Operator Coverage Matrix

Generated from operators.json version 0.62.4 on 2026-11-02.

Rows are Operator.function pairs. A gap is marked in source with @NotImplemented(reason = …, issue = …) (or @InProgress) and shows up here as ❌ (or 🚧) linked to its tracking issue; the PR that closes the gap removes the annotation. See Issue taxonomy to pick one up.

  • Validated — whether the function’s documentation has been DARC-validated by a reviewer (see DARC workflow).

  • Lowering — the compile-path forms the function declares with @Lowering (composite = a stablehlo.composite when a target keeps it fused, decompose = StableHLO/CHLO primitives, native = a 1:1 target op); — means undeclared.

  • StableHLO — ✅ the StableHLO converter can lower the op; ❌ export fails with "No converter found".

  • Backward — for @Diff ops: ✅ a backward rule is wired; ❌ the rule is missing or a stub returning null gradients; — the op is not differentiable.

  • Backend columns — ✅ the backend overrides the function; ↪ the backend relies on the TensorOps default; 🚧 in progress; ❌ not implemented; — the backend makes no claim (unknown, not "unsupported").

Operator.function Validated Lowering StableHLO Backward apple cpu wasm

TensorOps.add

✖

—

✅

✅

—

✅

—

TensorOps.subtract

✖

—

✅

✅

—

✅

—

TensorOps.multiply

✖

—

✅

✅

—

✅

—

TensorOps.divide

✖

—

✅

✅

—

✅

—

TensorOps.addScalar

✖

—

✅

✅

—

✅

—

TensorOps.subScalar

✖

—

✅

✅

—

✅

—

TensorOps.mulScalar

✖

—

✅

✅

—

✅

—

TensorOps.divScalar

✖

—

✅

✅

—

✅

—

TensorOps.rsubScalar

✖

—

✅

✅

—

✅

—

TensorOps.rdivScalar

✖

—

✅

✅

—

✅

—

TensorOps.matmul

✅

—

✅

✅

—

✅

—

TensorOps.matmulWeightTransposed

✖

—

❌

—

—

✅

—

TensorOps.relayoutPackedWeightForKernels

✖

—

❌

—

—

✅

—

TensorOps.transpose

✖

—

✅

✅

—

✅

—

TensorOps.permute

✖

—

✅

✅

—

✅

—

TensorOps.conv1d

✖

—

✅

✅

—

✅

—

TensorOps.conv2d

✖

—

✅

✅

—

✅

—

TensorOps.conv3d

✖

—

❌

✅

—

✅

—

TensorOps.convTranspose1d

✖

—

❌

✅

—

✅

—

TensorOps.convTranspose2d

✖

—

✅

✅

—

✅

—

TensorOps.maxPool2d

✖

—

✅

✅

—

✅

—

TensorOps.avgPool2d

✖

—

✅

✅

—

✅

—

TensorOps.upsample2d

✖

—

✅

✅

—

✅

—

TensorOps.reshape

✖

—

✅

✅

—

✅

—

TensorOps.flatten

✖

—

✅

✅

—

✅

—

TensorOps.concat

✖

—

✅

✅

—

✅

—

TensorOps.split

✖

—

✅

✅

—

✅

—

TensorOps.squeeze

✖

—

✅

✅

—

✅

—

TensorOps.unsqueeze

✖

—

✅

✅

—

✅

—

TensorOps.relu

✖

—

✅

✅

—

✅

—

TensorOps.leakyRelu

✖

—

❌

✅

—

✅

—

TensorOps.elu

✖

—

❌

✅

—

✅

—

TensorOps.softmax

✖

—

✅

✅

—

✅

—

TensorOps.logSoftmax

✖

—

❌

✅

—

✅

—

TensorOps.sigmoid

✖

—

✅

✅

—

✅

—

TensorOps.tanh

✖

—

✅

✅

—

✅

—

TensorOps.silu

✖

—

✅

✅

—

✅

—

TensorOps.gelu

✖

composite · decompose

✅

✅

—

✅

—

TensorOps.sum

✖

—

✅

✅

—

✅

—

TensorOps.mean

✖

—

✅

✅

—

✅

—

TensorOps.variance

✖

—

✅

✅

—

✅

—

TensorOps.argMax

⚠

—

✅

—

—

✅

—

TensorOps.sqrt

✖

—

✅

✅

—

✅

—

TensorOps.pow

✖

—

✅

✅

—

✅

—

TensorOps.powScalar

✖

—

❌

✅

—

✅

—

TensorOps.log

✖

—

✅

✅

—

✅

—

TensorOps.log2

✖

—

❌

✅

—

✅

—

TensorOps.log10

✖

—

❌

✅

—

✅

—

TensorOps.abs

✖

—

✅

✅

—

✅

—

TensorOps.sign

✖

—

✅

—

—

✅

—

TensorOps.clamp

✖

—

✅

✅

—

✅

—

TensorOps.narrow

✖

—

✅

✅

—

✅

—

TensorOps.pad2d

✖

—

❌

✅

—

✅

—

TensorOps.unfold

✖

—

❌

✅

—

✅

—

TensorOps.lt

✖

—

❌

—

—

✅

—

TensorOps.ge

✖

—

❌

—

—

✅

—

TensorOps.tril

✖

—

❌

✅

—

✅

—

TensorOps.convert

✖

—

✅

—

—

✅

—

TensorOps.gather

✖

—

✅

✅

—

✅

—

TensorOps.indexSelect

✖

—

✅

✅

—

✅

—

TensorOps.exp

✖

—

✅

✅

—

✅

—

TensorOps.expm1

✖

—

✅

✅

—

✅

—

TensorOps.erf

✖

composite · decompose

✅

✅

—

✅

—

TensorOps.sin

✖

—

✅

✅

—

✅

—

TensorOps.cos

✖

—

✅

✅

—

✅

—

TensorOps.scaledDotProductAttention

✖

composite · decompose

✅

❌ stub

—

✅

—

Similarity.cosineDistance

✖

—

❌

—

✅

✅

✅

Done

1 / 67

3 / 67

51 / 67

58 / 59

1 / 67

67 / 67

1 / 67

Per-function detail including notes lives in Operator reference.

NN layers

Every concrete sk.ainet.lang.nn Module subclass. ❌ means the layer’s forward throws today; the class carries @NotImplemented.

Layer Package Status Missing

AvgPool2d

sk.ainet.lang.nn

✅

—

Conv1d

sk.ainet.lang.nn

✅

—

Conv2d

sk.ainet.lang.nn

✅

—

Conv3d

sk.ainet.lang.nn

✅

—

DepthwiseSeparableConv2d

sk.ainet.lang.nn

✅

—

DilatedConv2d

sk.ainet.lang.nn

✅

—

Flatten

sk.ainet.lang.nn

✅

—

GroupedConv2d

sk.ainet.lang.nn

✅

—

Gru

sk.ainet.lang.nn

✅

—

Input

sk.ainet.lang.nn

✅

—

LayerScale

sk.ainet.lang.nn

✅

—

Linear

sk.ainet.lang.nn

✅

—

Lstm

sk.ainet.lang.nn

✅

—

MaxPool2d

sk.ainet.lang.nn

✅

—

TransposedConv1d

sk.ainet.lang.nn

✅

—

TransposedConv2d

sk.ainet.lang.nn

✅

—

Upsample2d

sk.ainet.lang.nn

✅

—

ActivationsWrapperModule

sk.ainet.lang.nn.activations

✅

—

ELU

sk.ainet.lang.nn.activations

✅

—

GELU

sk.ainet.lang.nn.activations

✅

—

LeakyReLU

sk.ainet.lang.nn.activations

✅

—

ReLU

sk.ainet.lang.nn.activations

✅

—

SiLU

sk.ainet.lang.nn.activations

✅

—

Sigmoid

sk.ainet.lang.nn.activations

✅

—

Snake

sk.ainet.lang.nn.activations

✅

—

Softmax

sk.ainet.lang.nn.activations

✅

—

Dropout

sk.ainet.lang.nn.layers

✅

—

LossModule

sk.ainet.lang.nn.loss

✅

—

BatchNormalization

sk.ainet.lang.nn.normalization

✅

—

GroupNormalization

sk.ainet.lang.nn.normalization

✅

—

LayerNormalization

sk.ainet.lang.nn.normalization

✅

—

MLP

sk.ainet.lang.nn.topology

✅

—

Done

32 / 32