You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
Hi, I find when I use env.observation I only get one image. However, the CMA model should use panorams, right ? Is ther someone konws how to get the panoramas ?
The text was updated successfully, but these errors were encountered:
Hi @Mingxiao-Li, neither the Seq2Seq or CMA models from the VLN-CE paper use panoramas. Both just use the egocentric single frame that you mention. If you would like to build different models that observe panoramas, I suggest looking into recent and ongoing development in Hab-Sim and Hab-Lab. In particular, there is some 360 support using the pipeline developed at facebookresearch/habitat-lab#478 and a fisheye camera here facebookresearch/habitat-lab#486. The API for these features may change as this is an active area of development.
Hi, I find when I use env.observation I only get one image. However, the CMA model should use panorams, right ? Is ther someone konws how to get the panoramas ?
The text was updated successfully, but these errors were encountered: