Jordi Salvador

The current supported way to get metrics for validation tasks is to run the test mode with those tasks, but we'll consider this for future releases.

Hi @pioneer-innovation. Thank you so much for your interest in ProcTHOR and AllenAct. ProcTHOR support in AllenAct is being worked on, so we'll hopefully provide an initial version soon.

Let's discuss about the usage of Memory to make sure it's sufficient for ANM.

Hi @suvaansh, Just for a sanity check, can you try running with `CUDA_VISIBLE_DEVICES=0,1,2,3,4,5,6,7`?

May I also ask you to share the output of `nvidia-smi` and `ps aux | grep Xorg` after `sudo ai2thor-xorg start`?

That looks good 👍 I would try to start training without headless mode, if that's an option. I would use a subset of the available x_displays in https://github.com/allenai/allenact/blob/474fb84789fcdd3917fbc006365653971994e93f/projects/objectnav_baselines/experiments/objectnav_thor_base.py#L237 In my...

`worker_devices` sounds great to me (in engine we also use `worker_id`)

ProcTHOR is officially supported by the [latest release of AI2THOR](https://github.com/allenai/ai2thor/releases/tag/5.0.0). Even though ProcTHOR support in AllenAct was prototyped, I ended up convinced that external projects are a better solution moving...

Hi @mariiak2021. It's currently not possible to use the `ProcTHOR`-related code base, since the underlying 10k houses dataset isn't yet publicly available. We're hoping to release everything soon, though, so...

I think @mattdeitke or @Lucaweihs would have more accurate answers to these questions, but here's what I know: The currently missing dataset is composed of 10,000 ProcTHOR one and two-room...