6. Container release notes

6.1. Triton Inference Server

6.1.1. New features

Initial public release of Docker image containing the Triton Inference Server with the Poplar Triton Backend.

6.1.2. Bug fixes

N/A

6.1.3. Other improvements

N/A

6.1.4. Known issues

N/A

6.1.5. Compatibility changes

N/A

6.2. TensorFlow Serving

6.2.1. New features

Provide a compiled version of TensorFlow Serving that supports a TensorFlow custom operator called tensorflow_model_runtime. This custom operator is written using the Poplar Model Runtime API and makes it possible to run PopEF files on TensorFlow Serving.

6.2.2. Bug fixes

N/A

6.2.3. Other improvements

N/A

6.2.4. Known issues

N/A

6.2.5. Compatibility changes

N/A

Search help