```python
import torch
import torch.nn as nn
from torch.utils.cpp_extension import load_inline


_cuda_source = r"""
#include <torch/extension.h>
#include <ATen/cuda/CUDAContext.h>
#include <c10/cuda/CUDAGuard.h>
#include <c10/cuda/CUDAException.h>
#include <cuda_runtime.h>
#include <cmath>
#include <cstdint>

template<int FIXED_CHANNELS>
__global__ void fused_postprocess_kernel(
    const float* __restrict__ input,
    const float* __restrict__ bias,
    float* __restrict__ output,
    int64_t spatial,
    int channels,
    int64_t count
) {
    const int64_t index =
        static_cast<int64_t>(blockIdx.x) * blockDim.x + threadIdx.x;
    if (index >= count) return;

    const int64_t batch = index / spatial;
    const int64_t position = index - batch * spatial;
    const int64_t base = batch * channels * spatial + position;

    float maximum = -INFINITY;
    float sum = 0.0f;

    if constexpr (FIXED_CHANNELS > 0) {
        float values[FIXED_CHANNELS];

        #pragma unroll
        for (int c = 0; c < FIXED_CHANNELS; ++c) {
            values[c] = input[base + static_cast<int64_t>(c) * spatial];
            maximum = fmaxf(maximum, values[c]);
        }

        // Match logsumexp's handling of infinite maxima.
        if (isinf(maximum)) maximum = 0.0f;

        #pragma unroll
        for (int c = 0; c < FIXED_CHANNELS; ++c) {
            sum += __expf(values[c] - maximum);
        }
    } else {
        for (int c = 0; c < channels; ++c) {
            maximum = fmaxf(
                maximum, input[base + static_cast<int64_t>(c) * spatial]
            );
        }

        if (isinf(maximum)) maximum = 0.0f;

        for (int c = 0; c < channels; ++c) {
            sum += __expf(
                input[base + static_cast<int64_t>(c) * spatial] - maximum
            );
        }
    }

    const float lse = maximum + logf(sum);
    const float sigmoid = 1.0f / (1.0f + __expf(-(lse + 3.0f)));
    float value = (lse * sigmoid) / 6.0f - bias[0];

    // Comparisons preserve NaNs, matching torch.clamp.
    value = value < -1.0f ? -1.0f : value;
    value = value > 1.0f ? 1.0f : value;
    output[index] = value;
}

torch::Tensor fused_postprocess(
    torch::Tensor input,
    torch::Tensor bias
) {
    TORCH_CHECK(input.is_cuda() && bias.is_cuda(),
                "Expected CUDA tensors");
    TORCH_CHECK(input.scalar_type() == at::kFloat &&
                bias.scalar_type() == at::kFloat,
                "Expected float32 tensors");
    TORCH_CHECK(input.dim() == 5 && input.is_contiguous(),
                "Expected contiguous NCDHW input");
    TORCH_CHECK(bias.numel() == 1 && input.device() == bias.device(),
                "Expected scalar bias on the input device");

    const c10::cuda::CUDAGuard guard(input.device());

    const int64_t spatial = input.size(2) * input.size(3) * input.size(4);
    const int64_t count = input.size(0) * spatial;
    const int channels = static_cast<int>(input.size(1));

    auto output = torch::empty(
        {input.size(0), 1, input.size(2), input.size(3), input.size(4)},
        input.options()
    );

    if (count == 0) return output;

    constexpr int threads = 256;
    const int blocks = static_cast<int>((count + threads - 1) / threads);
    const auto stream = at::cuda::getCurrentCUDAStream();

    if (channels == 16) {
        fused_postprocess_kernel<16><<<blocks, threads, 0, stream>>>(
            input.data_ptr<float>(), bias.data_ptr<float>(),
            output.data_ptr<float>(), spatial, channels, count
        );
    } else {
        fused_postprocess_kernel<0><<<blocks, threads, 0, stream>>>(
            input.data_ptr<float>(), bias.data_ptr<float>(),
            output.data_ptr<float>(), spatial, channels, count
        );
    }

    C10_CUDA_KERNEL_LAUNCH_CHECK();
    return output;
}
"""

_extension = load_inline(
    name="convtranspose3d_fused_lse_sigmoid_clamp",
    cpp_sources="""
        torch::Tensor fused_postprocess(
            torch::Tensor input, torch::Tensor bias);
    """,
    cuda_sources=_cuda_source,
    functions=["fused_postprocess"],
    extra_cflags=["-O3"],
    extra_cuda_cflags=["-O3", "--fmad=false"],
    verbose=False,
)


class ModelNew(nn.Module):
    def __init__(
        self, in_channels, out_channels, kernel_size, stride, padding, bias_shape
    ):
        super().__init__()
        self.conv_transpose = nn.ConvTranspose3d(
            in_channels,
            out_channels,
            kernel_size,
            stride=stride,
            padding=padding,
        )
        self.bias = nn.Parameter(torch.randn(1, 1, 1, 1))

    def forward(self, x):
        x = self.conv_transpose(x)
        if (
            x.is_cuda
            and x.dtype == torch.float32
            and self.bias.dtype == torch.float32
            and x.is_contiguous()
        ):
            return _extension.fused_postprocess(x, self.bias)

        x = torch.logsumexp(x, dim=1, keepdim=True)
        x = x * torch.sigmoid(x + 3) / 6
        return torch.clamp(x - self.bias, min=-1, max=1)
```