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 ObjectDetector:
from autogluon.vision import ObjectDetector
/home/ci/opt/venv/lib/python3.8/site-packages/gluoncv/__init__.py:40: UserWarning: Both mxnet==1.9.1 and torch==1.12.1+cu102 are installed. You might encounter increased GPU memory footprint if both framework are used at the same time.
warnings.warn(f'Both mxnet=={mx.__version__} and torch=={torch.__version__} are installed. '
INFO:matplotlib.font_manager:generated new fontManager
INFO:torch.distributed.nn.jit.instantiator:Created a temporary directory at /tmp/tmp7gmo9777
INFO:torch.distributed.nn.jit.instantiator:Writing /tmp/tmp7gmo9777/_remote_module_non_scriptable.py
INFO:root:Generating grammar tables from /usr/lib/python3.8/lib2to3/Grammar.txt
INFO:root:Generating grammar tables from /usr/lib/python3.8/lib2to3/PatternGrammar.txt
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')
Downloading /home/ci/.gluoncv/archive/tiny_motorbike.zip from https://autogluon.s3.amazonaws.com/datasets/tiny_motorbike.zip...
21273KB [00:01, 19119.50KB/s]
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}
hyperparameter_tune_kwargs={'num_trials': 2}
detector.fit(dataset_train, time_limit=time_limit, hyperparameters=hyperparameters, hyperparameter_tune_kwargs=hyperparameter_tune_kwargs)
=============================================================================
WARNING: ObjectDetector is deprecated as of v0.4.0 and may contain various bugs and issues!
In a future release ObjectDetector may be entirely reworked to use Torch as a backend.
This future change will likely be API breaking.Users should ensure they update their code that depends on ObjectDetector when upgrading to future AutoGluon releases.
For more information, refer to ObjectDetector refactor GitHub issue: https://github.com/awslabs/autogluon/issues/1559
=============================================================================
The number of requested GPUs is greater than the number of available GPUs.Reduce the number to 1
Randomly split train_data into train[150]/validation[20] splits.
Starting HPO experiments
0%| | 0/2 [00:00<?, ?it/s]
INFO:SSDEstimator:modified configs(<old> != <new>): {
INFO:SSDEstimator:root.train.epochs 20 != 5
INFO:SSDEstimator:root.train.early_stop_baseline 0.0 != -inf
INFO:SSDEstimator:root.train.seed 233 != 304
INFO:SSDEstimator:root.train.early_stop_max_value 1.0 != inf
INFO:SSDEstimator:root.train.batch_size 16 != 8
INFO:SSDEstimator:root.train.early_stop_patience -1 != 10
INFO:SSDEstimator:root.num_workers 4 != 8
INFO:SSDEstimator:root.gpus (0, 1, 2, 3) != (0,)
INFO:SSDEstimator:root.valid.batch_size 16 != 8
INFO:SSDEstimator:root.ssd.base_network vgg16_atrous != resnet50_v1
INFO:SSDEstimator:root.ssd.data_shape 300 != 512
INFO:SSDEstimator:root.dataset voc_tiny != auto
INFO:SSDEstimator:root.dataset_root ~/.mxnet/datasets/ != auto
INFO:SSDEstimator:}
INFO:SSDEstimator:Saved config to /home/ci/autogluon/docs/_build/eval/tutorials/object_detection/4bc97620/.trial_0/config.yaml
INFO:SSDEstimator:Using transfer learning from ssd_512_resnet50_v1_coco, the other network parameters are ignored.
INFO:root:Model file not found. Downloading.
Downloading /home/ci/.mxnet/models/ssd_512_resnet50_v1_coco-c4835162.zip from https://apache-mxnet.s3-accelerate.dualstack.amazonaws.com/gluon/models/ssd_512_resnet50_v1_coco-c4835162.zip...
0%| | 0/181188 [00:00<?, ?KB/s][A
0%| | 101/181188 [00:00<03:45, 804.02KB/s][A
0%| | 520/181188 [00:00<01:18, 2302.73KB/s][A
1%| | 2162/181188 [00:00<00:24, 7195.29KB/s][A
4%|▍ | 7937/181188 [00:00<00:07, 24443.31KB/s][A
8%|▊ | 14345/181188 [00:00<00:04, 37430.43KB/s][A
13%|█▎ | 23337/181188 [00:00<00:02, 54165.92KB/s][A
17%|█▋ | 30610/181188 [00:00<00:02, 59968.05KB/s][A
21%|██ | 37483/181188 [00:00<00:02, 61382.81KB/s][A
25%|██▌ | 46022/181188 [00:00<00:01, 68626.25KB/s][A
29%|██▉ | 53020/181188 [00:01<00:01, 68692.47KB/s][A
34%|███▍ | 61940/181188 [00:01<00:01, 74764.90KB/s][A
38%|███▊ | 69493/181188 [00:01<00:01, 73668.59KB/s][A
43%|████▎ | 77433/181188 [00:01<00:01, 74786.68KB/s][A
47%|████▋ | 85530/181188 [00:01<00:01, 76612.96KB/s][A
51%|█████▏ | 93298/181188 [00:01<00:01, 76924.35KB/s][A
56%|█████▌ | 101605/181188 [00:01<00:01, 78753.11KB/s][A
60%|██████ | 109498/181188 [00:01<00:00, 77986.89KB/s][A
65%|██████▍ | 117759/181188 [00:01<00:00, 78305.67KB/s][A
69%|██████▉ | 125724/181188 [00:02<00:00, 78700.45KB/s][A
74%|███████▎ | 133601/181188 [00:02<00:00, 78357.70KB/s][A
78%|███████▊ | 141690/181188 [00:02<00:00, 79107.26KB/s][A
83%|████████▎ | 149606/181188 [00:02<00:00, 78308.23KB/s][A
87%|████████▋ | 157970/181188 [00:02<00:00, 79484.91KB/s][A
92%|█████████▏| 165922/181188 [00:02<00:00, 78613.17KB/s][A
181189KB [00:02, 66955.99KB/s]
INFO:SSDEstimator:Start training from [Epoch 0]
INFO:SSDEstimator:[Epoch 0] Training cost: 11.064638, CrossEntropy=3.616203, SmoothL1=1.069762
INFO:SSDEstimator:[Epoch 0] Validation:
chair=nan
boat=nan
motorbike=0.7653565977429614
pottedplant=nan
car=0.6464646464646466
bus=nan
bicycle=0.03636363636363636
dog=nan
cow=nan
person=0.7877959688618874
mAP=0.5589952123582829
INFO:SSDEstimator:[Epoch 0] Current best map: 0.558995 vs previous 0.000000, saved to /home/ci/autogluon/docs/_build/eval/tutorials/object_detection/4bc97620/.trial_0/best_checkpoint.pkl
INFO:SSDEstimator:[Epoch 1] Training cost: 7.569960, CrossEntropy=2.491966, SmoothL1=1.054675
INFO:SSDEstimator:[Epoch 1] Validation:
chair=nan
boat=nan
motorbike=0.8190910444107375
pottedplant=nan
car=0.6590909090909091
bus=nan
bicycle=0.0
dog=nan
cow=nan
person=0.6106134623755007
mAP=0.5221988539692869
INFO:SSDEstimator:[Epoch 2] Training cost: 7.487073, CrossEntropy=2.431485, SmoothL1=1.071399
INFO:SSDEstimator:[Epoch 2] Validation:
chair=nan
boat=nan
motorbike=0.8317643925374864
pottedplant=nan
car=1.0000000000000002
bus=nan
bicycle=0.0
dog=nan
cow=nan
person=0.784931734931735
mAP=0.6541740318673054
INFO:SSDEstimator:[Epoch 2] Current best map: 0.654174 vs previous 0.558995, saved to /home/ci/autogluon/docs/_build/eval/tutorials/object_detection/4bc97620/.trial_0/best_checkpoint.pkl
INFO:SSDEstimator:[Epoch 3] Training cost: 7.550369, CrossEntropy=2.279991, SmoothL1=0.978284
INFO:SSDEstimator:[Epoch 3] Validation:
chair=nan
boat=nan
motorbike=0.872616020343293
pottedplant=nan
car=1.0000000000000002
bus=nan
bicycle=0.0
dog=nan
cow=nan
person=0.861377792623018
mAP=0.6834984532415778
INFO:SSDEstimator:[Epoch 3] Current best map: 0.683498 vs previous 0.654174, saved to /home/ci/autogluon/docs/_build/eval/tutorials/object_detection/4bc97620/.trial_0/best_checkpoint.pkl
INFO:SSDEstimator:[Epoch 4] Training cost: 7.039415, CrossEntropy=2.212785, SmoothL1=0.984403
INFO:SSDEstimator:[Epoch 4] Validation:
chair=nan
boat=nan
motorbike=0.9053679653679654
pottedplant=nan
car=0.8484848484848483
bus=nan
bicycle=0.0
dog=nan
cow=nan
person=0.7991626429126427
mAP=0.6382538641913641
INFO:SSDEstimator:Applying the state from the best checkpoint...
INFO:root:Model file not found. Downloading.
Downloading /home/ci/.mxnet/models/resnet50_v1-cc729d95.zip from https://apache-mxnet.s3-accelerate.dualstack.amazonaws.com/gluon/models/resnet50_v1-cc729d95.zip...
0%| | 0/57421 [00:00<?, ?KB/s][A
0%| | 91/57421 [00:00<01:16, 745.51KB/s][A
1%| | 518/57421 [00:00<00:24, 2338.38KB/s][A
4%|▍ | 2161/57421 [00:00<00:07, 7483.77KB/s][A
11%|█▏ | 6485/57421 [00:00<00:02, 19993.43KB/s][A
21%|██ | 11781/57421 [00:00<00:01, 30974.93KB/s][A
36%|███▌ | 20721/57421 [00:00<00:00, 49815.17KB/s][A
49%|████▉ | 28201/57421 [00:00<00:00, 57694.22KB/s][A
61%|██████ | 34746/57421 [00:00<00:00, 60098.93KB/s][A
76%|███████▌ | 43446/57421 [00:00<00:00, 68362.76KB/s][A
100%|██████████| 57421/57421 [00:01<00:00, 50370.79KB/s]
Finished, total runtime is 76.80 s
{ 'best_config': { 'dataset': 'auto',
'dataset_root': 'auto',
'estimator': <class 'gluoncv.auto.estimators.ssd.ssd.SSDEstimator'>,
'gpus': [0],
'horovod': False,
'num_workers': 8,
'resume': '',
'save_interval': 1,
'ssd': { 'amp': False,
'base_network': 'resnet50_v1',
'data_shape': 512,
'filters': None,
'nms_thresh': 0.45,
'nms_topk': 400,
'ratios': ( [1, 2, 0.5],
[1, 2, 0.5, 3, 0.3333333333333333],
[1, 2, 0.5, 3, 0.3333333333333333],
[1, 2, 0.5, 3, 0.3333333333333333],
[1, 2, 0.5],
[1, 2, 0.5]),
'sizes': (30, 60, 111, 162, 213, 264, 315),
'steps': (8, 16, 32, 64, 100, 300),
'syncbn': False,
'transfer': 'ssd_512_resnet50_v1_coco'},
'train': { 'batch_size': 8,
'dali': False,
'early_stop_baseline': -inf,
'early_stop_max_value': inf,
'early_stop_min_delta': 0.001,
'early_stop_patience': 10,
'epochs': 5,
'log_interval': 100,
'lr': 0.001,
'lr_decay': 0.1,
'lr_decay_epoch': (160, 200),
'momentum': 0.9,
'seed': 304,
'start_epoch': 0,
'wd': 0.0005},
'valid': { 'batch_size': 8,
'iou_thresh': 0.5,
'metric': 'voc07',
'val_interval': 1}},
'total_time': 76.79892778396606,
'train_map': 0.7013412996382214,
'valid_map': 0.6834984532415778}
<autogluon.vision.detector.detector.ObjectDetector at 0x7f2cdb6da940>
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.
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: 0.12157768403857265
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)
predict_class predict_score 0 motorbike 0.964290
1 person 0.901450
2 motorbike 0.379375
3 car 0.224108
4 person 0.151031
.. ... ...
76 person 0.026454
77 person 0.026297
78 person 0.026240
79 chair 0.025993
80 person 0.025979
predict_rois
0 {'xmin': 0.32511788606643677, 'ymin': 0.426943...
1 {'xmin': 0.38163241744041443, 'ymin': 0.279039...
2 {'xmin': 0.0, 'ymin': 0.6350289583206177, 'xma...
3 {'xmin': 0.0, 'ymin': 0.6296865940093994, 'xma...
4 {'xmin': 0.03611136972904205, 'ymin': 0.0, 'xm...
.. ...
76 {'xmin': 0.8196716904640198, 'ymin': 0.4491611...
77 {'xmin': 0.40028253197669983, 'ymin': 0.757062...
78 {'xmin': 0.9661840200424194, 'ymin': 0.2806696...
79 {'xmin': 0.11712463200092316, 'ymin': 0.011974...
80 {'xmin': 0.993757426738739, 'ymin': 0.08150030...
[81 rows x 3 columns]
Prediction with multiple images is permitted:
bulk_result = detector.predict(dataset_test)
print(bulk_result)
predict_class predict_score 0 motorbike 0.964290
1 person 0.901450
2 motorbike 0.379375
3 car 0.224108
4 person 0.151031
... ... ...
3760 motorbike 0.017063
3761 car 0.017042
3762 person 0.016949
3763 person 0.016934
3764 motorbike 0.016925
predict_rois 0 {'xmin': 0.32511788606643677, 'ymin': 0.426943...
1 {'xmin': 0.38163241744041443, 'ymin': 0.279039...
2 {'xmin': 0.0, 'ymin': 0.6350289583206177, 'xma...
3 {'xmin': 0.0, 'ymin': 0.6296865940093994, 'xma...
4 {'xmin': 0.03611136972904205, 'ymin': 0.0, 'xm...
... ...
3760 {'xmin': 0.11219224333763123, 'ymin': 0.560805...
3761 {'xmin': 0.8976275324821472, 'ymin': 0.7462039...
3762 {'xmin': 0.3027859032154083, 'ymin': 0.4321423...
3763 {'xmin': 0.7102004289627075, 'ymin': 0.2931949...
3764 {'xmin': 0.7111496925354004, 'ymin': 0.8699753...
image
0 /home/ci/.gluoncv/datasets/tiny_motorbike/tiny...
1 /home/ci/.gluoncv/datasets/tiny_motorbike/tiny...
2 /home/ci/.gluoncv/datasets/tiny_motorbike/tiny...
3 /home/ci/.gluoncv/datasets/tiny_motorbike/tiny...
4 /home/ci/.gluoncv/datasets/tiny_motorbike/tiny...
... ...
3760 /home/ci/.gluoncv/datasets/tiny_motorbike/tiny...
3761 /home/ci/.gluoncv/datasets/tiny_motorbike/tiny...
3762 /home/ci/.gluoncv/datasets/tiny_motorbike/tiny...
3763 /home/ci/.gluoncv/datasets/tiny_motorbike/tiny...
3764 /home/ci/.gluoncv/datasets/tiny_motorbike/tiny...
[3765 rows x 4 columns]
We can also save the trained model, and use it later.
Warning
ObjectDetector.load() used pickle module implicitly, which is
known to be insecure. It is possible to construct malicious pickle
data which will execute arbitrary code during unpickling. Never load
data that could have come from an untrusted source, or that could
have been tampered with. Only load data you trust.
savefile = 'detector.ag'
detector.save(savefile)
new_detector = ObjectDetector.load(savefile)
/home/ci/opt/venv/lib/python3.8/site-packages/mxnet/gluon/block.py:1784: UserWarning: Cannot decide type for the following arguments. Consider providing them as input:
data: None
input_sym_arg_type = in_param.infer_type()[0]