Can you train a Ludwig model on Google Cloud ML Engine?

ludwig-ai / ludwig

Low-code framework for building custom LLMs, neural networks, and other AI models

http://ludwig.ai

Apache License 2.0

11.09k stars 1.19k forks source link

Can you train a Ludwig model on Google Cloud ML Engine? #208

Open Aashish-1008 opened 5 years ago

w4nderlust commented 5 years ago

I have never tried, I'm not sure what is required to be able to do that. Let's see if someone has tried to do it.

Aashish-1008 commented 5 years ago

I will try to use lugwig on Google Cloud ML Engine and host the model in cloud ml engine. Hope it works.

dawhudlaide commented 5 years ago

I am currently doing sentiment analysis for reviews data using ludwig. I already have my model on the localhost. I'm currently trying it out on GCP hopefully it works.

gogasca commented 5 years ago

In order to train a Ludwig model in Google Cloud ML Engine (AI Platform Training now), you need to create a folder structure and Python code which is compatible with Cloud ML Engine format. (Check example here)

In regards serving If Ludwig generates a [SavedModel] after training (https://www.tensorflow.org/guide/saved_model#build_and_load_a_savedmodel) format, you can configure Serving in Cloud ML Engine.