dmx.compressor.fx.transformer.configuration_transformer

Classes

ConfigurationTransformer(module, scopeDict)

A transformer that changes the format of the ops according to the cfg file
class dmx.compressor.fx.transformer.configuration_transformer.ConfigurationTransformer(module: GraphModule, scopeDict: dict, cfg=None)

Bases: Transformer

A transformer that changes the format of the ops according to the cfg file

Parameters:

  • module (fx.GraphModule) – module to be added with additional ops.

  • scopeDict (Optional[dict]) – Dictionary that maps node name to scope. Defaults to None.

  • cfg (Optional[str]) – config file for setting the added ops formats. Defaults to None.

module

module to be added with additional ops.

Type:

fx.GraphModule

scopeDict

Dictionary that maps node name to scope.

Type:

Optional[str]

cfg

config file for setting the added ops formats.

Type:

Optional[str]

call_function(target: Target, args: Tuple[tuple[Argument, ...] | Sequence[Argument] | Mapping[str, Argument] | slice | range | Node | str | int | float | bool | complex | dtype | Tensor | device | memory_format | layout | OpOverload | SymInt | SymBool | SymFloat | None, ...], kwargs: Dict[str, Any]) → Any

Execute a call_function node. Modifies the format of cast according to cfg

Parameters:

  • target (Target) – The call target for this node. See Node for details on semantics

  • args (Tuple) – Tuple of positional args for this invocation

  • kwargs (Dict) – Dict of keyword arguments for this invocation

Returns:

A Proxy containing the node and the tracer of the new graph

get_attr(target: Target, args: Tuple[tuple[Argument, ...] | Sequence[Argument] | Mapping[str, Argument] | slice | range | Node | str | int | float | bool | complex | dtype | Tensor | device | memory_format | layout | OpOverload | SymInt | SymBool | SymFloat | None, ...], kwargs: Dict[str, Any]) → Proxy

Execute a get_attr node. Modifies the format of cast/sparsifier according to cfg

Parameters:

  • target (Target) – The call target for this node. See Node for details on semantics

  • args (Tuple) – Tuple of positional args for this invocation

  • kwargs (Dict) – Dict of keyword arguments for this invocation

Returns:

A Proxy containing the node and the tracer of the new graph

has_module(target: str) → bool

A function that checks if self.module contains a specific target

Parameters:

target (str) – the target to look up

Returns:

True if self.module contains a node with the same target

output(target: Target, args: Tuple[tuple[Argument, ...] | Sequence[Argument] | Mapping[str, Argument] | slice | range | Node | str | int | float | bool | complex | dtype | Tensor | device | memory_format | layout | OpOverload | SymInt | SymBool | SymFloat | None, ...], kwargs: Dict[str, Any]) → Any

Execute a output node. Modifies the format of output_cast according to cfg

Parameters:

  • target (Target) – The call target for this node. See Node for details on semantics

  • args (Tuple) – Tuple of positional args for this invocation

  • kwargs (Dict) – Dict of keyword arguments for this invocation

Returns:

A Proxy containing the node and the tracer of the new graph

placeholder(target: Target, args: Tuple[tuple[Argument, ...] | Sequence[Argument] | Mapping[str, Argument] | slice | range | Node | str | int | float | bool | complex | dtype | Tensor | device | memory_format | layout | OpOverload | SymInt | SymBool | SymFloat | None, ...], kwargs: Dict[str, Any]) → Proxy

Execute a placeholder node. Modifies the format of input_cast according to cfg

Parameters:

  • target (Target) – The call target for this node. See Node for details on semantics

  • args (Tuple) – Tuple of positional args for this invocation

  • kwargs (Dict) – Dict of keyword arguments for this invocation

Returns:

A Proxy containing the node and the tracer of the new graph

transform()

Transform self.module and return the transformed``GraphModule``.

Returns:

A GraphModule with updated ops formats.