Describe the bug
When my network is run with too large of an input shape, nothing happens and the program doesn't return any information or errors, it just keeps running. Smaller input shapes work fine, its just after a random size that it stops working.

No such problem occurs when using the CPU back-end so I assume its a Loihi dependent issue.

To reproduce current behavior
Here is the file which causes the problem
lif_conv_model.txt

But here is a snippet for context

# Seems to stop working at some random input size
# Current working limit is (57, 57, 2)
# At another point in time, up to (98, 98, 2) was working
shape = (58, 58, 2)

# If just lif_in layer, then code works fine regardless of input size
lif_in = LIF(shape=shape, vth=50, du=0, dv=0,
          bias_mant=25)

# Addition of conv/lif_out layer seems to be the problem
conv = Conv(
        weight=np.random.randint(256, size=(4, 3, 3, 2))-128,
        input_shape=shape,
        padding=(0,0),
        stride=(2,2),
      )
      
lif_out = LIF(shape=conv.output_shape)

lif_in.s_out.connect(conv.s_in)
conv.a_out.connect(lif_out.a_in)

2. I get this error ...

Program runs forever with no error

Expected behavior
When run, the network should be compiled/partitioned and then run for the specified number of steps.

Screenshots
Expected Output

Current Output

Environment (please complete the following information):

• Device: Intel Cloud - Loihi 2
• OS: Linux
• Lava-Loihi version 0.6.0
• Lava-nc version 0.9.0

Additional context
This is the first issue I have opened on this repository so I apologize if anything is incorrect. Please let me know if I need to add or change anything.