Hi,
Firstly thank you for the TLT general release! For those of us (me including) who couldn’t make it to the Early Access Program, this is a wait which is finally over and I am really excited!
After downloading and installing the TLT docker image and preparing my dataset as per the instructions in the getting started guide, I ran a training cycle using the “tlt_resnet18_detectnet_v2_v1” model.
I am fine-tuning the resnet18 model on a single class called ‘customer’.
However at the end of the first eval cycle I get this error:
2019-09-26 21:22:13,642 [INFO] tensorflow: global_step/sec: 2.55173
INFO:tensorflow:epoch = 0.9527027027027027, loss = 0.008371737, step = 141 (5.488 sec)
2019-09-26 21:22:14,041 [INFO] tensorflow: epoch = 0.9527027027027027, loss = 0.008371737, step = 141 (5.488 sec)
2019-09-26 21:22:16,915 [INFO] iva.detectnet_v2.evaluation.evaluation: step 0 / 36, 0.00s/step
/usr/local/lib/python2.7/dist-packages/iva/detectnet_v2/evaluation/metadata.py:38: UserWarning: One or more metadata field(s) are missing from ground_truth batch_data, and will be replaced with defaults: ['frame/camera_location']
2019-09-26 21:22:28,510 [INFO] iva.detectnet_v2.evaluation.evaluation: step 10 / 36, 1.16s/step
2019-09-26 21:22:36,094 [INFO] iva.detectnet_v2.evaluation.evaluation: step 20 / 36, 0.76s/step
2019-09-26 21:22:43,573 [INFO] iva.detectnet_v2.evaluation.evaluation: step 30 / 36, 0.75s/step
Traceback (most recent call last):
File "/usr/local/bin/tlt-train-g1", line 10, in <module>
sys.exit(main())
File "./common/magnet_train.py", line 37, in main
File "</usr/local/lib/python2.7/dist-packages/decorator.pyc:decorator-gen-2>", line 2, in main
File "./detectnet_v2/utilities/timer.py", line 46, in wrapped_fn
File "./detectnet_v2/scripts/train.py", line 632, in main
File "./detectnet_v2/scripts/train.py", line 556, in run_experiment
File "./detectnet_v2/scripts/train.py", line 490, in train_gridbox
File "./detectnet_v2/scripts/train.py", line 136, in run_training_loop
File "/usr/local/lib/python2.7/dist-packages/tensorflow/python/training/monitored_session.py", line 676, in run
run_metadata=run_metadata)
File "/usr/local/lib/python2.7/dist-packages/tensorflow/python/training/monitored_session.py", line 1270, in run
raise six.reraise(*original_exc_info)
File "/usr/local/lib/python2.7/dist-packages/tensorflow/python/training/monitored_session.py", line 1255, in run
return self._sess.run(*args, **kwargs)
File "/usr/local/lib/python2.7/dist-packages/tensorflow/python/training/monitored_session.py", line 1335, in run
run_metadata=run_metadata))
File "./detectnet_v2/tfhooks/validation_hook.py", line 69, in after_run
File "./detectnet_v2/tfhooks/validation_hook.py", line 75, in validate
File "./detectnet_v2/evaluation/evaluation.py", line 179, in evaluate
File "./detectnet_v2/evaluation/compute_metrics.py", line 161, in __init__
File "./detectnet_v2/evaluation/compute_metrics.py", line 343, in _prepare_internal_structures
File "./detectnet_v2/evaluation/compute_metrics.py", line 301, in _check_if_bbox_is_valid
KeyError: 'customer'
my training commnad is:
tlt-train detectnet_v2 --gpus 1 \
-r /workspace/nvidia_experiment/training_output \
-e /workspace/nvidia_experiment/training.config \
-n nvidia_experiment_1 \
-k $MY_API_KEY
and my training config is:
dataset_config {
data_sources: {
tfrecords_path: "/workspace/nvidia_experiment/dataset/tfrecords/*"
image_directory_path: "/workspace/nvidia_experiment/dataset/"
}
image_extension: "jpeg"
target_class_mapping {
key: "customer"
value: "customer"
}
validation_fold: 0
}
augmentation_config {
preprocessing {
output_image_width: 640
output_image_height: 480
output_image_channel: 3
min_bbox_width: 1.0
min_bbox_height: 1.0
}
spatial_augmentation {
hflip_probability: 0.5
zoom_min: 1.0
zoom_max: 1.0
translate_max_x: 8.0
translate_max_y: 8.0
}
color_augmentation {
hue_rotation_max: 25.0
saturation_shift_max: 0.2
contrast_scale_max: 0.1
contrast_center: 0.5
}
}
model_config {
# Model architecture can be chosen from:
# ['resnet', 'vgg', 'googlenet', 'alexnet', 'mobilenet_v1', 'mobilenet_v2', 'squeezenet']
arch: "resnet"
pretrained_model_file: "/workspace/nvidia_experiment/tlt_resnet18_detectnet_v2_v1/resnet18.hdf5"
#we are freezing the first two conv block to maintain features learnt in pretraining
freeze_blocks: 0
freeze_blocks: 1
all_projections: True
# for resnet --> n_layers can be [10, 18, 50]
# for vgg --> n_layers can be [16, 19]
num_layers: 18
#use_bias: True
use_pooling: False
use_batch_norm: True
dropout_rate: 0.0
freeze_bn: False
training_precision: {
backend_floatx: FLOAT32
}
objective_set: {
cov {}
bbox {
scale: 35.0
offset: 0.5
}
}
}
training_config {
batch_size_per_gpu: 16
num_epochs: 80
learning_rate {
soft_start_annealing_schedule {
min_learning_rate: 5e-6
max_learning_rate: 5e-4
soft_start: 0.1
annealing: 0.7
}
}
regularizer {
type: L1
weight: 3e-9
}
optimizer {
adam {
epsilon: 1e-08
beta1: 0.9
beta2: 0.999
}
}
cost_scaling {
enabled: False
initial_exponent: 20.0
increment: 0.005
decrement: 1.0
}
checkpoint_interval: 10
}
evaluation_config {
average_precision_mode: INTEGRATE
validation_period_during_training: 10
first_validation_epoch: 1
minimum_detection_ground_truth_overlap {
key: "customer"
#value is the IoU value
value: 0.5
}
}
bbox_rasterizer_config {
target_class_config {
key: "customer"
value {
cov_center_x: 0.5
cov_center_y: 0.5
cov_radius_x: 0.4
cov_radius_y: 0.4
bbox_min_radius: 1.0
}
}
deadzone_radius: 0.67
}
postprocessing_config {
target_class_config {
key: "customer"
value {
clustering_config {
coverage_threshold: 0.005
dbscan_eps: 0.13
dbscan_min_samples: 0.05
minimum_bounding_box_height: 4
}
}
}
}
cost_function_config {
target_classes {
name: "customer"
class_weight: 1.0
coverage_foreground_weight: 0.05
objectives {
name: "cov"
initial_weight: 1.0
weight_target: 1.0
}
objectives {
name: "bbox"
initial_weight: 10.0
weight_target: 10.0
}
}
enable_autoweighting: True
max_objective_weight: 0.9999
min_objective_weight: 0.0001
}
I am unable to find where this KeyError is happening… please help!