Object Detection - Quick Start¶
Note: AutoGluon ObjectDetector will be deprecated in v0.7. Please try our AutoGluon MultiModalPredictor for more functionalities and better support for your object detection need.
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/tmp21vc7_5d
INFO:torch.distributed.nn.jit.instantiator:Writing /tmp/tmp21vc7_5d/_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, 17169.42KB/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/autogluon/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[154]/validation[16] splits.
Starting HPO experiments
0%| | 0/2 [00:00<?, ?it/s]
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 voc_tiny != auto
INFO:SSDEstimator:root.dataset_root ~/.mxnet/datasets/ != auto
INFO:SSDEstimator:root.gpus (0, 1, 2, 3) != (0,)
INFO:SSDEstimator:root.ssd.base_network vgg16_atrous != resnet50_v1
INFO:SSDEstimator:root.ssd.data_shape 300 != 512
INFO:SSDEstimator:root.train.early_stop_baseline 0.0 != -inf
INFO:SSDEstimator:root.train.epochs 20 != 5
INFO:SSDEstimator:root.train.seed 233 != 240
INFO:SSDEstimator:root.train.batch_size 16 != 8
INFO:SSDEstimator:root.train.early_stop_patience -1 != 10
INFO:SSDEstimator:root.train.early_stop_max_value 1.0 != inf
INFO:SSDEstimator:}
INFO:SSDEstimator:Saved config to /home/ci/autogluon/docs/_build/eval/tutorials/object_detection/34847973/.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%| | 93/181188 [00:00<04:06, 735.90KB/s][A
0%| | 518/181188 [00:00<01:21, 2226.78KB/s][A
1%| | 2174/181188 [00:00<00:25, 7099.11KB/s][A
4%|▍ | 7884/181188 [00:00<00:07, 23927.15KB/s][A
8%|▊ | 13687/181188 [00:00<00:04, 34999.21KB/s][A
12%|█▏ | 22157/181188 [00:00<00:03, 50746.09KB/s][A
16%|█▌ | 28977/181188 [00:00<00:02, 56183.85KB/s][A
21%|██ | 37500/181188 [00:00<00:02, 62173.87KB/s][A
25%|██▌ | 46073/181188 [00:01<00:01, 69139.82KB/s][A
29%|██▉ | 53115/181188 [00:01<00:01, 69283.02KB/s][A
34%|███▍ | 62034/181188 [00:01<00:01, 75197.68KB/s][A
38%|███▊ | 69633/181188 [00:01<00:01, 74227.63KB/s][A
43%|████▎ | 77545/181188 [00:01<00:01, 73948.74KB/s][A
47%|████▋ | 86011/181188 [00:01<00:01, 77068.52KB/s][A
52%|█████▏ | 93757/181188 [00:01<00:01, 75282.36KB/s][A
57%|█████▋ | 102609/181188 [00:01<00:01, 78448.83KB/s][A
61%|██████ | 110482/181188 [00:01<00:00, 77121.88KB/s][A
65%|██████▌ | 118215/181188 [00:01<00:00, 76153.47KB/s][A
70%|██████▉ | 126622/181188 [00:02<00:00, 78449.86KB/s][A
74%|███████▍ | 134485/181188 [00:02<00:00, 76899.15KB/s][A
79%|███████▉ | 142767/181188 [00:02<00:00, 78620.62KB/s][A
83%|████████▎ | 150646/181188 [00:02<00:00, 76708.55KB/s][A
88%|████████▊ | 158621/181188 [00:02<00:00, 77542.94KB/s][A
92%|█████████▏| 166391/181188 [00:02<00:00, 76310.85KB/s][A
181189KB [00:02, 65796.69KB/s]
INFO:SSDEstimator:Start training from [Epoch 0]
INFO:SSDEstimator:[Epoch 0] Training cost: 11.337567, CrossEntropy=3.617623, SmoothL1=0.974323
INFO:SSDEstimator:[Epoch 0] Validation:
bicycle=nan
motorbike=0.8395863395863397
cow=nan
person=0.6386015880656871
bus=1.0000000000000002
pottedplant=0.0
boat=nan
chair=nan
dog=nan
car=1.0000000000000002
mAP=0.6956375855304054
INFO:SSDEstimator:[Epoch 0] Current best map: 0.695638 vs previous 0.000000, saved to /home/ci/autogluon/docs/_build/eval/tutorials/object_detection/34847973/.trial_0/best_checkpoint.pkl
INFO:SSDEstimator:[Epoch 1] Training cost: 7.683115, CrossEntropy=2.668093, SmoothL1=1.196020
INFO:SSDEstimator:[Epoch 1] Validation:
bicycle=nan
motorbike=0.7963052989790959
cow=nan
person=0.6720987230852619
bus=1.0000000000000002
pottedplant=0.0
boat=nan
chair=nan
dog=nan
car=1.0000000000000002
mAP=0.6936808044128716
INFO:SSDEstimator:[Epoch 2] Training cost: 7.572500, CrossEntropy=2.433229, SmoothL1=1.208440
INFO:SSDEstimator:[Epoch 2] Validation:
bicycle=nan
motorbike=0.8468805704099821
cow=nan
person=0.6317966903073285
bus=1.0000000000000002
pottedplant=0.0
boat=nan
chair=nan
dog=nan
car=0.16666666666666663
mAP=0.5290687854767955
INFO:SSDEstimator:[Epoch 3] Training cost: 8.068636, CrossEntropy=2.492503, SmoothL1=1.155709
INFO:SSDEstimator:[Epoch 3] Validation:
bicycle=nan
motorbike=0.8552139037433155
cow=nan
person=0.7063568010936432
bus=1.0000000000000002
pottedplant=0.0
boat=nan
chair=nan
dog=nan
car=0.0
mAP=0.5123141409673918
INFO:SSDEstimator:[Epoch 4] Training cost: 7.803893, CrossEntropy=2.241806, SmoothL1=0.968488
INFO:SSDEstimator:[Epoch 4] Validation:
bicycle=nan
motorbike=0.8622009569377991
cow=nan
person=0.7100792615146682
bus=1.0000000000000002
pottedplant=0.0
boat=nan
chair=nan
dog=nan
car=1.0000000000000002
mAP=0.7144560436904935
INFO:SSDEstimator:[Epoch 4] Current best map: 0.714456 vs previous 0.695638, saved to /home/ci/autogluon/docs/_build/eval/tutorials/object_detection/34847973/.trial_0/best_checkpoint.pkl
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%| | 101/57421 [00:00<01:11, 803.05KB/s][A
1%| | 510/57421 [00:00<00:25, 2211.67KB/s][A
4%|▍ | 2176/57421 [00:00<00:07, 7118.13KB/s][A
14%|█▎ | 7797/57421 [00:00<00:02, 23653.24KB/s][A
23%|██▎ | 13452/57421 [00:00<00:01, 34339.60KB/s][A
37%|███▋ | 21451/57421 [00:00<00:00, 45410.88KB/s][A
52%|█████▏ | 30075/57421 [00:00<00:00, 57512.05KB/s][A
64%|██████▍ | 36643/57421 [00:00<00:00, 59940.45KB/s][A
79%|███████▉ | 45232/57421 [00:01<00:00, 67677.16KB/s][A
100%|██████████| 57421/57421 [00:01<00:00, 48761.81KB/s]
Finished, total runtime is 76.11 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': 240,
'start_epoch': 0,
'wd': 0.0005},
'valid': { 'batch_size': 8,
'iou_thresh': 0.5,
'metric': 'voc07',
'val_interval': 1}},
'total_time': 76.10566902160645,
'train_map': 0.693483757046729,
'valid_map': 0.7144560436904935}
<autogluon.vision.detector.detector.ObjectDetector at 0x7f55202cc0d0>
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: 0.10378962435414049
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 0.994273
1 person 0.990845
2 motorbike 0.272468
3 motorbike 0.220999
4 bicycle 0.114779
.. ... ...
91 person 0.019631
92 person 0.019560
93 person 0.019326
94 person 0.019282
95 person 0.019236
predict_rois
0 {'xmin': 0.31301018595695496, 'ymin': 0.440578...
1 {'xmin': 0.39104267954826355, 'ymin': 0.270204...
2 {'xmin': 0.37182092666625977, 'ymin': 0.318126...
3 {'xmin': 0.0, 'ymin': 0.6348890066146851, 'xma...
4 {'xmin': 0.3112749755382538, 'ymin': 0.4515416...
.. ...
91 {'xmin': 0.05895491689443588, 'ymin': 0.018557...
92 {'xmin': 0.5334900617599487, 'ymin': 0.1607223...
93 {'xmin': 0.040014903992414474, 'ymin': 0.0, 'x...
94 {'xmin': 0.9102627635002136, 'ymin': 0.0030383...
95 {'xmin': 0.39613157510757446, 'ymin': 0.273160...
[96 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 0.994273
1 person 0.990845
2 motorbike 0.272468
3 motorbike 0.220999
4 bicycle 0.114779
... ... ...
4361 person 0.024184
4362 person 0.024164
4363 bicycle 0.024148
4364 motorbike 0.023968
4365 person 0.023958
predict_rois 0 {'xmin': 0.31301018595695496, 'ymin': 0.440578...
1 {'xmin': 0.39104267954826355, 'ymin': 0.270204...
2 {'xmin': 0.37182092666625977, 'ymin': 0.318126...
3 {'xmin': 0.0, 'ymin': 0.6348890066146851, 'xma...
4 {'xmin': 0.3112749755382538, 'ymin': 0.4515416...
... ...
4361 {'xmin': 0.3682926893234253, 'ymin': 0.3581345...
4362 {'xmin': 0.9676966071128845, 'ymin': 0.6243187...
4363 {'xmin': 0.8265968561172485, 'ymin': 0.8293898...
4364 {'xmin': 0.15024471282958984, 'ymin': 0.608118...
4365 {'xmin': 0.16833484172821045, 'ymin': 0.428479...
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...
... ...
4361 /home/ci/.gluoncv/datasets/tiny_motorbike/tiny...
4362 /home/ci/.gluoncv/datasets/tiny_motorbike/tiny...
4363 /home/ci/.gluoncv/datasets/tiny_motorbike/tiny...
4364 /home/ci/.gluoncv/datasets/tiny_motorbike/tiny...
4365 /home/ci/.gluoncv/datasets/tiny_motorbike/tiny...
[4366 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]