RuntimeError: Internal: Unsupported data type in custom op handler: -476681008Node number 1 (EdgeTpuDelegateForCustomOp) failed to prepare

guichristmann / edge-tpu-tiny-yolo

Run Tiny YOLO-v3 on Google's Edge TPU USB Accelerator.

MIT License

102 stars 31 forks source link

RuntimeError: Internal: Unsupported data type in custom op handler: -476681008Node number 1 (EdgeTpuDelegateForCustomOp) failed to prepare #24

Closed AkkiSony closed 3 years ago

AkkiSony commented 3 years ago

I am getting this error when I try to run Inference.py script on my yolov3 model.

coral-TPU-error-26

My current conda environment consists of:

Tensorflow: 1.15.0 Keras: 2.2.4 numpy: 1.21.2 opencv-python: 4.5.3.56

I am working on windows 10 with Python: 3.7.3

Any leads would really help me alot. Thanks in advance!

AkkiSony commented 3 years ago

I could over come the error by following: google-coral/tflite#37 (comment)

But however, detection on the images are very poor in some cases and on the other cases, theere is no predicitons at all. Is it necessary to modify the script?

Thnks in advance! :)