mkshing e4t-diffusion issues

mkshing / e4t-diffusion

Implementation of Encoder-based Domain Tuning for Fast Personalization of Text-to-Image Models

https://arxiv.org/abs/2302.12228

MIT License

317 stars 24 forks source link

issues

Newest

Newest Most commented Recently updated Oldest Least commented Least recently updated

images and resources for pre-training

#31 yudongjian opened 7 months ago
0
Question about settings and results.

#30 zhanjiahui opened 10 months ago
0
{placeholder_token} vs *s in inference

#29 bsun0802 opened 10 months ago
1
Problem with running the code both in colab and local machine

#28 alimohammadiamirhossein opened 11 months ago
1
Upgrade to SD 1.5 or later

#27 blistick opened 1 year ago
0
can you release the cat model

#26 silence401 opened 1 year ago
0
Need help... OOM with 2 RTX3090 (bs=2)

#25 CHR-ray opened 1 year ago
1
BUG in inference.py

#24 CHR-ray opened 1 year ago
1
Bugs in the pretrain script

#23 shenzy08 opened 1 year ago
0
when I finetuned the model

#22 zhangqizky opened 1 year ago
2
how can i get the unet.pt file?

#21 Parisa-Bagherzadeh closed 1 year ago
1
can a get a cat model?

#20 redredbluee closed 1 year ago
1
Handle case where ckpt_path is None, i.e. when running pretrain_e4t.py

#19 brianfitzgerald opened 1 year ago
0
I fine-tuned the same yannlecun image as yours, and the generated results are different from yours. I used the pre-trained model e4t-diffusion-ffhq-celebahq-v1 you released. Why are the results different?

#18 wangzhaolei closed 1 year ago
8
Prompt used during pretraining

#17 Actis92 opened 1 year ago
1
finetuning-free personalized text-to-image model

#16 iRedBean closed 1 year ago
1
The performance of my pretraining model is very bad.

#15 dong-y21 closed 1 year ago
7
CUDA out of memory when running pretrain of branch v1 even if I set batch_size to 1. I wonder how much GPU memory is required to train this model. And does this mean I can't train on the V100(32G)?

#14 dong-y21 closed 1 year ago
5
I tried to train faces use FFHQ and celebA dataset. Why is the image sampled in the training completely different from the input image face? When I input an image during the inference, the generated face is completely different from the input face. And why domain_embed_scale=0.1 in training？ Thankyou！

#13 wangzhaolei closed 1 year ago
2
big changes for better results

#12 mkshing closed 1 year ago
0
fix a bug in pipe

#11 mkshing closed 1 year ago
0
question about inference.py

#10 laolongboy closed 1 year ago
2
how to process dataset

#9 lsjiiia closed 1 year ago
1
request a pretrained model

#8 laolongboy closed 1 year ago
2
a Pre-training problem

#7 YibooZhao closed 1 year ago
13
Problem due to wrong dimensions in E4TEncoder

#6 Actis92 closed 1 year ago
2
Can you specify versions in requirements.txt?

#5 cutlass90 closed 1 year ago
2
fix clip pooling part to align with the paper

#4 mkshing closed 1 year ago
0
difference in the structure of encoder

#3 XavierXiao closed 1 year ago
6
When can we use the pertained model?

#2 songyiren98 closed 1 year ago
1
Create LICENSE

#1 mkshing closed 1 year ago
0