lipurple / grounded-diffusion Goto Github PK

Thanks a lot for your great work! May I know what is the inference speed for generating grounded images?

when I run the code

python train.py --class_split 1 --train_data random --save_name pascal_1_random

FileNotFoundError: mmdetection/checkpoint/mask_rcnn_swin-s-p4-w7_fpn_fp16_ms-crop-3x_coco_20210903_104808-b92c91f1.pth can not be found.

Thanks for the great work!

At the moment, the provided train.py seems to be hardwired to train on the Pascal VOC dataset. Is there a plan to release the COCO training script that can be used to reproduce results in the paper?

when i run Inference,
python test.py --sd_ckpt 'lipurple/stable_diffusion.ckpt'
--grounding_ckpt 'lipurple/grounding_module.pth'
--prompt "a photo of a lion on a mountain top at sunset"
--category "lion"

and will occur ：RuntimeError: PytorchStreamReader failed reading zip archive: failed finding central directory.

can you give me some advices?

thanks

Thanks for your excellent work!

I am confused about the definition of open-vocabulary segmentation from two aspects:

1. I note that the segmentation model (i.e., maskformer in the paper) is trained on full categories of PASCAL VOC and COCO while the data are synthetic from the Stable Diffusion.
2. Can open-vocabulary segmentation protocol access the complete categories during training? In my opinion, the unseen(novel) class name should only be available at the test instead of training time. Otherwise, it is not really open-vocabulary.

Hope the authors could give me some help to make me better understand this paper!

Thanks!

I follow the readme use python train.py --class_split 1 --train_data random --save_name pascal_1_random ' to train the model and generate the checkpoints；now how to evaluate them? I dont find the evalution code in you project.

python test.py --sd_ckpt 'xxx/stable_diffusion.ckpt'
--grounding_ckpt 'xxx/grounding_module.pth'
--prompt "a photo of a lion on a mountain top at sunset"
--category "lion"

您好，执行下面这行代码出现上述错误，可能是stable_diffusion.ckpt该ckpt文件有问题。
能否帮忙确认一下，谢谢

Hello, Thank you for your excellent work.I am very interested in the teaser input images, it is the following pictures,

Can you post them?

How long does it take to retrain this model?
是否有并行训练这个模型的方法？一张卡上训练的话好像需要很久很久；

python train.py --class_split 1 --train_data random --save_name pascal_1_random,
and No such file or directory: 'models/ldm/stable-diffusion-v1/model.ckpt',

can you give me some advices?

which version of torch is it？