search

sherwinbahmani / 4dfy

4D-fy: Text-to-4D Generation Using Hybrid Score Distillation Sampling
https://sherwinbahmani.github.io/4dfy/
Apache License 2.0
288 stars 7 forks source link

train_dynamic_camera #10

Closed ByChelsea closed 4 months ago

ByChelsea commented 4 months ago

Hi, could you please explain the purpose of the code related to "train_dynamic_camera"? It appears to have never been used?

sherwinbahmani commented 4 months ago

Hi, the plan was to simulate dynamic cameras, as in the text-to-video model the videos have camera and scene motion. So I tried sampling a sequence of nearby cameras. But I could not get it working but left it in the code if someone wants to get it working

ByChelsea commented 4 months ago

I understand, thank you for your prompt reply.