I used custom vision to train my object detction model, but my exported tensorflow tflite model cannot loaded rightly in tensorflow example?

bofei li 20 信誉分
2024-02-17T18:00:14.7733333+00:00

I used custom vision(https://www.customvision.ai/) to train my object detction model, but my exported tensorflow tflite model cannot loaded rightly in tensorflow example(https://github.com/tensorflow/examples/tree/cbe96424b6b930fd74dbd4ef2b1e826a031a3554/lite/examples/object_detection). It seems lack of metadata. If I use metadata_writers py script to add metadata to tflite model and load model in the object_detection example, an error like "Mobile SSD models are expected to have exactly 4 outputs, found 3" will occured.

0 个注释 无注释
{count} 票

你的答案

问题作者可以将答案标记为“接受的答案”,这有助于用户了解已解决作者问题的答案。