run_128_B.sh

#23 wd255 opened 4 months ago
0
Training phase sanity check fails by loading "../../data/imagenet/val" as an image

#22 wd255 opened 4 months ago
0
Can we write `_sigmoid_cross_entropy_with_logits` in a simple way?

#21 AIWanderer-X closed 2 months ago
1
`logits_real = logits_fake.reshape(B, -1)`: is this on purpose or just a bug?

#20 AIWanderer-X closed 4 months ago
1
Auto-Regressive Generative Model with MagVitV2 Tokenizer

#19 youngsheen closed 4 months ago
1
Training loss curves?

#18 shinshiner closed 2 months ago
5
Training phase sanity check fails by loading "../../data/imagenet/val" as an image

#17 wd255 closed 4 months ago
1
Create setup.py file to make it easy to import magvit2 into other projects

#16 ericjang closed 4 months ago
1
Initializing distributed halts for a long time and then timeout

#15 wd255 closed 4 months ago
1
Could you Vs VQGAN-LC benchmark?

#14 xesdiny closed 4 months ago
1
Open-MAGVIT2 achieves SOTA??? I just cannot reproduce the numbers reported in the main page.

#13 juncongmoo closed 4 months ago
16
Noisy points in reconstruction results

#12 shinshiner closed 5 months ago
4
about the result of magvit2

#11 yyyouy closed 5 months ago
1
[BUG] reconstruct from index number instead of quant

#10 eisneim closed 5 months ago
1
ckpt ffhq

#9 kepengxu closed 5 months ago
1
Is it possible to also align magvit2 with a text encoder, as what you do in seed?

#8 StarCycle closed 5 months ago
5
How long does it take to train this model?

#7 StarCycle closed 5 months ago
3
encoder和decoder的结果不一致。

#6 Mr-Harry closed 5 months ago
3
[question] any plans to train higher compression ratio?

#5 eisneim closed 5 months ago
3
Inquiry Regarding Codebook Selection Method in Open-MAGVIT2

#4 zhaohm14 closed 5 months ago
2
Do you compare your ckpt with stability's?

#3 StarCycle closed 5 months ago
4
Will consider using Web images train a more robust version?

#2 MonolithFoundation closed 5 months ago
11
Reconstruction result

#1 Epiphqny closed 5 months ago
2