yobx.xoptim.patterns.onnx_split — yet-another-onnx-builder 0.1.0 documentation

class yobx.xoptim.patterns.onnx_split.GathersSplitPattern(verbose: int = 0, priority: int = 1, min_opset: int = 1)[source]#

Merges multiple parallel gather into a split followed by unsqueeze.

Model with nodes to be fused:

        graph TD

    classDef ioNode fill:#dfd,stroke:#333,color:#333
    classDef initNode fill:#cccc00,stroke:#333,color:#333
    classDef constNode fill:#f9f,stroke:#333,stroke-width:2px,color:#333
    classDef opNode fill:#bbf,stroke:#333,stroke-width:2px,color:#333

    I_X(["X FLOAT(a, 2)"])

    Gather_0[["Gather(., 0, axis=1)"]]
    Gather_1[["Gather(., 1, axis=1)"]]

    I_X -->|"FLOAT(a, 2)"| Gather_0
    I_X -->|"FLOAT(a, 2)"| Gather_1

    O_x2(["x2 FLOAT(a)"])
    Gather_1 --> O_x2
    O_x1(["x1 FLOAT(a)"])
    Gather_0 --> O_x1

    class I_X,O_x2,O_x1 ioNode
    class Gather_0,Gather_1 opNode

Outcome of the fusion:

        graph TD

    classDef ioNode fill:#dfd,stroke:#333,color:#333
    classDef initNode fill:#cccc00,stroke:#333,color:#333
    classDef constNode fill:#f9f,stroke:#333,stroke-width:2px,color:#333
    classDef opNode fill:#bbf,stroke:#333,stroke-width:2px,color:#333

    I_X(["X FLOAT(a, 2)"])

    Split_0[["Split(., axis=1)"]]
    Squeeze_1[["Squeeze(., [1])"]]
    Squeeze_2[["Squeeze(., [1])"]]

    I_X -->|"FLOAT(a, 2)"| Split_0
    Split_0 -->|"FLOAT(a, 1)"| Squeeze_1
    Split_0 -->|"FLOAT(a, 1)"| Squeeze_2

    O_x2(["x2 FLOAT(a)"])
    Squeeze_2 --> O_x2
    O_x1(["x1 FLOAT(a)"])
    Squeeze_1 --> O_x1

    class I_X,O_x2,O_x1 ioNode
    class Split_0,Squeeze_1,Squeeze_2 opNode

apply(g: GraphBuilder, *gather_nodes: NodeProto) → List[NodeProto][source]#

The method does the rewriting. It assumes it can happen. It takes a list of nodes impacted by the rewriting assumes no other pattern optimizer will be modify them. It receives the list of nodes returned by method apply. Since it is a list of argument, method match can include None values. The method returns the new nodes. The optimizer considers that any node given to this function is removed from the graph, and any node returned by it are added. If a received node must be kept, it must be added to the list of returned node.

Parameters:: nodes – nodes returned by method match, there are then removed
Returns:: nodes to add to graph.

match(g: GraphBuilderPatternOptimization, node: NodeProto, matched: List[MatchResult]) → MatchResult | None[source]#

Determines nodes around node which can be rewritten.

Parameters:

g – is a GraphBuilderPatternOptimization, it holds all the existing nodes, is able to return any information about type, shape, the node before, the node after another one.
node – the matching must determine if some nodes around this one are part of set of nodes this pattern optimizer can rewrite. From there, the function explores wherever it needs, checking any condition it needs.
matched – usually unused, it returns of nodes already matching a pattern

The method must not modify the graph. The method returns None if no match is found or an instance of class MatchResult. It must contain:

a list of nodes involved in the rewriting. It does not mean all of them will be removed but all of them are needed to do the rewriting and must not be impacted by other pattern optimizer.
A function doing the rewriting (usually method apply of the pattern class).
An existing node where the rewritten nodes can be inserted. Knowing it makes it faster to rewriter. If not specified, the optimizer will automatically determine the position of the new nodes.

class yobx.xoptim.patterns.onnx_split.SlicesSplitPattern(verbose: int = 0, priority: int = 1, min_opset: int = 1)[source]#

Merges multiple parallel slices into a split.

Model with nodes to be fused:

        graph TD

    classDef ioNode fill:#dfd,stroke:#333,color:#333
    classDef initNode fill:#cccc00,stroke:#333,color:#333
    classDef constNode fill:#f9f,stroke:#333,stroke-width:2px,color:#333
    classDef opNode fill:#bbf,stroke:#333,stroke-width:2px,color:#333

    I_transpose_1(["transpose_1 FLOAT16(2, 2, 1024, 512)"])

    Slice_0[["Slice(., [0], [256], [3])"]]
    Slice_1[["Slice(., [256], [9223372036854775807], [3])"]]

    I_transpose_1 -->|"FLOAT16(2, 2, 1024, 512)"| Slice_0
    I_transpose_1 -->|"FLOAT16(2, 2, 1024, 512)"| Slice_1

    O_slice_11(["slice_11 FLOAT16(2, 2, 1024, 256)"])
    Slice_0 --> O_slice_11
    O_slice_12(["slice_12 FLOAT16(2, 2, 1024, 256)"])
    Slice_1 --> O_slice_12

    class I_transpose_1,O_slice_11,O_slice_12 ioNode
    class Slice_0,Slice_1 opNode

Outcome of the fusion:

        graph TD

    classDef ioNode fill:#dfd,stroke:#333,color:#333
    classDef initNode fill:#cccc00,stroke:#333,color:#333
    classDef constNode fill:#f9f,stroke:#333,stroke-width:2px,color:#333
    classDef opNode fill:#bbf,stroke:#333,stroke-width:2px,color:#333

    I_transpose_1(["transpose_1 FLOAT16(2, 2, 1024, 512)"])

    Split_0[["Split(., [256, 256], axis=3)"]]

    I_transpose_1 -->|"FLOAT16(2, 2, 1024, 512)"| Split_0

    O_slice_11(["slice_11 FLOAT16(2, 2, 1024, 256)"])
    Split_0 --> O_slice_11
    O_slice_12(["slice_12 FLOAT16(2, 2, 1024, 256)"])
    Split_0 --> O_slice_12

    class I_transpose_1,O_slice_11,O_slice_12 ioNode
    class Split_0 opNode

apply(g: GraphBuilder, *nodes: NodeProto) → List[NodeProto][source]#

The method does the rewriting. It assumes it can happen. It takes a list of nodes impacted by the rewriting assumes no other pattern optimizer will be modify them. It receives the list of nodes returned by method apply. Since it is a list of argument, method match can include None values. The method returns the new nodes. The optimizer considers that any node given to this function is removed from the graph, and any node returned by it are added. If a received node must be kept, it must be added to the list of returned node.

Parameters:: nodes – nodes returned by method match, there are then removed
Returns:: nodes to add to graph.

match(g: GraphBuilderPatternOptimization, node: NodeProto, matched: List[MatchResult]) → MatchResult | None[source]#

Determines nodes around node which can be rewritten.

Parameters:

g – is a GraphBuilderPatternOptimization, it holds all the existing nodes, is able to return any information about type, shape, the node before, the node after another one.
node – the matching must determine if some nodes around this one are part of set of nodes this pattern optimizer can rewrite. From there, the function explores wherever it needs, checking any condition it needs.
matched – usually unused, it returns of nodes already matching a pattern

The method must not modify the graph. The method returns None if no match is found or an instance of class MatchResult. It must contain:

a list of nodes involved in the rewriting. It does not mean all of them will be removed but all of them are needed to do the rewriting and must not be impacted by other pattern optimizer.
A function doing the rewriting (usually method apply of the pattern class).
An existing node where the rewritten nodes can be inserted. Knowing it makes it faster to rewriter. If not specified, the optimizer will automatically determine the position of the new nodes.

class yobx.xoptim.patterns.onnx_split.SplitConcatPattern(verbose: int = 0, priority: int = 1, min_opset: int = 1)[source]#

Replaces Split + Concat into identity if this is equivalent.

Model with nodes to be fused:

        graph TD

    classDef ioNode fill:#dfd,stroke:#333,color:#333
    classDef initNode fill:#cccc00,stroke:#333,color:#333
    classDef constNode fill:#f9f,stroke:#333,stroke-width:2px,color:#333
    classDef opNode fill:#bbf,stroke:#333,stroke-width:2px,color:#333

    I_X(["X FLOAT(a, b)"])

    Split_0[["Split(., axis=-1)"]]
    Concat_1[["Concat(., ., axis=-1)"]]

    I_X -->|"FLOAT(a, b)"| Split_0
    Split_0 -->|"FLOAT(a, CeilToInt(b,2))"| Concat_1
    Split_0 -->|"FLOAT(a, b-CeilToInt(b,2))"| Concat_1

    O_Y(["Y FLOAT(a, b)"])
    Concat_1 --> O_Y

    class I_X,O_Y ioNode
    class Split_0,Concat_1 opNode

Outcome of the fusion:

        graph TD

    classDef ioNode fill:#dfd,stroke:#333,color:#333
    classDef initNode fill:#cccc00,stroke:#333,color:#333
    classDef constNode fill:#f9f,stroke:#333,stroke-width:2px,color:#333
    classDef opNode fill:#bbf,stroke:#333,stroke-width:2px,color:#333

    I_X(["X FLOAT(a, b)"])

    Identity_0[["Identity(.)"]]

    I_X -->|"FLOAT(a, b)"| Identity_0

    O_Y(["Y FLOAT(a, b)"])
    Identity_0 --> O_Y

    class I_X,O_Y ioNode
    class Identity_0 opNode

apply(g: GraphBuilder, split_node: NodeProto, concat_node: NodeProto) → List[NodeProto][source]#

The method does the rewriting. It assumes it can happen. It takes a list of nodes impacted by the rewriting assumes no other pattern optimizer will be modify them. It receives the list of nodes returned by method apply. Since it is a list of argument, method match can include None values. The method returns the new nodes. The optimizer considers that any node given to this function is removed from the graph, and any node returned by it are added. If a received node must be kept, it must be added to the list of returned node.

Parameters:: nodes – nodes returned by method match, there are then removed
Returns:: nodes to add to graph.

match(g: GraphBuilderPatternOptimization, node: NodeProto, matched: List[MatchResult]) → MatchResult | None[source]#

Determines nodes around node which can be rewritten.

Parameters:

g – is a GraphBuilderPatternOptimization, it holds all the existing nodes, is able to return any information about type, shape, the node before, the node after another one.
node – the matching must determine if some nodes around this one are part of set of nodes this pattern optimizer can rewrite. From there, the function explores wherever it needs, checking any condition it needs.
matched – usually unused, it returns of nodes already matching a pattern

The method must not modify the graph. The method returns None if no match is found or an instance of class MatchResult. It must contain:

a list of nodes involved in the rewriting. It does not mean all of them will be removed but all of them are needed to do the rewriting and must not be impacted by other pattern optimizer.
A function doing the rewriting (usually method apply of the pattern class).
An existing node where the rewritten nodes can be inserted. Knowing it makes it faster to rewriter. If not specified, the optimizer will automatically determine the position of the new nodes.

yobx.xoptim.patterns.onnx_split#

This Page