# Add GST plug-in to support tensors in and out 

Source: [https://docs.qualcomm.com/doc/80-88500-3/topic/75_Add_GST_plug_in_to_support_tensors_in_and_out___.html](https://docs.qualcomm.com/doc/80-88500-3/topic/75_Add_GST_plug_in_to_support_tensors_in_and_out___.html)

## About this task

You can add an ML plug-in to support generic tensors in and tensors out. The decouple
        preprocessing and postprocessing are separate GST components. Hence, you can run customized
        postprocessing after inferencing. You can also configure it in a separate GST plug-in or
        within the docker application.

To add the GST plug-in to support tensors, perform the following:

## Procedure

1. Set up adb root and adb shell.
2. Push models and label files on the device. The models can be downloaded from the TensorFlow Lite webpage.
        
adb push  detect.tflite /data/ 
        adb push detection.labels /data/
        adb push mobilenet_v2_1.0_224_quant.tflite /data/
        adb push classification.labels /data/
        adb push dv3_argmax_int32.tflite /data/
        adb push segmentation.labels /data/Copy to clipboard
3. Weston must be running for display use cases. See [Verify Weston](https://docs.qualcomm.com/doc/80-88500-3/topic/10_Verify_Weston.html#Verify_Weston_10) for more information.

**Parent Topic:** [Configure AI/ML modules](https://docs.qualcomm.com/doc/80-88500-3/topic/66_Configure_AI_ML_modules.html)

Last Published: Sep 26, 2023

[Previous Topic
Test Qualcomm Neural Processing SDK](https://docs.qualcomm.com/bundle/publicresource/80-88500-3/topics/73_Test_SNPE_.md) [Next Topic
Object detection use case](https://docs.qualcomm.com/bundle/publicresource/80-88500-3/topics/object_detection_use_case.md)