This function packages the ImageNet library trained on the Inception V3 model which is bundled with TensorFlow 2.0.
You can adapt the code to load other pre-trained models for serving and scaling over HTTP/REST.
Example with image from WWF being classified
Requires:
- AMD64/Intel Operating System and CPU
- Or 64-bit ARM Operating System like Ubuntu 22.04
faas-cli deploy
curl -d https://upload.wikimedia.org/wikipedia/commons/6/61/Humpback_Whale_underwater_shot.jpg \
http://127.0.0.1:8080/function/imagenetSample output:
[
{
"name": "sea_lion",
"score": "0.7445793"
},
{
"name": "great_white_shark",
"score": "0.16700004"
},
{
"name": "grey_whale",
"score": "0.029751772"
}
]To get the timing, run curl -i and look for the header:
X-Duration-Seconds: 0.189897
Asynchronous invocations are also supported, with a Callback-Url if you want to capture the result:
curl -d https://upload.wikimedia.org/wikipedia/commons/6/61/Humpback_Whale_underwater_shot.jpg \
-H "X-Callback-Url: https://http.bin/example" \
http://127.0.0.1:8080/async-function/imagenet