Open jkennedyvz opened 5 months ago
Most of our users won't be able to upload all of their evidence to a 3rd party for inference. This worker uses the Microsoft phi-3-vision model to answer pre-defined questions of the image evidence you submit.
This works locally, but the init lasts longer than the maximum for aws lambda (10 seconds). I'll package this as an ECS task as well.
Most of our users won't be able to upload all of their evidence to a 3rd party for inference. This worker uses the Microsoft phi-3-vision model to answer pre-defined questions of the image evidence you submit.