Image Prediction - Quick Start¶
In this quick start, we’ll use the task of image classification to illustrate how to use AutoGluon’s APIs. This tutorial demonstrates how to load images and corresponding labels into AutoGluon and use this data to obtain a neural network that can classify new images. This is different from traditional machine learning where we need to manually define the neural network and then specify the hyperparameters in the training process. Instead, with just a single call to AutoGluon’s fit function, AutoGluon automatically trains many models with different hyperparameter configurations and returns the model that achieved the highest level of accuracy.
import autogluon.core as ag
from autogluon.vision import ImagePredictor
Create Image Dataset¶
For demonstration purposes, we use a subset of the Shopee-IET
dataset
from Kaggle. Each image in this data depicts a clothing item and the
corresponding label specifies its clothing category. Our subset of the
data contains the following possible labels: BabyPants
,
BabyShirt
, womencasualshoes
, womenchiffontop
.
We can load a dataset by downloading a url data automatically:
train_dataset, _, test_dataset = ImagePredictor.Dataset.from_folders('https://autogluon.s3.amazonaws.com/datasets/shopee-iet.zip')
print(train_dataset)
data/
├── test/
└── train/
image label
0 /var/lib/jenkins/.gluoncv/datasets/shopee-iet/... 0
1 /var/lib/jenkins/.gluoncv/datasets/shopee-iet/... 0
2 /var/lib/jenkins/.gluoncv/datasets/shopee-iet/... 0
3 /var/lib/jenkins/.gluoncv/datasets/shopee-iet/... 0
4 /var/lib/jenkins/.gluoncv/datasets/shopee-iet/... 0
.. ... ...
795 /var/lib/jenkins/.gluoncv/datasets/shopee-iet/... 3
796 /var/lib/jenkins/.gluoncv/datasets/shopee-iet/... 3
797 /var/lib/jenkins/.gluoncv/datasets/shopee-iet/... 3
798 /var/lib/jenkins/.gluoncv/datasets/shopee-iet/... 3
799 /var/lib/jenkins/.gluoncv/datasets/shopee-iet/... 3
[800 rows x 2 columns]
Use AutoGluon to Fit Models¶
Now, we fit a classifier using AutoGluon as follows:
predictor = ImagePredictor()
# since the original dataset does not provide validation split, the `fit` function splits it randomly with 90/10 ratio
predictor.fit(train_dataset, hyperparameters={'epochs': 2}) # you can trust the default config, we reduce the # epoch to save some build time
INFO:root:time_limit=auto set to time_limit=7200. WARNING:gluoncv.auto.tasks.image_classification:The number of requested GPUs is greater than the number of available GPUs.Reduce the number to 1 INFO:gluoncv.auto.tasks.image_classification:Randomly split train_data into train[719]/validation[81] splits. INFO:gluoncv.auto.tasks.image_classification:Starting fit without HPO INFO:ImageClassificationEstimator:modified configs(<old> != <new>): { INFO:ImageClassificationEstimator:root.valid.batch_size 128 != 16 INFO:ImageClassificationEstimator:root.valid.num_workers 4 != 8 INFO:ImageClassificationEstimator:root.img_cls.model resnet50_v1 != resnet50_v1b INFO:ImageClassificationEstimator:root.train.epochs 10 != 2 INFO:ImageClassificationEstimator:root.train.rec_val_idx ~/.mxnet/datasets/imagenet/rec/val.idx != auto INFO:ImageClassificationEstimator:root.train.rec_train ~/.mxnet/datasets/imagenet/rec/train.rec != auto INFO:ImageClassificationEstimator:root.train.lr 0.1 != 0.01 INFO:ImageClassificationEstimator:root.train.rec_train_idx ~/.mxnet/datasets/imagenet/rec/train.idx != auto INFO:ImageClassificationEstimator:root.train.rec_val ~/.mxnet/datasets/imagenet/rec/val.rec != auto INFO:ImageClassificationEstimator:root.train.num_workers 4 != 8 INFO:ImageClassificationEstimator:root.train.num_training_samples 1281167 != -1 INFO:ImageClassificationEstimator:root.train.data_dir ~/.mxnet/datasets/imagenet != auto INFO:ImageClassificationEstimator:root.train.batch_size 128 != 16 INFO:ImageClassificationEstimator:} INFO:ImageClassificationEstimator:Saved config to /var/lib/jenkins/workspace/workspace/autogluon-tutorial-image-classification-v3/docs/_build/eval/tutorials/image_prediction/2457e1d5/.trial_0/config.yaml INFO:ImageClassificationEstimator:Start training from [Epoch 0] INFO:ImageClassificationEstimator:Epoch[0] Batch [49] Speed: 95.150201 samples/sec accuracy=0.503750 lr=0.010000 INFO:ImageClassificationEstimator:[Epoch 0] training: accuracy=0.503750 INFO:ImageClassificationEstimator:[Epoch 0] speed: 93 samples/sec time cost: 11.436843 INFO:ImageClassificationEstimator:[Epoch 0] validation: top1=0.745000 top5=1.000000 INFO:ImageClassificationEstimator:[Epoch 0] Current best top-1: 0.745000 vs previous 0.000000, saved to /var/lib/jenkins/workspace/workspace/autogluon-tutorial-image-classification-v3/docs/_build/eval/tutorials/image_prediction/2457e1d5/.trial_0/best_checkpoint.pkl INFO:ImageClassificationEstimator:Pickled to /var/lib/jenkins/workspace/workspace/autogluon-tutorial-image-classification-v3/docs/_build/eval/tutorials/image_prediction/2457e1d5/.trial_0/best_checkpoint.pkl INFO:ImageClassificationEstimator:Epoch[1] Batch [49] Speed: 96.998034 samples/sec accuracy=0.693750 lr=0.010000 INFO:ImageClassificationEstimator:[Epoch 1] training: accuracy=0.693750 INFO:ImageClassificationEstimator:[Epoch 1] speed: 95 samples/sec time cost: 11.234721 INFO:ImageClassificationEstimator:[Epoch 1] validation: top1=0.846250 top5=1.000000 INFO:ImageClassificationEstimator:[Epoch 1] Current best top-1: 0.846250 vs previous 0.745000, saved to /var/lib/jenkins/workspace/workspace/autogluon-tutorial-image-classification-v3/docs/_build/eval/tutorials/image_prediction/2457e1d5/.trial_0/best_checkpoint.pkl INFO:ImageClassificationEstimator:Pickled to /var/lib/jenkins/workspace/workspace/autogluon-tutorial-image-classification-v3/docs/_build/eval/tutorials/image_prediction/2457e1d5/.trial_0/best_checkpoint.pkl INFO:gluoncv.auto.tasks.image_classification:Finished, total runtime is 33.95 s INFO:gluoncv.auto.tasks.image_classification:{ 'best_config': { 'batch_size': 16, 'dist_ip_addrs': None, 'epochs': 2, 'estimator': <class 'gluoncv.auto.estimators.image_classification.image_classification.ImageClassificationEstimator'>, 'final_fit': False, 'gpus': [0], 'log_dir': '/var/lib/jenkins/workspace/workspace/autogluon-tutorial-image-classification-v3/docs/_build/eval/tutorials/image_prediction/2457e1d5', 'lr': 0.01, 'model': 'resnet50_v1b', 'ngpus_per_trial': 8, 'nthreads_per_trial': 128, 'num_trials': 1, 'num_workers': 8, 'search_strategy': 'random', 'seed': 332, 'time_limits': 7200, 'wall_clock_tick': 1614636384.3292763}, 'total_time': 24.781545400619507, 'train_acc': 0.69375, 'valid_acc': 0.84625}
<autogluon.vision.predictor.predictor.ImagePredictor at 0x7f80e7964390>
Within fit
, the dataset is automatically split into training and
validation sets. The model with the best hyperparameter configuration is
selected based on its performance on the validation set. The best model
is finally retrained on our entire dataset (i.e., merging
training+validation) using the best configuration.
The best Top-1 accuracy achieved on the validation set is as follows:
fit_result = predictor.fit_summary()
print('Top-1 train acc: %.3f, val acc: %.3f' %(fit_result['train_acc'], fit_result['valid_acc']))
Top-1 train acc: 0.694, val acc: 0.846
Predict on a New Image¶
Given an example image, we can easily use the final model to predict
the label (and the conditional class-probability denoted as score
):
image_path = test_dataset.iloc[0]['image']
result = predictor.predict(image_path)
print(result)
class score id
0 BabyShirt 0.783843 1
If probabilities of all categories are needed, you can call
predict_proba
:
proba = predictor.predict_proba(image_path)
print(proba)
class score id
0 BabyShirt 0.783843 1
1 BabyPants 0.132135 0
2 womencasualshoes 0.043968 2
3 womenchiffontop 0.040053 3
You can also feed in multiple images all together, let’s use images in test dataset as an example:
bulk_result = predictor.predict(test_dataset)
print(bulk_result)
class score id 0 BabyShirt 0.783843 1 1 BabyShirt 0.946820 1 2 BabyShirt 0.450992 1 3 BabyShirt 0.445897 1 4 BabyShirt 0.711610 1 .. ... ... .. 75 womenchiffontop 0.740012 3 76 BabyPants 0.357589 0 77 womenchiffontop 0.757476 3 78 womenchiffontop 0.401652 3 79 womenchiffontop 0.652606 3 image 0 /var/lib/jenkins/.gluoncv/datasets/shopee-iet/... 1 /var/lib/jenkins/.gluoncv/datasets/shopee-iet/... 2 /var/lib/jenkins/.gluoncv/datasets/shopee-iet/... 3 /var/lib/jenkins/.gluoncv/datasets/shopee-iet/... 4 /var/lib/jenkins/.gluoncv/datasets/shopee-iet/... .. ... 75 /var/lib/jenkins/.gluoncv/datasets/shopee-iet/... 76 /var/lib/jenkins/.gluoncv/datasets/shopee-iet/... 77 /var/lib/jenkins/.gluoncv/datasets/shopee-iet/... 78 /var/lib/jenkins/.gluoncv/datasets/shopee-iet/... 79 /var/lib/jenkins/.gluoncv/datasets/shopee-iet/... [80 rows x 4 columns]
An extra column will be included in bulk prediction, indicate the
corresponding image for the row. There will be (# image) rows in the
result, each row includes class
, score
, id
and image
for
prediction class, prediction confidence, class id, and image path
respectively.
Generate image features with a classifier¶
Extracting representation from the whole image learned by a model is
also very useful. We provide predict_feature
function to allow
predictor to return the N-dimensional image feature where N
depends
on the model(usually a 512 to 2048 length vector)
image_path = test_dataset.iloc[0]['image']
feature = predictor.predict_feature(image_path)
print(feature)
image_feature
0 [0.36375043, 0.0, 0.40905723, 1.6886753, 0.402...
Evaluate on Test Dataset¶
You can evaluate the classifier on a test dataset rather than retrieving the predictions.
The validation and test top-1 accuracy are:
test_acc, _ = predictor.evaluate(test_dataset)
print('Top-1 test acc: %.3f' % test_acc)
Top-1 test acc: 0.787
Save and load classifiers¶
You can directly save the instances of classifiers:
filename = 'predictor.ag'
predictor.save(filename)
predictor_loaded = ImagePredictor.load(filename)
# use predictor_loaded as usual
result = predictor_loaded.predict(image_path)
print(result)
class score id
0 BabyShirt 0.783843 1