Object Detection - Quick Start¶
Object detection is the process of identifying and localizing objects in an image and is an important task in computer vision. Follow this tutorial to learn how to use AutoGluon for object detection.
Tip: If you are new to AutoGluon, review Image Prediction - Quick Start first to learn the basics of the AutoGluon API.
Our goal is to detect motorbike in images by YOLOv3 model. A tiny dataset is collected from VOC dataset, which only contains the motorbike category. The model pretrained on the COCO dataset is used to fine-tune our small dataset. With the help of AutoGluon, we are able to try many models with different hyperparameters automatically, and return the best one as our final model.
To start, import autogluon.vision and ObjectDetector:
import autogluon.core as ag
from autogluon.vision import ObjectDetector
/var/lib/jenkins/workspace/workspace/autogluon-tutorial-object-detection-v3/core/src/autogluon/core/scheduler/jobs.py:132: SyntaxWarning: "is" with a literal. Did you mean "=="?
file = sys.stderr if out is 'err' else sys.stdout
Tiny_motorbike Dataset¶
We collect a toy dataset for detecting motorbikes in images. From the VOC dataset, images are randomly selected for training, validation, and testing - 120 images for training, 50 images for validation, and 50 for testing. This tiny dataset follows the same format as VOC.
Using the commands below, we can download this dataset, which is only
23M. The name of unzipped folder is called tiny_motorbike. Anyway,
the task dataset helper can perform the download and extraction
automatically, and load the dataset according to the detection formats.
url = 'https://autogluon.s3.amazonaws.com/datasets/tiny_motorbike.zip'
dataset_train = ObjectDetector.Dataset.from_voc(url, splits='trainval')
tiny_motorbike/
├── Annotations/
├── ImageSets/
└── JPEGImages/
Fit Models by AutoGluon¶
In this section, we demonstrate how to apply AutoGluon to fit our detection models. We use mobilenet as the backbone for the YOLOv3 model. Two different learning rates are used to fine-tune the network. The best model is the one that obtains the best performance on the validation dataset. You can also try using more networks and hyperparameters to create a larger searching space.
We fit a classifier using AutoGluon as follows. In each experiment
(one trial in our searching space), we train the model for 5 epochs to
avoid bursting our tutorial runtime.
time_limit = 60*30 # at most 0.5 hour
detector = ObjectDetector()
hyperparameters = {'epochs': 5, 'batch_size': 8}
hyperparamter_tune_kwargs={'num_trials': 2}
detector.fit(dataset_train, time_limit=time_limit, hyperparameters=hyperparameters, hyperparamter_tune_kwargs=hyperparamter_tune_kwargs)
WARNING:gluoncv.auto.tasks.object_detection:The number of requested GPUs is greater than the number of available GPUs.Reduce the number to 1
INFO:gluoncv.auto.tasks.object_detection:Randomly split train_data into train[153]/validation[17] splits.
INFO:gluoncv.auto.tasks.object_detection:Starting fit without HPO
INFO:SSDEstimator:modified configs(<old> != <new>): {
INFO:SSDEstimator:root.valid.batch_size 16 != 8
INFO:SSDEstimator:root.num_workers 4 != 8
INFO:SSDEstimator:root.dataset_root ~/.mxnet/datasets/ != auto
INFO:SSDEstimator:root.train.seed 233 != 518
INFO:SSDEstimator:root.train.batch_size 16 != 8
INFO:SSDEstimator:root.train.epochs 20 != 5
INFO:SSDEstimator:root.dataset voc_tiny != auto
INFO:SSDEstimator:root.gpus (0, 1, 2, 3) != (0,)
INFO:SSDEstimator:root.ssd.data_shape 300 != 512
INFO:SSDEstimator:root.ssd.base_network vgg16_atrous != resnet50_v1
INFO:SSDEstimator:}
INFO:SSDEstimator:Saved config to /var/lib/jenkins/workspace/workspace/autogluon-tutorial-object-detection-v3/docs/_build/eval/tutorials/object_detection/05c48225/.trial_0/config.yaml
INFO:SSDEstimator:Using transfer learning from ssd_512_resnet50_v1_coco, the other network parameters are ignored.
INFO:SSDEstimator:Start training from [Epoch 0]
INFO:SSDEstimator:[Epoch 0] Training cost: 10.095721, CrossEntropy=3.505902, SmoothL1=1.049549
INFO:SSDEstimator:[Epoch 0] Validation:
dog=1.0000000000000002
chair=0.0
cow=0.909090909090909
bicycle=0.06436337901306054
bus=0.31818181818181823
person=0.7422855054197233
motorbike=0.6067141109174189
pottedplant=0.0
boat=1.0000000000000002
car=0.7809041688253078
mAP=0.5421539891448239
INFO:SSDEstimator:[Epoch 0] Current best map: 0.542154 vs previous 0.000000, saved to /var/lib/jenkins/workspace/workspace/autogluon-tutorial-object-detection-v3/docs/_build/eval/tutorials/object_detection/05c48225/.trial_0/best_checkpoint.pkl
INFO:SSDEstimator:Pickled to /var/lib/jenkins/workspace/workspace/autogluon-tutorial-object-detection-v3/docs/_build/eval/tutorials/object_detection/05c48225/.trial_0/best_checkpoint.pkl
INFO:SSDEstimator:[Epoch 1] Training cost: 9.202750, CrossEntropy=2.617644, SmoothL1=1.162525
INFO:SSDEstimator:[Epoch 1] Validation:
dog=1.0000000000000002
chair=0.10000000000000002
cow=1.0000000000000002
bicycle=0.36366578819435563
bus=0.6363636363636365
person=0.7353234805617082
motorbike=0.8644176832704251
pottedplant=0.0
boat=1.0000000000000002
car=0.6797905525846702
mAP=0.6379561140974797
INFO:SSDEstimator:[Epoch 1] Current best map: 0.637956 vs previous 0.542154, saved to /var/lib/jenkins/workspace/workspace/autogluon-tutorial-object-detection-v3/docs/_build/eval/tutorials/object_detection/05c48225/.trial_0/best_checkpoint.pkl
INFO:SSDEstimator:Pickled to /var/lib/jenkins/workspace/workspace/autogluon-tutorial-object-detection-v3/docs/_build/eval/tutorials/object_detection/05c48225/.trial_0/best_checkpoint.pkl
INFO:SSDEstimator:[Epoch 2] Training cost: 8.853812, CrossEntropy=2.517284, SmoothL1=1.173569
INFO:SSDEstimator:[Epoch 2] Validation:
dog=1.0000000000000002
chair=1.0000000000000002
cow=0.5587583148558757
bicycle=0.44866373296125367
bus=0.6472727272727272
person=0.7827678247643114
motorbike=0.8525577209333013
pottedplant=0.017595307917888565
boat=0.009433962264150943
car=0.8111666954171699
mAP=0.6128216286386678
INFO:SSDEstimator:[Epoch 3] Training cost: 9.413587, CrossEntropy=2.267903, SmoothL1=1.077149
INFO:SSDEstimator:[Epoch 3] Validation:
dog=1.0000000000000002
chair=1.0000000000000002
cow=0.909090909090909
bicycle=0.5454545454545455
bus=0.5454545454545454
person=0.7529320695779278
motorbike=0.8822863768806136
pottedplant=0.0
boat=1.0000000000000002
car=0.7772610464571088
mAP=0.741247949291565
INFO:SSDEstimator:[Epoch 3] Current best map: 0.741248 vs previous 0.637956, saved to /var/lib/jenkins/workspace/workspace/autogluon-tutorial-object-detection-v3/docs/_build/eval/tutorials/object_detection/05c48225/.trial_0/best_checkpoint.pkl
INFO:SSDEstimator:Pickled to /var/lib/jenkins/workspace/workspace/autogluon-tutorial-object-detection-v3/docs/_build/eval/tutorials/object_detection/05c48225/.trial_0/best_checkpoint.pkl
INFO:SSDEstimator:[Epoch 4] Training cost: 9.059360, CrossEntropy=2.242657, SmoothL1=0.959251
INFO:SSDEstimator:[Epoch 4] Validation:
dog=1.0000000000000002
chair=1.0000000000000002
cow=1.0000000000000002
bicycle=0.7030303030303031
bus=1.0000000000000002
person=0.8510013223518652
motorbike=0.8882810877374476
pottedplant=0.003896103896103896
boat=1.0000000000000002
car=0.8521074013092383
mAP=0.8298316218324959
INFO:SSDEstimator:[Epoch 4] Current best map: 0.829832 vs previous 0.741248, saved to /var/lib/jenkins/workspace/workspace/autogluon-tutorial-object-detection-v3/docs/_build/eval/tutorials/object_detection/05c48225/.trial_0/best_checkpoint.pkl
INFO:SSDEstimator:Pickled to /var/lib/jenkins/workspace/workspace/autogluon-tutorial-object-detection-v3/docs/_build/eval/tutorials/object_detection/05c48225/.trial_0/best_checkpoint.pkl
INFO:gluoncv.auto.tasks.object_detection:Finished, total runtime is 82.77 s
INFO:gluoncv.auto.tasks.object_detection:{ 'best_config': { 'batch_size': 8,
'dist_ip_addrs': None,
'epochs': 5,
'final_fit': False,
'gpus': [0],
'log_dir': '/var/lib/jenkins/workspace/workspace/autogluon-tutorial-object-detection-v3/docs/_build/eval/tutorials/object_detection/05c48225',
'lr': 0.001,
'ngpus_per_trial': 8,
'nthreads_per_trial': 128,
'num_trials': 1,
'num_workers': 8,
'search_strategy': 'random',
'seed': 518,
'time_limits': 1800,
'transfer': 'ssd_512_resnet50_v1_coco',
'wall_clock_tick': 1614127622.607981},
'total_time': 69.39395928382874,
'train_map': 0.8298316218324959,
'valid_map': 0.8298316218324959}
<autogluon.vision.detector.detector.ObjectDetector at 0x7f69d72eceb0>
Note that num_trials=2 above is only used to speed up the tutorial.
In normal practice, it is common to only use time_limit and drop
num_trials. Also note that hyperparameter tuning defaults to random
search. Model-based variants, such as search_strategy='bayesopt' or
search_strategy='bayesopt_hyperband' can be a lot more
sample-efficient.
After fitting, AutoGluon automatically returns the best model among all models in the searching space. From the output, we know the best model is the one trained with the second learning rate. To see how well the returned model performed on test dataset, call detector.evaluate().
dataset_test = ObjectDetector.Dataset.from_voc(url, splits='test')
test_map = detector.evaluate(dataset_test)
print("mAP on test dataset: {}".format(test_map[1][-1]))
tiny_motorbike/
├── Annotations/
├── ImageSets/
└── JPEGImages/
mAP on test dataset: 0.05792196933501283
Below, we randomly select an image from test dataset and show the
predicted class, box and probability over the origin image, stored in
predict_class, predict_rois and predict_score columns,
respectively. You can interpret predict_rois as a dict of (xmin,
ymin, xmax, ymax) proportional to original image size.
image_path = dataset_test.iloc[0]['image']
result = detector.predict(image_path)
print(result)
INFO:numexpr.utils:NumExpr defaulting to 8 threads.
predict_class predict_score 0 person 0.977397
1 motorbike 0.961676
2 car 0.557639
3 motorbike 0.483465
4 person 0.169022
.. ... ...
87 chair 0.029470
88 car 0.029415
89 person 0.029415
90 person 0.029240
91 person 0.029232
predict_rois
0 {'xmin': 0.3849811255931854, 'ymin': 0.3002845...
1 {'xmin': 0.3170837163925171, 'ymin': 0.4358768...
2 {'xmin': 0.004333138465881348, 'ymin': 0.63187...
3 {'xmin': 0.3680243492126465, 'ymin': 0.3390576...
4 {'xmin': 0.0605580173432827, 'ymin': 0.0173151...
.. ...
87 {'xmin': 0.8222059011459351, 'ymin': 0.3000083...
88 {'xmin': 0.00212657917290926, 'ymin': 0.459672...
89 {'xmin': 0.05945102870464325, 'ymin': 0.075048...
90 {'xmin': 0.5079175233840942, 'ymin': 0.2999778...
91 {'xmin': 0.6806942224502563, 'ymin': 0.2688137...
[92 rows x 3 columns]
Prediction with multiple images is permitted:
bulk_result = detector.predict(dataset_test)
print(bulk_result)
predict_class predict_score 0 person 0.977397
1 motorbike 0.961676
2 car 0.557639
3 motorbike 0.483465
4 person 0.169022
... ... ...
4046 person 0.023806
4047 motorbike 0.023552
4048 person 0.023442
4049 person 0.023383
4050 motorbike 0.023344
predict_rois 0 {'xmin': 0.3849811255931854, 'ymin': 0.3002845...
1 {'xmin': 0.3170837163925171, 'ymin': 0.4358768...
2 {'xmin': 0.004333138465881348, 'ymin': 0.63187...
3 {'xmin': 0.3680243492126465, 'ymin': 0.3390576...
4 {'xmin': 0.0605580173432827, 'ymin': 0.0173151...
... ...
4046 {'xmin': 0.543327808380127, 'ymin': 0.61410915...
4047 {'xmin': 0.4615864157676697, 'ymin': 0.2539878...
4048 {'xmin': 0.49686524271965027, 'ymin': 0.190620...
4049 {'xmin': 0.3686455488204956, 'ymin': 0.1449364...
4050 {'xmin': 0.8133487701416016, 'ymin': 0.8839820...
image
0 /var/lib/jenkins/.gluoncv/datasets/tiny_motorb...
1 /var/lib/jenkins/.gluoncv/datasets/tiny_motorb...
2 /var/lib/jenkins/.gluoncv/datasets/tiny_motorb...
3 /var/lib/jenkins/.gluoncv/datasets/tiny_motorb...
4 /var/lib/jenkins/.gluoncv/datasets/tiny_motorb...
... ...
4046 /var/lib/jenkins/.gluoncv/datasets/tiny_motorb...
4047 /var/lib/jenkins/.gluoncv/datasets/tiny_motorb...
4048 /var/lib/jenkins/.gluoncv/datasets/tiny_motorb...
4049 /var/lib/jenkins/.gluoncv/datasets/tiny_motorb...
4050 /var/lib/jenkins/.gluoncv/datasets/tiny_motorb...
[4051 rows x 4 columns]
We can also save the trained model, and use it later.
savefile = 'detector.ag'
detector.save(savefile)
new_detector = ObjectDetector.load(savefile)
/var/lib/jenkins/workspace/workspace/autogluon-tutorial-object-detection-v3/venv/lib/python3.8/site-packages/mxnet/gluon/block.py:1512: UserWarning: Cannot decide type for the following arguments. Consider providing them as input:
data: None
input_sym_arg_type = in_param.infer_type()[0]