main.py: error: argument action: invalid choice: '--tasks semantic scene instance orientation' (choose from 'resume')

[shllgtca]

I'm trying to run the evaluation on NYUv2 (without pretraining on Hypersim), but I keep receiving this error, I had the problem discussed in #16

So I commented the line, run it again and receivedthe error:

main.py: error: argument action: invalid choice: '--tasks semantic scene instance orientation' (choose from 'resume')

It looks like the "--tasks" which is assigned as "semantic scene instance orientation", but it should be resume, correct?

[Tripton]

It appears that the command itself may be broken, and the issue might not be related to the "--tasks" argument. Could you please provide the full command you're trying to execute so that we can understand the problem?

[shllgtca]

The exact same command that is suggested on the readme file. The fist command.

python main.py
--dataset nyuv2
--dataset-path ./datasets/nyuv2
--tasks semantic scene instance orientation
--enable-panoptic
--input-modalities rgb depth
--rgb-encoder-backbone resnet34
--rgb-encoder-backbone-block nonbottleneck1d
--depth-encoder-backbone resnet34
--depth-encoder-backbone-block nonbottleneck1d
--no-pretrained-backbone
--weights-filepath ./trained_models/nyuv2/r34_NBt1D.pth
--checkpointing-metrics valid_semantic_miou bacc mae_gt_deg panoptic_deeplab_semantic_miou panoptic_all_deeplab_pq
--validation-batch-size 16
--validation-only
--skip-sanity-check
--wandb-mode disabled

[Tripton]

Could it be that you changed the --dataset-path from what's mentioned in the readme file? The error you encountered might be due to incorrect escaping of newline or whitespace characters in the command (using \ or "", respectively). Since the --tasks argument cannot be parsed correctly, I assume the error to be before that part of the arguments.

If this doesn't help, could you please provide more details about your setup (e.g. the operating system you are using, python version, ...)?