Nail/fission-src
1
0
Fork 0
Code Issues Pull Requests Actions Packages Projects Releases Wiki Activity

Allow to set API type for tensorflow serving environment (#1371)

Browse Source 
Tensorflow Serving supports three kinds of API: predict, classify, regress. In order to fully support possible API types, this PR adds code to split the entry point with separator ":" to get model name and the API type separately. And append the API type to proxy URL when receiving the requests.

$ fission fn create --name t1 --pkg <pkg name> \
    --env tensorflow --entrypoint "half_plus_two:regress"
This commit is contained in:
Ta-Ching Chen
2019-11-02 23:51:31 +08:00
committed by GitHub
parent ef0a2fa340
commit 8f6887280c
2 changed files with 25 additions and 4 deletions
Download Patch File Download Diff File Expand all files Collapse all files
examples/tensorflow-serving/README.md
+3 -2
View File
@@ -15,9 +15,10 @@ $ fission pkg create --env tensorflow --deploy half_plus_two.zip
## Create Function
Here, the `--entrypoint` represents the name of top directory contains trained model.
Here, the `--entrypoint` represents the name of the top directory contains the trained model and what kind of API the model supports.
Currently, three API are supported: `predict`, `classify`, `regress`. `predict` API will be used if no API kind was given.
```bash
$ fission fn create --name t1 --pkg <pkg name> --env tensorflow --entrypoint "half_plus_two"
$ fission fn create --name t1 --pkg <pkg name> --env tensorflow --entrypoint "half_plus_two:predict"
$ fission fn test --name t1 --body '{"instances": [1.0, 2.0, 0.0]}' --method POST
```