Repository files navigation
🧱 True Layout of the Model
State Dict Key
Component Type
Likely Layer & Role
network.0.weight / bias
Classical Layer
First Convolutional layer (Conv2d) extracting initial spatial patterns from the image.
network.2.weight / bias
Classical Layer
Second Convolutional layer (Conv2d) capturing deeper feature combinations.
network.5.weight / bias
Classical Layer
Third Convolutional layer (Conv2d).
network.7.weight / bias
Classical Layer
Fourth Convolutional layer (Conv2d).
network.10.weight / bias
Classical Layer
Fifth Convolutional layer (Conv2d).
network.12.weight / bias
Classical Layer
Sixth Convolutional layer (Conv2d).
network.16.weight / bias
Classical Layer
Transition Linear layer (Linear) flattening the feature map down.
network.18.weight / bias
Classical Layer
A classical projection layer narrowing down the dimensions.
network.20.weight / bias
Classical Layer
The final classical embedding bottleneck feeding straight into the quantum state.
qnn.weights
Quantum Layer
The variational quantum circuit parameters (Euler rotation angles for the qubits).
fc4.weight / bias
Classical Layer
The final output classification layer mapping quantum outputs to the 10 target categories.
You can’t perform that action at this time.