google-research/causal-agents

Google Research

This repository contains code released by Google Research.

All datasets in this repository are released under the CC BY 4.0 International license, which can be found here: https://creativecommons.org/licenses/by/4.0/legalcode. All source files in this repository are released under the Apache 2.0 license, the text of which can be found in the LICENSE file.

---

Because the repo is large, we recommend you download only the subdirectory of interest:

• Use GitHub editor to open the project. To open the editor change the url from github.com to github.dev in the address bar.
• In the left navigation panel, right-click on the folder of interest and select download.

If you'd like to submit a pull request, you'll need to clone the repository; we recommend making a shallow clone (without history).

git clone [email protected]:google-research/google-research.git --depth=1

---

Disclaimer: This is not an official Google product.

Updated in 2023.

causal-agents's Issues

How to read the causal-label?

Issue Description:

Hi, thanks for your amazing work! I'm encountering an issue while parsing TFRecord files using TensorFlow. I'm unable to parse the data within them, and I'm getting the following error message:

Error Message:

2023-09-08 15:07:54.635147: W tensorflow/core/framework/op_kernel.cc:1830] OP_REQUIRES failed at example_parsing_ops.cc:94: INVALID_ARGUMENT: Could not parse example input, value: '
...
InvalidArgumentError: {{function_node _wrapped__ParseExampleV2_Tdense_0_num_sparse_2_ragged_split_types_0_ragged_value_types_0_sparse_types_2_device/job:localhost/replica:0/task:0/device:CPU:0}} Could not parse example input, value: '

Specifications:

• TensorFlow Version: 2.0.0
• Operating System: Ubuntu 20.04

Background:

I'm encountering an issue while attempting to parse TFRecord files using TensorFlow. The error message suggests a problem with parsing the example input.

Issue Details:

I've defined the dataset description, including variable-length features (VarLenFeature), but it seems there's an issue during parsing. Here's a code snippet:

filename = "causal_labels.tfrecord"
dataset = tf.data.TFRecordDataset(filename, compression_type='')
data = next(dataset.as_numpy_iterator())

features_description = {}
causal_features = {
    'scenario_id': tf.io.VarLenFeature(tf.string),
    'labeler_results': tf.io.VarLenFeature(tf.string),
}

features_description.update(causal_features)
parsed = tf.io.parse_single_example(data, features_description)

Expected Outcome:

I hope to successfully parse the data within the TFRecord file for further data processing and analysis.

Please help me identify the cause of the parsing issue and how to resolve it. Thank you for your assistance!

Cannot load the perturbed dataset

Behavior

I am using the Waymo Open Dataset Motion Tutorial to visualize the perturbed validation dataset, but when creating the dataset I got the error below, which I suspect that the roadgraph_samples/* features are missing, is this expected?

---------------------------------------------------------------------------
InvalidArgumentError                      Traceback (most recent call last)
[<ipython-input-27-64935651141a>](https://localhost:8080/#) in <cell line: 3>()
      1 dataset = tf.data.TFRecordDataset(FILENAME, compression_type='')
      2 data = next(dataset.as_numpy_iterator())
----> 3 parsed = tf.io.parse_single_example(data, features_description)

1 frames
[/usr/local/lib/python3.10/dist-packages/tensorflow/python/eager/execute.py](https://localhost:8080/#) in quick_execute(op_name, num_outputs, inputs, attrs, ctx, name)
     50   try:
     51     ctx.ensure_initialized()
---> 52     tensors = pywrap_tfe.TFE_Py_Execute(ctx._handle, device_name, op_name,
     53                                         inputs, attrs, num_outputs)
     54   except core._NotOkStatusException as e:

InvalidArgumentError: {{function_node __wrapped__ParseExampleV2_Tdense_55_num_sparse_0_ragged_split_types_0_ragged_value_types_0_sparse_types_0_device_/job:localhost/replica:0/task:0/device:CPU:0}} Key: roadgraph_samples/xyz.  Can't parse serialized Example. [Op:ParseExampleV2]

Steps to Reproduce the Problem

1. Just to make sure the file is not corrupted I downloaded one single file directly from v1.1/validation/RemoveNoncausal/RemoveNoncausalAgents.tfrecord-00000-of-01127 and uploaded to this tutorial to run.
2. I tried with the dataset from WODM: validation_tfexample.tfrecord-00000-of-00150 and it worked.

Dropping all objects in the `state/tracks_to_predict` field

Hi,

I have just realized that in some of the scenes, all of the objects in the state/tracks_to_predict are dropped as a result, if that is the case people can't simply use those prediction tags to filter out the agents to predict right, we need to label some of our own agents to predict.

Thanks