Error while Compiling Keras to to TVM Run time

Questions
1

I am trying to Complile the Keras Model to Tvm Runtime, Also to generate the Deploypack which can be made run on Embedded Board Further. But when I try to compile the model, I was given out error as below:

TVMError: Check failed:
checked_type.as>IncompleteTypeNode>() == nullptr: Cannot resolve type of Var(lstm_1_input) at (nullptr)

Do anyone know how to fix this

2

can you provide more details on what Keras model you’re trying to compile; and what platform you’re targeting? Are you using NNVM or Relay model importer?

3

This is most likely due to the fact that you incorrectly labeled your input node. Check your Keras input node name with model.input, which should return something like <tf.Tensor 'input_1:0' shape=(?, 227, 227, 3) dtype=float32>. In this case the input node name is input_1, which you have to add to the shape dictionary when calling relay.frontend.from_keras.

4

Thanks lot for your replay, i just fixed this issue