Kai Wang

@hydradrianfyh Do you have any updates? I met some similar problem like:  Thank you.

@abhigoku10 thanks for your question, though I'm not the owner of this repo. For (1) I would say, yes. just refer to the code in src/dataset and custom your dataset...

@abhigoku10 For Q1, Depth is not required for panorama generation. As for a minimal dataset implementation, Prompts, Image GT is required which are key inputs for SD. If you wanna...

@abhigoku10 1. yes. you have some multi-view data (like shot from different angle at on scene), right? in practice you can use some other method (for example, cv2 findHomography) to...

This answer would help: https://stackoverflow.com/questions/63232589/compute-homography-matrix-based-on-intrinsic-and-extrinsic-camera-parameters

hi @Sodayeyeye, to generate longer music under current setup is actually pretty doable. Tranfering a super simple "copy & paste" trick that I introduced in this [paper](https://arxiv.org/abs/2407.19468) (wasn't for audio...

@dorienh yes, please check this forked [repo](https://github.com/kkaiwwana/mustango), I provided many generated [examples](https://github.com/kkaiwwana/mustango/tree/main/assets/demo_audio).

fgsm stands for Fast-Gradient-Sign-Method, which is an optimization/updating method for white-box attack. It generates adversarail patterns based on gradient values (specifically, its sign, i.e. positive or negtive). Hope you find...

I checked the random state (specifically, it's `py_random` object in this very specific case) of `Compose` object, and it truly didn't pass on to next state when excuted in a...

Thanks. Here are some extra information, let me quote my comment at #2473 first, > pytorch indeed distributes unique yet controllable seeds (if you have called torch.manual_seed()) to different workers...