Run_Classification_Sample.md

Run Classification Sample Application with the Optimized SqueezeNet v1.1 Model

In this lab, we are going to run a classification Python sample application with the optimized SqueezeNet v1.1 Model we converted in Lab1 - Optimize a Caffe* Classification Model - SqueezeNet v1.1.

1. Navigate to the directory of the Python Sample Application

cd /opt/intel/openvino_2021/deployment_tools/inference_engine/samples/python

2. Navigate to the directory of classification_sample_async and check the usage of this sample application

cd classification_sample_async
python3 classification_sample_async.py -h

3. Run classification_sample_async Sample Application with SqueezeNet v1.1 model on CPU with FP32 precision model

python3 classification_sample_async.py \
-m /opt/intel/workshop/Squeezenet/FP32/squeezenet1.1_fp32.xml \
-i /opt/intel/workshop/smart-video-workshop/Labs/daisy.jpg \
-d CPU \
--labels /opt/intel/workshop/smart-video-workshop/Labs/squeezetnet_label.txt 

The output would be something like this:

classid probability
------- -----------
985: 'daisy',0.9918603
309: 'bee',0.0062189
108: 'sea anemone, anemone',0.0002938
973: 'coral reef',0.0001974
308: 'fly',0.0001486
301: 'ladybug, ladybeetle, lady beetle, ladybird, ladybird beetle',0.0001268
324: 'cabbage butterfly',0.0001079
947: 'mushroom',0.0001005
995: 'earthstar',0.0000976
506: 'coil, spiral, volute, whorl, helix',0.0000964

4. Switch to another picture and run the same application on GPU with FP16 data precision model

python3 classification_sample_async.py \
-m /opt/intel/workshop/Squeezenet/FP16/squeezenet1.1_fp16.xml \
-i /opt/intel/workshop/smart-video-workshop/Labs/puppy.jpg \
-d GPU \
--labels /opt/intel/workshop/smart-video-workshop/Labs/squeezetnet_label.txt 

You will see the output like this:

classid probability
------- -----------
208: 'Labrador retriever',0.5244141
852: 'tennis ball',0.4282227
805: 'soccer ball',0.0158691
207: 'golden retriever',0.0066185
222: 'kuvasz',0.0064087
178: 'Weimaraner',0.0055656
180: 'American Staffordshire terrier, Staffordshire terrier, American pit bull terrier, pit bull terrier',0.0020828
190: 'Sealyham terrier, Sealyham',0.0020485
251: 'dalmatian, coach dog, carriage dog',0.0012798
176: 'Saluki, gazelle hound',0.0012589

You can download different pictures from the internet, and try the same application with the optimized SqueezeNet v1.1 model.

5. Explore the source code

Now let's take a look at the sourcecode, and learn more about the Inference Engine API, here's the Inference Engine Workflow, you can try to find those functions from the source code:

cd /opt/intel/openvino_2021/deployment_tools/inference_engine/samples/python/classification_sample_async
gedit classification_sample_async.py

Further Reading

To learn more about Image Classification Python* Sample Async, please refer to this OpenVINO documentation Image Classification Python* Sample Async