Autoscaling TensorFlow Model Deployments with TF Serving and Kubernetes Reviews

Hicham M. · Reviewed около 4 лет ago

Refdt H. · Reviewed около 4 лет ago

Feda N. · Reviewed около 4 лет ago

Kinsi L. · Reviewed около 4 лет ago

Hicham M. · Reviewed около 4 лет ago

Prof G. · Reviewed около 4 лет ago

ibrahimmetehanergun i. · Reviewed около 4 лет ago

Diogo F. · Reviewed около 4 лет ago

Pimple WOW P. · Reviewed около 4 лет ago

Easy to understand

Pedro F. · Reviewed около 4 лет ago

Zlira C. · Reviewed около 4 лет ago

Used tf-serving/*.yaml should've been tfserving-gke/tfserving-gke-autoscaling/tf-serving/*.yaml You could change where users cd and be explicit about it in the instructions, change the overall file structure, or just edit the commands to use this long path _______________________________________ For the following: EXTERNAL_IP=[YOUR_SERVICE_IP] curl -d @locust/request-body.json -X POST http://${EXTERNAL_IP}:8501/v1/models/image_classifier:predict Should be using IP here from `kubectl get svc image-classifier `, which then forwards to 8501: image-classifier LoadBalancer 10.95.246.190 35.222.143.46 8500:30393/TCP,8501:31892/TCP 3m23s Due to directory stucture, the curl command may fail and give the following warnings: Warning: Couldn't read data from file "locust/request-body.json", this makes Warning: an empty POST. _____________________________________ The last sections needed a lot of work, since my resources were not available from the start for my image-classifier model.

Walt H. · Reviewed около 4 лет ago

Phoenix Y. · Reviewed около 4 лет ago

Maria Z. · Reviewed около 4 лет ago

Xavier R. · Reviewed около 4 лет ago

dafa s. · Reviewed около 4 лет ago

dafa s. · Reviewed около 4 лет ago

JaeHwan K. · Reviewed около 4 лет ago

nice

wileji7288 t. · Reviewed около 4 лет ago

Typo in instructions. When we update the configmap.yaml file, we need to change `MODEL_PATH`, not `MODEL_NAME`. It's not difficult to understand what the author means, but the instructions are still wrong. I've noticed A LOT of mistakes in the instructions and have pointed them out in the comments, but as far as I can tell no one is reading our comments. Very frustrating that Qwiklabs encourages us to leave comments, but then doesn't take the time to go through them.

Matthew J. · Reviewed около 4 лет ago

Mike T. · Reviewed около 4 лет ago

Ross M. · Reviewed около 4 лет ago

Emile S. · Reviewed около 4 лет ago

bhi b. · Reviewed около 4 лет ago

Laima K. · Reviewed около 4 лет ago