How Cloud Functions Work in GCP
Cloud Functions help developers and data scientists to apply standard algorithms to train the model and deploy the model on the edge networking devices without maintaining any infrastructure.
Cloud Function offers an event-driven serverless computing platform, which gives provision to the developers to code locally or in the cloud without managing the server, as it is a serverless approach. It can be integrated with Cloud Firestore to run the backend services and to run the labeled data for training and deployment on the edge, using edge computing.
Benefits of Cloud Functions in GCP
- No provisioning of server
- Auto Scale functionalities based on high load traffic
- Level based security functions
- High networking capabilities for multi-cloud
R Interface to Google CloudML
- Google Cloud ML function provides an R interface that helps in the following ways to train the labeled data
- Training of models built with the TensorFlow and R packages, which provides full access to the computational graph
- Hyper Parameterized tuning optimizes the key attributes of model architectures maximizing predictive accuracy
Hyperparameter Tuning in Cloud ML
It is the automated model enhancer provided by ML engine. Hyperparameter tuning offers models to configure when training the actual model.
Benefits of using Hyperparameter in Cloud ML are:
- Code for implementing the inference, for which the developers are building the custom logic and standard algorithms
- Higher Scalability with maximum accuracy
- Developers can use custom versions of different frameworks PyTorch, Keras, and TensorFlow, just to name a few
Mechanism of Cloud Functions
The model data is trained locally and then these are uploaded to Cloud Storage. Then the cloud functions will be invoked through an API request to download the model and a test image from Cloud Storage. A cluster will be used as an inference that might be used for the model.
This above diagram clarifies how requests are generated to train the ML Model.
