Add python/tensorrt to benchmark

Change-Id: Ib6b852fb08d3714bd30c3ada81b0c09118843ae7
deepjavalibrary · Sep 28, 2021 · cee2ef6 · cee2ef6
1 parent 6c67b00
commit cee2ef6
Show file tree

Hide file tree

Showing 2 changed files with 7 additions and 1 deletion.
diff --git a/extensions/benchmark/README.md b/extensions/benchmark/README.md
@@ -22,8 +22,9 @@ djl-bench currently support benchmark the following type of models:
 - PaddlePaddle model
 - TFLite model
 - TensorRT model
-- Neo DLR (TVM) model
 - XGBoost model
+- Python script model
+- Neo DLR (TVM) model
 
 You can build djl-bench from source if you need to benchmark fastText/BlazingText/Sentencepiece models.
 
@@ -188,6 +189,7 @@ By default, the above script will use MXNet as the default Engine, but you can a
 -e TensorRT # TensorRT
 -e DLR # Neo DLR
 -e XGBoost # XGBoost
+-e Python # Python script
 ```
 
 ### Step 2: Identify the source of your model

diff --git a/extensions/benchmark/build.gradle b/extensions/benchmark/build.gradle
@@ -33,6 +33,10 @@ dependencies {
         runtimeOnly "ai.djl.dlr:dlr-native-auto"
 
         runtimeOnly "ai.djl.ml.xgboost:xgboost"
+
+        runtimeOnly "ai.djl.python:python"
+
+        runtimeOnly "ai.djl.tensorrt:tensorrt"
     } else {
         implementation project(":model-zoo")