LDLMIn construction

Contextual evidence from Time embedding

The timestep embedding: cosines and sines of the time multiplied by frequencies falling geometrically from 1 to nearly 1/max_period.

The implementation, pinned at revision e18c42acdfb7, not yet public

Contextual evidence from Time embedding

The denoiser owns the time embedding’s two affine layers, one rotary table for all blocks, an affine projection of the concatenated input, and the stack.

The implementation, pinned at revision e18c42acdfb7, not yet public

Contextual evidence from Time embedding

The time embedding passes its MLP and a SiLU; the noised latents are concatenated with the previous clean estimate, zeros when there is none, and projected; the stack runs in bfloat16.

The implementation, pinned at revision e18c42acdfb7, not yet public

Contextual evidence from Time embedding

Diffusion: self-conditioning on, `t_min` 0.05 and `eps` 10⁻⁵; the adaptive time sampler has 100 bins, a minimum weight of 0.0001, an EMA of 0.9 and updates every 5,000 steps; the denoiser’s EMA decay is 0.9999; scheduler `tanh` with d = 3, solver `euler`, `ode_sampling` false, N = 128 by default; the denoiser has 12 blocks of width 768 with 12 heads, time conditioning `adaln`, a standard MLP, softcap 50 and self-conditioning by concatenation; the timestep embedding’s `max_period` is 10.

The implementation, pinned at revision e18c42acdfb7, not yet public

Contextual evidence from Denoiser block

A block owns two gain-only LayerNorms, a fused query–key–value projection and an output projection without bias, an MLP of the configured kind, and, with modulation, an affine layer from the time condition to six vectors, initialized at zero.

The implementation, pinned at revision e18c42acdfb7, not yet public

Contextual evidence from Denoiser block

A block normalizes, modulates by the time condition’s shift and scale, projects to queries, keys and values, rotates queries and keys, attends over all positions with softcap, and adds the projected output scaled by a gate; the MLP branch does the same with its own shift, scale and gate.

The implementation, pinned at revision e18c42acdfb7, not yet public

Contextual evidence from Denoiser block

The denoiser’s stack applies the blocks in order and then the output layer, each with the time condition.

The implementation, pinned at revision e18c42acdfb7, not yet public

Contextual evidence from Time projection

The output layer: gain-only LayerNorm, a shift and a scale from the time condition, and an affine projection initialized at zero.

The implementation, pinned at revision e18c42acdfb7, not yet public

Contextual evidence from SiLU

PyTorch SiLU applies x × sigmoid(x) element-wise.

token-silu · pytorch/pytorch / torch/nn/modules/activation.py
Pinned revision ee1b6804381c ↗ · lines 368–406

class SiLU(Module):
    r"""Applies the Sigmoid Linear Unit (SiLU) function, element-wise.

    The SiLU function is also known as the swish function.

    .. math::
        \text{silu}(x) = x * \sigma(x), \text{where } \sigma(x) \text{ is the logistic sigmoid.}

    .. note::
        See `Gaussian Error Linear Units (GELUs) <https://arxiv.org/abs/1606.08415>`_
        where the SiLU (Sigmoid Linear Unit) was originally coined, and see
        `Sigmoid-Weighted Linear Units for Neural Network Function Approximation
        in Reinforcement Learning <https://arxiv.org/abs/1702.03118>`_ and `Swish:
        a Self-Gated Activation Function <https://arxiv.org/abs/1710.05941v1>`_
        where the SiLU was experimented with later.

    Shape:
        - Input: :math:`(*)`, where :math:`*` means any number of dimensions.
        - Output: :math:`(*)`, same shape as the input.

    .. image:: ../scripts/activation_images/SiLU.png

    Examples::

        >>> m = nn.SiLU()
        >>> input = torch.randn(2)
        >>> output = m(input)
    """

    __constants__ = ['inplace']
    inplace: bool

    def __init__(self, inplace: bool = False):
        super().__init__()
        self.inplace = inplace

    def forward(self, input: Tensor) -> Tensor:
        return F.silu(input, inplace=self.inplace)