Unable to Generate TF Records

Hi,

Trying to train TLT ResNet18SSD model on a custom data-set. Converted the Pascal VOC Annotations to Kitti Format using xml2kitti. Here is the reference used

TF-records kitti txt

kitti_config {
root_directory_path: “/workspace/tlt-experiments/data/training”
image_dir_name: “image_2”
label_dir_name: “label_2”
image_extension: “.jpg”
partition_mode: “random”
num_partitions: 2
val_split: 14
num_shards: 10
}
image_directory_path: “/workspace/tlt-experiments/data/training”

The following error occurred while running in the jupyter notebook.
!tlt-dataset-convert -d $SPECS_DIR/ssd_tfrecords_kitti_trainval.txt
-o $USER_EXPERIMENT_DIR/tfrecords/kitti_trainval

Traceback (most recent call last):
File “/usr/local/bin/tlt-dataset-convert”, line 10, in
sys.exit(main())
File “./detectnet_v2/scripts/dataset_convert.py”, line 64, in main
File “./detectnet_v2/dataio/dataset_converter_lib.py”, line 74, in convert
File “./detectnet_v2/dataio/dataset_converter_lib.py”, line 108, in _write_partitions
File “./detectnet_v2/dataio/dataset_converter_lib.py”, line 149, in _write_shard
File “./detectnet_v2/dataio/kitti_converter_lib.py”, line 169, in _create_example_proto
File “./detectnet_v2/dataio/kitti_converter_lib.py”, line 272, in _add_targets
AssertionError: Ground truth kitti labels should have only 15 fields.

Please help me with this issue.

Thanks in Advance

Hi pagr,
Ground truth kitti labels should have only 15 fields. Please double check your label file.

Just a small confirmation Is it 15 fields including labels or it should have 15 after label

Hi psgr,
Could you please see tlt doc section 4.2.2? There is a description about KITTI format.