optimize_weight_bits is increasing the weight matrix scale?

Describe the bug I was observing the synapse weight values stored in a netx model I had exported and they seemed scaled up rather than down. See example:

To reproduce current behavior Steps to reproduce the behavior:

When I run this code (add code or minimum test case) ...


weights = np.array([-1, 1])
optimized_weights, num_weight_bits, weight_exponent, sign_mode = optimize_weight_bits(weights)
print(f"  Original Weights:  {weights}")
print(f"  Optimized Weights: {optimized_weights}")
print(f"  Number of Weight Bits: {num_weight_bits}")
print(f"  Weight Exponent: {weight_exponent}")
print(f"  Sign Mode: {sign_mode}\n")
print("--------------------------------------------")

2. I get this result ...

Original Weights: [-1 1] Optimized Weights: [-128 128] Number of Weight Bits: 2 Weight Exponent: -7 Sign Mode: 1



**Expected behavior**

- optimized_weights should match the original weights I believe

- weight_exponent should be zero. 

- The rest are okay

From what I understand this function should be trying to compress the weights as much as possible. Then we scale the currents by `2**weight_exponent`. 

In short:

```np.all(weight == optimized_weights * 2**weight_exponent)```

Please tell me if this is not the case and will close the issue.

Additional Potential Failure Example:

  Original Weights:  [-1 -2 -3 -4]
  Optimized Weights: [ -64 -128 -192 -256]
  Number of Weight Bits: 2
  Weight Exponent: -6
  Sign Mode: 3

**Environment (please complete the following information):**
 - Device: Local
 - OS: Windows
 - Lava version : main branch

lava-nc / lava-dl

optimize_weight_bits is increasing the weight matrix scale? #304