SmartDJ-Editor Weights

[melo1998]

Awesome work, do you have pan to Release inference code for SmartDJ-Editor (diffusion editor)

[ZitongLan]

Hi @melo1998 thanks for your interest. We have released the weight at hugging face. You can run the following command to download it. let me know if you have any question:

mkdir -p pretrained_models
hf download ztlan/SmartDJ --local-dir pretrained_models

beyond this, we also release the inference code for editor, it is in editor_demo.py and we make it a interactive usage. You can simply run bash ./script/interactive_edit_editor.sh

[ZitongLan]

I have also added some detailed command guides, you can check the most updated commit 9770544

[melo1998]

OK, thank you very much!

[melo1998]

hello, have you any plan to release inference code for SmartDJ-Planer (ALM planer) and dataset synthesis pipeline?

[ZitongLan]

I’m currently tied up with a few upcoming deadlines, so progress on the release has been a bit slower than I’d like. That said, I’m actively working on it. My plan is to release: (i) the ALM planner inference code (including both our ALM and a proprietary model version), and (ii) the dataset synthesis pipeline before the end of this month. Thanks

[melo1998]

I’m currently tied up with a few upcoming deadlines, so progress on the release has been a bit slower than I’d like. That said, I’m actively working on it. My plan is to release: (i) the ALM planner inference code (including both our ALM and a proprietary model version), and (ii) the dataset synthesis pipeline before the end of this month. Thanks
Thank you very much for your feedback，I truly appreciate it and am excited to see your work!

[ZitongLan]

Hi, thank you for your patience and interest. I have updated a proprietary ALM model version for the whole pipeline in the new repo.

[melo1998]

Hi, thank you for your patience and interest. I have updated a proprietary ALM model version for the whole pipeline in the new repo.

Thank you so much for your great work and your valuable contributions to open source! We really appreciate it.
We’re also looking forward to the release of the data pipeline and the training/fine-tuning pipeline!