Write a custom CUDA kernel to optimize `LaLU` (Laplace Linear Unit).

Formula:
  f(x) = x * (1 - 0.5 * exp(-x))   if x >= 0
  f(x) = x * (0.5 * exp(x))        if x < 0

Problem Analysis:
1. Memory Bound: This is a point-wise activation with low arithmetic intensity.
2. Operator Chaining: The PyTorch implementation using `torch.where` creates intermediate tensors.

Optimization Strategy: Fused Element-wise Kernel with Vectorization

1. One-Thread-per-Element: Map each element to a CUDA thread.

2. Vectorized Loads (float4): Use `float4` to process 128 bits per memory transaction.

3. Fused Branching Logic:
   - For each element `x`, check `if (x >= 0)`.
   - If true, compute `x * (1.0f - 0.5f * __expf(-x))`.
   - If false, compute `x * (0.5f * __expf(x))`.

4. One-Pass: Fuse all steps into a single read-compute-write kernel.
  
Here's an example to show you the syntax of inline embedding custom CUDA operators in torch: The example given architecture is:   
  
```python
import torch
import torch.nn as nn

BATCH_SIZE = 4096
HIDDEN_DIM = 4096
SHAPE = (BATCH_SIZE, HIDDEN_DIM)

class LaLU(nn.Module):
    """
    Laplace Linear Unit (LaLU).
    reference:The Adaptive Quadratic Linear Unit (AQuLU): Adaptive Non Monotonic Piecewise Activation Function
    https://hrcak.srce.hr/file/444170
    Formula:
      f(x) = x * (1 - 0.5 * exp(-x))   if x >= 0
      f(x) = x * (0.5 * exp(x))        if x < 0
    """
    def __init__(self):
        super(LaLU, self).__init__()

    def forward(self, x: torch.Tensor) -> torch.Tensor:
        pos_part = x * (1.0 - 0.5 * torch.exp(-x))
        neg_part = x * (0.5 * torch.exp(x))
        return torch.where(x >= 0, pos_part, neg_part)

class Model(nn.Module):
    def __init__(self):
        super(Model, self).__init__()
        self.act = LaLU()
    
    def forward(self, x):
        return self.act(x)

def get_inputs():
    input_tensor = torch.randn(SHAPE, dtype=torch.float32) * 5.0
    return [input_tensor.contiguous()]

def get_init_inputs():
    return []