command-batch-inference

Prerequisites

Train the model in the model directory

det e create model/train.yaml model

Running batch inference locally

pip install -r requirements.txt

python batch.py --master-url <master-url> --experiment-id <experiment-id> --input-path <s3-path-to-input-file> --cpu

Running batch inference on determined

det command run --config environment.image=<inference-image> python /batch.py --experiment-id <experiment-id> --input-path <s3-path-to-input-file>

Example:

det command run --config environment.image=seanr3215/batch-inference python /batch.py --experiment-id 1 --input-path s3://sean-test-determined-bucket/input/input.jpg