You write custom CUDA kernels to replace the pytorch operators in the given architecture to get speedups. 

    You have complete freedom to choose the set of operators you want to replace. You may make the decision to replace some operators with custom CUDA kernels and leave others unchanged. You may replace multiple operators with custom implementations, consider operator fusion opportunities (combining multiple operators into a single kernel, for example, combining matmul+relu), or algorithmic changes (such as online softmax). You are only limited by your imagination.


        Here's an example to show you the syntax of inline embedding custom CUDA operators in torch: The example given architecture is: 

        ``` 

        # Copyright (c) Meta Platforms, Inc. and affiliates.
# All rights reserved.
#
# This source code is licensed under the license found in the
# LICENSE file in the root directory of this source tree.

import torch
import torch.nn as nn
import torch.nn.functional as F


class Model(nn.Module):
    def __init__(self) -> None:
        super().__init__()

    def forward(self, a, b):
        return a + b


def get_inputs():
    # randomly generate input tensors based on the model architecture
    a = torch.randn(1, 128).cuda()
    b = torch.randn(1, 128).cuda()
    return [a, b]


def get_init_inputs():
    # randomly generate tensors required for initialization based on the model architecture
    return []

        ``` 

        The example new arch with custom CUDA kernels looks like this:
        ```
        # Copyright (c) Meta Platforms, Inc. and affiliates.
# All rights reserved.
#
# This source code is licensed under the license found in the
# LICENSE file in the root directory of this source tree.

import torch
import torch.nn as nn
import torch.nn.functional as F
from torch.utils.cpp_extension import load_inline

# Define the custom CUDA kernel for element-wise addition
elementwise_add_source = """
#include <torch/extension.h>
#include <cuda_runtime.h>

__global__ void elementwise_add_kernel(const float* a, const float* b, float* out, int size) {
    int idx = blockIdx.x * blockDim.x + threadIdx.x;
    if (idx < size) {
        out[idx] = a[idx] + b[idx];
    }
}

torch::Tensor elementwise_add_cuda(torch::Tensor a, torch::Tensor b) {
    auto size = a.numel();
    auto out = torch::zeros_like(a);

    const int block_size = 256;
    const int num_blocks = (size + block_size - 1) / block_size;

    elementwise_add_kernel<<<num_blocks, block_size>>>(a.data_ptr<float>(), b.data_ptr<float>(), out.data_ptr<float>(), size);

    return out;
}
"""

elementwise_add_cpp_source = (
    "torch::Tensor elementwise_add_cuda(torch::Tensor a, torch::Tensor b);"
)

# Compile the inline CUDA code for element-wise addition
elementwise_add = load_inline(
    name="elementwise_add",
    cpp_sources=elementwise_add_cpp_source,
    cuda_sources=elementwise_add_source,
    functions=["elementwise_add_cuda"],
    verbose=True,
    extra_cflags=[""],
    extra_ldflags=[""],
)


class ModelNew(nn.Module):
    def __init__(self) -> None:
        super().__init__()
        self.elementwise_add = elementwise_add

    def forward(self, a, b):
        return self.elementwise_add.elementwise_add_cuda(a, b)

        ``` 

        
    You are given the following architecture: 

    ```
    # Copyright (c) Meta Platforms, Inc. and affiliates.
# All rights reserved.
#
# This source code is licensed under the license found in the
# LICENSE file in the root directory of this source tree.

import torch
import torch.nn as nn

class Model(nn.Module):
    """
    This model performs a convolution, average pooling, applies sigmoid, and sums the result.
    """
    def __init__(self, in_channels, out_channels, kernel_size, pool_kernel_size):
        super(Model, self).__init__()
        self.conv = nn.Conv2d(in_channels, out_channels, kernel_size)
        self.avg_pool = nn.AvgPool2d(pool_kernel_size)

    def forward(self, x):
        x = self.conv(x)
        x = self.avg_pool(x)
        x = torch.sigmoid(x)
        x = torch.sum(x, dim=[1,2,3]) # Sum over all spatial dimensions
        return x

batch_size = 128
in_channels = 8
out_channels = 64
height, width = 384, 384
kernel_size = 3
pool_kernel_size = 4

def get_inputs():
    return [torch.rand(batch_size, in_channels, height, width)]

def get_init_inputs():
    return [in_channels, out_channels, kernel_size, pool_kernel_size]
    ```
    
Optimize the architecture named Model with custom CUDA operators! Name your optimized output architecture ModelNew. Output the new code in codeblocks. Please generate real code, NOT pseudocode, make sure the code compiles and is fully functional. Just output the new model code, no other text, and NO testing code! 



This is turn 3 of 6 for this task. Your previous solution (turn 2) was:
```python
# Copyright (c) Meta Platforms, Inc. and affiliates.
# All rights reserved.
#
# This source code is licensed under the license found in the
# LICENSE file in the root directory of this source tree.

import torch
import torch.nn as nn
import torch.nn.functional as F
from torch.utils.cpp_extension import load_inline

# Let cuDNN autotune over ALL candidate engines (cached after warmup).
torch.backends.cudnn.benchmark = True
torch.backends.cudnn.benchmark_limit = 0

# Custom CUDA ops:
#   1) nchw_to_nhwc_fp16: fuses fp32->fp16 cast + NCHW->NHWC transpose in one
#      pass so the convolution can run natively in NHWC (channels_last). This
#      eliminates cuDNN's expensive NHWC->NCHW output re-transpose that
#      dominated the previous version.
#   2) fused_pool_sigmoid_sum_nhwc: non-overlapping AvgPool + sigmoid + global
#      sum, reading the NHWC conv output exactly once with fully coalesced
#      16-byte loads.
#   3) fused_pool_sigmoid_sum_nchw: same fusion for NCHW layout (safety net).
_cuda_src = r"""
#include <torch/extension.h>
#include <ATen/cuda/CUDAContext.h>
#include <cuda_runtime.h>
#include <cuda_fp16.h>

namespace {

__device__ __forceinline__ void block_reduce_atomic(float acc, float* out) {
    #pragma unroll
    for (int off = 16; off > 0; off >>= 1)
        acc += __shfl_down_sync(0xffffffffu, acc, off);
    __shared__ float warp_sums[32];
    const int lane = threadIdx.x & 31;
    const int wid  = threadIdx.x >> 5;
    if (lane == 0) warp_sums[wid] = acc;
    __syncthreads();
    if (wid == 0) {
        const int nwarps = (blockDim.x + 31) >> 5;
        float v = (lane < nwarps) ? warp_sums[lane] : 0.0f;
        #pragma unroll
        for (int off = 16; off > 0; off >>= 1)
            v += __shfl_down_sync(0xffffffffu, v, off);
        if (lane == 0) atomicAdd(out, v);
    }
}

// fp32 NCHW -> fp16 NHWC, one thread per pixel, C == 8 (vectorized 16B store).
__global__ void nchw_to_nhwc_fp16_c8_kernel(
        const float* __restrict__ in, __half* __restrict__ out,
        const size_t HW, const size_t total_pixels) {
    const size_t p = (size_t)blockIdx.x * blockDim.x + threadIdx.x;
    if (p >= total_pixels) return;
    const size_t n = p / HW;
    const size_t s = p - n * HW;
    const float* base = in + n * 8 * HW + s;
    uint4 packed;
    __half2* ph = reinterpret_cast<__half2*>(&packed);
    #pragma unroll
    for (int c = 0; c < 4; ++c) {
        ph[c] = __floats2half2_rn(base[(size_t)(2 * c) * HW],
                                  base[(size_t)(2 * c + 1) * HW]);
    }
    *reinterpret_cast<uint4*>(out + p * 8) = packed;
}

// Generic-C variant (one thread per output element).
__global__ void nchw_to_nhwc_fp16_generic_kernel(
        const float* __restrict__ in, __half* __restrict__ out,
        const int C, const size_t HW, const size_t total) {
    const size_t idx = (size_t)blockIdx.x * blockDim.x + threadIdx.x;
    if (idx >= total) return;
    const int c = (int)(idx % (size_t)C);
    const size_t p = idx / (size_t)C;
    const size_t n = p / HW;
    const size_t s = p - n * HW;
    out[idx] = __float2half_rn(in[(n * (size_t)C + c) * HW + s]);
}

// AvgPool(non-overlapping) + sigmoid + sum over [C,H,W], NHWC input.
// Each thread handles one pooling window for a group of 8 channels using
// 16-byte vectorized loads; consecutive threads read consecutive memory.
template <int POOL>
__global__ void fused_pool_sigmoid_sum_nhwc_kernel(
        const __half* __restrict__ x, float* __restrict__ out,
        const int C, const int H, const int W,
        const int Hp, const int Wp, const int units_per_image) {
    const int n = blockIdx.y;
    const __half* __restrict__ img = x + (size_t)n * ((size_t)H * W * C);
    const int cg_count = C >> 3;

    float acc = 0.0f;
    const int stride = gridDim.x * blockDim.x;
    for (int v = blockIdx.x * blockDim.x + threadIdx.x; v < units_per_image; v += stride) {
        const int cg = v % cg_count;
        const int t  = v / cg_count;
        const int wx = t % Wp;
        const int wy = t / Wp;
        const int c0 = cg << 3;

        const __half* base = img + ((size_t)(wy * POOL) * W + (size_t)(wx * POOL)) * C + c0;
        float sums[8] = {0.f, 0.f, 0.f, 0.f, 0.f, 0.f, 0.f, 0.f};
        #pragma unroll
        for (int r = 0; r < POOL; ++r) {
            const __half* row = base + (size_t)r * W * C;
            #pragma unroll
            for (int q = 0; q < POOL; ++q) {
                const uint4 raw = *reinterpret_cast<const uint4*>(row + (size_t)q * C);
                const __half2* h2 = reinterpret_cast<const __half2*>(&raw);
                #pragma unroll
                for (int i = 0; i < 4; ++i) {
                    const float2 f = __half22float2(h2[i]);
                    sums[2 * i]     += f.x;
                    sums[2 * i + 1] += f.y;
                }
            }
        }
        constexpr float inv = 1.0f / (float)(POOL * POOL);
        #pragma unroll
        for (int i = 0; i < 8; ++i) {
            const float m = sums[i] * inv;
            acc += 1.0f / (1.0f + expf(-m));
        }
    }
    block_reduce_atomic(acc, &out[n]);
}

__device__ __forceinline__ float2 load_two(const float* p) {
    return *reinterpret_cast<const float2*>(p);
}
__device__ __forceinline__ float2 load_two(const __half* p) {
    return __half22float2(*reinterpret_cast<const __half2*>(p));
}

// AvgPool(non-overlapping) + sigmoid + sum over [C,H,W], NCHW input (backup).
template <typename scalar_t, int POOL>
__global__ void fused_pool_sigmoid_sum_nchw_kernel(
        const scalar_t* __restrict__ x, float* __restrict__ out,
        const int C, const int H, const int W,
        const int Hp, const int Wp, const int num_windows) {
    const int n = blockIdx.y;
    const scalar_t* __restrict__ img = x + (size_t)n * C * H * W;

    float acc = 0.0f;
    const int stride = gridDim.x * blockDim.x;
    constexpr float inv = 1.0f / (float)(POOL * POOL);

    for (int v = blockIdx.x * blockDim.x + threadIdx.x; v < num_windows; v += stride) {
        const int wx = v % Wp;
        const int t  = v / Wp;
        const int wy = t % Hp;
        const int c  = t / Hp;

        const scalar_t* p = img + ((size_t)c * H + (size_t)wy * POOL) * W + (size_t)wx * POOL;
        float s = 0.0f;
        #pragma unroll
        for (int r = 0; r < POOL; ++r) {
            #pragma unroll
            for (int q = 0; q < POOL / 2; ++q) {
                const float2 val = load_two(p + 2 * q);
                s += val.x + val.y;
            }
            p += W;
        }
        const float m = s * inv;
        acc += 1.0f / (1.0f + expf(-m));
    }
    block_reduce_atomic(acc, &out[n]);
}

} // namespace

torch::Tensor nchw_to_nhwc_fp16(torch::Tensor x) {
    TORCH_CHECK(x.is_cuda() && x.dim() == 4, "input must be CUDA NCHW");
    TORCH_CHECK(x.scalar_type() == torch::kFloat, "input must be fp32");
    auto xc = x.contiguous();
    const int64_t N = xc.size(0), C = xc.size(1), H = xc.size(2), W = xc.size(3);

    auto out = torch::empty({N, C, H, W},
        xc.options().dtype(torch::kHalf).memory_format(torch::MemoryFormat::ChannelsLast));

    const size_t HW = (size_t)H * (size_t)W;
    const size_t total_pixels = (size_t)N * HW;
    if (total_pixels == 0) return out;

    auto stream = at::cuda::getCurrentCUDAStream();
    const int threads = 256;
    __half* out_ptr = reinterpret_cast<__half*>(out.data_ptr<at::Half>());

    if (C == 8) {
        const size_t blocks = (total_pixels + threads - 1) / threads;
        nchw_to_nhwc_fp16_c8_kernel<<<blocks, threads, 0, stream>>>(
            xc.data_ptr<float>(), out_ptr, HW, total_pixels);
    } else {
        const size_t total = total_pixels * (size_t)C;
        const size_t blocks = (total + threads - 1) / threads;
        nchw_to_nhwc_fp16_generic_kernel<<<blocks, threads, 0, stream>>>(
            xc.data_ptr<float>(), out_ptr, (int)C, HW, total);
    }
    return out;
}

torch::Tensor fused_pool_sigmoid_sum_nhwc(torch::Tensor x, int64_t pool) {
    TORCH_CHECK(x.is_cuda() && x.dim() == 4, "input must be CUDA NCHW4d");
    TORCH_CHECK(x.scalar_type() == torch::kHalf, "input must be fp16");
    TORCH_CHECK(x.is_contiguous(at::MemoryFormat::ChannelsLast), "input must be channels_last");

    const int N = (int)x.size(0);
    const int C = (int)x.size(1);
    const int H = (int)x.size(2);
    const int W = (int)x.size(3);
    TORCH_CHECK(C % 8 == 0, "channels must be a multiple of 8");

    const int Hp = H / (int)pool;
    const int Wp = W / (int)pool;
    auto out = torch::zeros({N}, x.options().dtype(torch::kFloat32));
    if (N == 0 || C == 0 || Hp == 0 || Wp == 0) return out;

    const int units = Hp * Wp * (C >> 3);
    const int threads = 256;
    int blocks = (units + threads - 1) / threads;
    if (blocks > 1024) blocks = 1024;
    dim3 grid(blocks, N);
    auto stream = at::cuda::getCurrentCUDAStream();
    const __half* ptr = reinterpret_cast<const __half*>(x.data_ptr<at::Half>());

    if (pool == 4) {
        fused_pool_sigmoid_sum_nhwc_kernel<4><<<grid, threads, 0, stream>>>(
            ptr, out.data_ptr<float>(), C, H, W, Hp, Wp, units);
    } else if (pool == 2) {
        fused_pool_sigmoid_sum_nhwc_kernel<2><<<grid, threads, 0, stream>>>(
            ptr, out.data_ptr<float>(), C, H, W, Hp, Wp, units);
    } else {
        TORCH_CHECK(false, "unsupported pool size");
    }
    return out;
}

torch::Tensor fused_pool_sigmoid_sum_nchw(torch::Tensor x, int64_t pool) {
    TORCH_CHECK(x.is_cuda() && x.dim() == 4, "input must be CUDA NCHW");
    auto xc = x.contiguous();

    const int N = (int)xc.size(0);
    const int C = (int)xc.size(1);
    const int H = (int)xc.size(2);
    const int W = (int)xc.size(3);
    TORCH_CHECK(W % 2 == 0, "width must be even");

    const int Hp = H / (int)pool;
    const int Wp = W / (int)pool;
    auto out = torch::zeros({N}, xc.options().dtype(torch::kFloat32));
    if (N == 0 || C == 0 || Hp == 0 || Wp == 0) return out;

    const int num_windows = C * Hp * Wp;
    const int threads = 256;
    int blocks = (num_windows + threads - 1) / threads;
    if (blocks > 1024) blocks = 1024;
    dim3 grid(blocks, N);
    auto stream = at::cuda::getCurrentCUDAStream();

    if (xc.scalar_type() == torch::kHalf) {
        const __half* ptr = reinterpret_cast<const __half*>(xc.data_ptr<at::Half>());
        if (pool == 4) {
            fused_pool_sigmoid_sum_nchw_kernel<__half, 4><<<grid, threads, 0, stream>>>(
                ptr, out.data_ptr<float>(), C, H, W, Hp, Wp, num_windows);
        } else if (pool == 2) {
            fused_pool_sigmoid_sum_nchw_kernel<__half, 2><<<grid, threads, 0, stream>>>(
                ptr, out.data_ptr<float>(), C, H, W, Hp, Wp, num_windows);
        } else {
            TORCH_CHECK(false, "unsupported pool size");
        }
    } else if (xc.scalar_type() == torch::kFloat) {
        const float* ptr = xc.data_ptr<float>();
        if (pool == 4) {
            fused_pool_sigmoid_sum_nchw_kernel<float, 4><<<grid, threads, 0, stream>>>(
                ptr, out.data_ptr<float>(), C, H, W, Hp, Wp, num_windows);
        } else if (pool == 2) {
            fused_pool_sigmoid_sum_nchw_kernel<float, 2><<<grid, threads, 0, stream>>>(
                ptr, out.data_ptr<float>(), C, H, W, Hp, Wp, num_windows);
        } else {
            TORCH_CHECK(false, "unsupported pool size");
        }
    } else {
        TORCH_CHECK(false, "unsupported dtype");
    }
    return out;
}
"""

_cpp_src = (
    "torch::Tensor nchw_to_nhwc_fp16(torch::Tensor x);\n"
    "torch::Tensor fused_pool_sigmoid_sum_nhwc(torch::Tensor x, int64_t pool);\n"
    "torch::Tensor fused_pool_sigmoid_sum_nchw(torch::Tensor x, int64_t pool);"
)

_ext = load_inline(
    name="conv_pool_sigmoid_fused_v2",
    cpp_sources=_cpp_src,
    cuda_sources=_cuda_src,
    functions=[
        "nchw_to_nhwc_fp16",
        "fused_pool_sigmoid_sum_nhwc",
        "fused_pool_sigmoid_sum_nchw",
    ],
    verbose=False,
    extra_cuda_cflags=["-O3"],
)


class ModelNew(nn.Module):
    """
    Optimized version:
      * The fp32->fp16 cast and NCHW->NHWC transpose of the input are fused into
        a single custom kernel, so the convolution runs natively in
        channels_last (NHWC) on tensor cores with NO layout re-transposes
        (the NHWC->NCHW output transpose dominated the previous version).
      * fp16 weights in channels_last are cached across calls.
      * AvgPool + sigmoid + global sum are fused into one custom CUDA kernel
        that reads the NHWC conv output exactly once with 16-byte coalesced
        loads and emits one fp32 scalar per batch element.
    Falls back to the original PyTorch ops for unsupported configurations.
    """
    def __init__(self, in_channels, out_channels, kernel_size, pool_kernel_size):
        super(ModelNew, self).__init__()
        self.conv = nn.Conv2d(in_channels, out_channels, kernel_size)
        self.avg_pool = nn.AvgPool2d(pool_kernel_size)
        self.pool_kernel_size = pool_kernel_size
        self._w16 = None
        self._b16 = None
        self._w_key = None

    def _half_weights(self):
        w = self.conv.weight
        key = (w.data_ptr(), w._version)
        if self._w16 is None or self._w_key != key:
            self._w16 = w.detach().half().contiguous(memory_format=torch.channels_last)
            b = self.conv.bias
            self._b16 = b.detach().half() if b is not None else None
            self._w_key = key
        return self._w16, self._b16

    def forward(self, x):
        pool = self.pool_kernel_size
        fast = (
            x.is_cuda
            and x.dtype == torch.float32
            and isinstance(pool, int)
            and pool in (2, 4)
            and self.conv.out_channels % 8 == 0
        )
        if fast:
            try:
                w16, b16 = self._half_weights()
                x16 = _ext.nchw_to_nhwc_fp16(x)
                y = F.conv2d(
                    x16,
                    w16,
                    b16,
                    self.conv.stride,
                    self.conv.padding,
                    self.conv.dilation,
                    self.conv.groups,
                )
                if y.is_contiguous(memory_format=torch.channels_last):
                    return _ext.fused_pool_sigmoid_sum_nhwc(y, pool)
                if y.is_contiguous() and y.shape[-1] % 2 == 0:
                    return _ext.fused_pool_sigmoid_sum_nchw(y, pool)
                # Unexpected layout: cheap torch fallback on the conv output.
                y = self.avg_pool(y.float())
                return torch.sum(torch.sigmoid(y), dim=[1, 2, 3])
            except RuntimeError:
                pass

        x = self.conv(x)
        x = self.avg_pool(x)
        x = torch.sigmoid(x)
        return torch.sum(x, dim=[1, 2, 3])
```
Result of evaluating it:
The solution is correct on the test inputs. Mean runtime 11.5 ms; the PyTorch reference (with TF32 enabled) takes 36.9 ms; speedup 3.209x.
Profiler, top device kernels of one forward pass of your solution:
| device kernel | total device time (us) | calls |
|---|---|---|
| void at::native::elementwise_kernel<128, 4, at::native::gpu_kernel_impl_nocast<at::native: | 5764.0 | 2 |
| sm80_xmma_fprop_implicit_gemm_indexed_wo_smem_f16f16_f16f32_f32_nhwckrsc_nhwc_tilesize128x | 3566.9 | 1 |
| void (anonymous namespace)::fused_pool_sigmoid_sum_nhwc_kernel<4>(__half const*, float*, i | 1607.1 | 1 |
| (anonymous namespace)::nchw_to_nhwc_fp16_c8_kernel(float const*, __half*, unsigned long, u | 668.7 | 1 |
| void at::native::vectorized_elementwise_kernel<4, at::native::FillFunctor<float>, std::arr | 2.1 | 1 |

Write an improved complete ModelNew: fix any error, and if the solution is already correct, make it faster. Output the new code in one ```python codeblock. Just the model code, no other text, and NO testing code.
