Converting a tflite model to onnx model
I developed a model using tf/keras and then converted it into a tflite model before running the following code. It seems that Toolchain cannot open the tflite version of the model. I checked with netron.app and the model seems to be fine. Any help for identifying the root cause will be appreciated.
Thanks
The discussion has been closed due to inactivity. To continue with the topic, please feel free to post a new discussion.
Comments
Update: I took off the AUC metric from the line where I compile the model. However, now I get a error message; "killed". Please see the screenshot.
@Anoop Mishra
Hi Anoop,
You can try to check if your model operators conform to the Kneron hardware architecture, cut out the OPs that don't fit the Kneron hardware architecture, and post-process them on the PC here is the reference link (2.3 Supported operators) :
http://doc.kneron.com/docs/#toolchain/manual/#2-toolchain-docker-overview