MNIST Tutorial errors out when creating object of class PyImageClassifierModel()

[JuliaA369]

Objective of issue: Repair lava/tutorials/end_to_end/tutorial01_mnist_digit_classification.ipynb so that it runs without issues

Lava version:

• 0.4.0 (current version)

I'm submitting a ...

• bug report

Current behavior:

• Errors our during run time

In the 9th cell , during execution of:

mnist_clf.run(
    condition=RunSteps(num_steps=num_steps_per_image),
    run_cfg=Loihi1SimCfg(select_sub_proc_model=True,
                         select_tag='fixed_pt'))

We get the following error

Expected behavior:

Steps to reproduce:

• Run the tutorial lava/tutorials/end_to_end/tutorial01_mnist_digit_classification.ipynb

Related code:
lava/tutorials/end_to_end/tutorial01_mnist_digit_classification.ipynb

Problem can be fixed by defining the shape of the Dense() layer as follows

class PyImageClassifierModel(AbstractSubProcessModel):
    def __init__(self, proc):
        self.dense0 = Dense(weights=proc.w_dense0.init, shape=proc.w_dense0.init.shape)
        self.lif1 = LIF(shape=(64,), bias_mant=proc.b_lif1.init, vth=400,
                        dv=0, du=4095)
        self.dense1 = Dense(weights=proc.w_dense1.init, shape=proc.w_dense1.init.shape)
        self.lif2 = LIF(shape=(64,), bias_mant=proc.b_lif2.init, vth=350,
                        dv=0, du=4095)
        self.dense2 = Dense(weights=proc.w_dense2.init, shape=proc.w_dense2.init.shape)
        self.output_lif = LIF(shape=(10,), bias_mant=proc.b_output_lif.init,
                              vth=1, dv=0, du=4095)

Other information:

insert the output from lava debug here

[JuliaA369]

This was a local pyenv problem. Somehow two different versions of lava were installed locally. When version 0.4.0 is properly installed and used, this problem no longer exists.

[drader]

I am getting the same problem;

the system is;
macOS
lava-nc v0.9.0
python 3.9.18

I've reinstalled and made pytest for lava. However, they still couldn't solve the problem. Would you happen to have any idea or solution for this issue? @JuliaA369