Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

INVALID_ARGUMENT: Feature: path_samples/arc_length (data type: float) is required but could not be found. #25

Open
Sephirex-X opened this issue Oct 27, 2023 · 3 comments

Comments

@Sephirex-X
Copy link

Hi, when I set my conifg like:

conf = DatasetConfig(
    path='My_path_to_dataset',
    max_num_rg_points=20000,
    data_format=DataFormat.TFRECORD,
    max_num_objects=max_num_objects,
    num_shards = 48,
    repeat= 1,
    include_sdc_paths=True,
    num_paths= 1,
    num_points_per_path= 1,
)

I got this error:

2023-10-27 16:50:13.885355: W tensorflow/core/framework/op_kernel.cc:1828] OP_REQUIRES failed at example_parsing_ops.cc:98 : INVALID_ARGUMENT: Feature: path_samples/arc_length (data type: float) is required but could not be found.

I located the occurance on waymax/waymax/dataloader/dataloader_utils.py,line229:

for example in dataset.as_numpy_iterator():
      yield postprocess_fn(example)

It would be great if you can help me out there! Thx

@justinjfu
Copy link
Contributor

Hi @Sephirex-X,

This feature is meant to support routes for the SDC, but it's not public in the Waymo Open Dataset yet. We should have marked this flag with a "NotImplementedError".

@Sephirex-X
Copy link
Author

Sephirex-X commented Oct 28, 2023

Hi @Sephirex-X,

This feature is meant to support routes for the SDC, but it's not public in the Waymo Open Dataset yet. We should have marked this flag with a "NotImplementedError".
@justinjfu
Yeah I see, thank u 4 ur instant response, it's a very great work, however there r few things trouble me a lot.

  1. How can I use the SDC reward provided by ur framework in this case?
  2. Is there any ways that I can use the expert log on SDC to do my own reward calculation?
  3. I'm really confused about the relationship between missed routes for SDC (used to calculate reward for SDC I suppose) and the logged expert trajectory for SDC, is it possible u can provided some detail description?
  4. And finally, based on the code and data provided right now, is it possible to run some RL algorithms on SDC (how to evaluate if SDC has reached destination within a particular scenario ?).
    Many thx 4 ur time!

@hz3014
Copy link

hz3014 commented Aug 15, 2024

Hi @Sephirex-X , Did you find the answer to the questions?
I am also trying to use Route functionalities but limited by lacking of simulator_state.sdc_paths.

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Labels
None yet
Projects
None yet
Development

No branches or pull requests

3 participants