DiffusionTAD
DiffusionTAD copied to clipboard
sauradip
When the code will be released?
Thank you for the interest in our work ! I am currently writing my thesis, so it's difficult to make time but expect the code release in the last week of September.
Hi @sauradip, I would be interested in the code, too. Do you have any estimate when it could be released? Thanks a lot!
Thanks all for the interest. November Last Week the code should be uploaded. Apologies for the delay !
Are there issues on code? It's been a while... Could you notice when the code is released?
Hi all thisbwork happened during my PhD ending phase and since i graduated i lost access to the system to test it properly. This is the reason i could not make the code public. I am very sorry for the delay , i am also trying hard to put up a file on the repo soon.
If you need a heads up i can tell you how to implement : you can clone the Actionformer paper, replace the CNN decoder with Transformer Decoder , and then use the codebase of DiffusionDet to do the noise to proposal denoising on this DETR style Actionformer model. For the selective conditioning, you can refer to React (ECCV 22) paper for the similarity part which will be implemented in the self-conditioning part. You can refer to BitDiffusion paper for self conditioning. Its a simple trick.
I'm really impressed by your work and am curious to know when you plan to open source the code. It would be incredibly helpful
Hi, thanks for reaching out.It´s a shame that you lost access to your own work. Since this was done during your Ph.D., such a work should be publicly available, since universities are financed from public money. Fingers crossed for putting it back tohether.Best, Jan.13. 1. 2024 v 10:48, Sauradip Nag @.***>: Hi all thisbwork happened during my PhD ending phase and since i graduated i lost access to the system to test it properly. This is the reason i could not make the code public. I am very sorry for the delay , i am also trying hard to put up a file on the repo soon. If you need a heads up i can tell you how to implement : you can clone the Actionformer paper, replace the CNN decoder with Transformer Decoder , and then use the codebase of DiffusionDet to do the noise to proposal denoising on this DETR style Actionformer model. For the selective conditioning, you can refer to React (ECCV 22) paper for the similarity part which will be implemented in the self-conditioning part. You can refer to BitDiffusion paper for self conditioning. Its a simple trick.
—Reply to this email directly, view it on GitHub, or unsubscribe.You are receiving this because you commented.Message ID: @.***>