This message was deleted.

We built a CI/CD using Gitlab pipelines with the following steps: 1. The model is trained locally, saved to the local Bento store and exported to the remote one on s3. 2. Then we have the pipeline that imports the model from s3, builds the bento and deploys it with bentoctl to our staging environment 3. We run some tests 4. Repeat 2 and 3 with the production environment

Hi Amit, currently, we don’t have an official example, but this is what we are currently working on. For now, an example workflow with GitHub actions that build a Bento and deploy to Yatai/BentoCloud as follow: 1. Save your training model with

bentoml.<framework>.save_model

2. Build the bento with

bentoml build

3. Login into Yatai with

bentoml yatai login --api_token ${{secrets.YATAI_API_TOKEN}} --endpoint ${{secrets.YATAI_API_ENDPOINT}}

4. Push to Yatai:

bentoml push …