Test end-to-end prediction pipeline with CPU-only machine

triton-inference-server / fil_backend

FIL backend for the Triton Inference Server

Apache License 2.0

68 stars 35 forks source link

Test end-to-end prediction pipeline with CPU-only machine #87

Closed hcho3 closed 3 years ago

hcho3 commented 3 years ago

Now that CPU prediction is on the docket (#82), we should ensure that the FIL backend can run in an environment with no visible GPU. The CI can test this scenario by launching the test container without the --gpus flag.

For now, I tested #82 locally by manually launching the Docker container with no visible GPU.