2. PopART Python API¶

2.1. Sessions¶

class popart.session.InferenceSession(fnModel, dataFlow, deviceInfo, inputShapeInfo=<popart_core.InputShapeInfo object>, patterns=None, userOptions=<popart_core.SessionOptions object>)¶

Bases: popart_core._InferenceSessionCore

Create a runtime class for executing an ONNX graph on a set of IPU hardware for inference.

A wrapper around the Session C++ class, renamed SessionCore in pybind, to enable more Pythonic use. See session.hpp for parameter descriptions.

Parameters

fnModel – ONNX model proto. Usually a loaded ONNX model, or from builder.getModelProto().
dataFlow – Configuration for the data feeds and fetches.
deviceInfo – DeviceInfo object specifying device type. (one of IPU, IPUModel or CPU) and count.
inputShapeInfo – Information about the shapes of input and output tensors. Default: popart.InputShapeInfo().
patterns – Patterns to be run for optimization etc. Note: default for patterns must not be popart.Patterns(). When import popart is run, the default arguments are created. If the user then loads a custom pattern using ctypes.cdll.LoadLibrary(custom_pattern_lib.so) then the already constructed popart.Patterns will not include the custom pattern. Default None.
userOptions – Session options to apply. Default: popart.SessionOptions().

compileAndExport(filename)¶

Compiles the graph and exports it to the specified file.

This will form the snap::Graph and compile the polar::Executable before exporting the executable and metadata.

Parameters

filename – Where to save the executable and metadata. If
does not exist (it) –
will be created. (it) –

Raises

popart.OutOfMemoryException – If an out of memory event occurs
OSError – Thrown in the event of any file system related errors during the export

Return type

None

initAnchorArrays()¶

Create the anchor arrays to feed data back into Python with.

Returns: Dict of anchor names and their relevant np arrays.
Return type: Dict[str, numpy.array]

prepareDevice(loadEngine=True)¶

Prepare the network for execution.

This will create the snap::Graph and poplar::Engine, and set up poplar::Streams.

Parameters: loadEngine (bool) – Load the engine and connect the streams once the device is ready.
Raises: popart.OutOfMemoryException – If an out of memory event occurs
Return type: None

exception popart.session.OutOfMemoryException(e)¶

Bases: popart_core.popart_exception

Parameters: e (popart_core.popart_exception) –
Return type: None

getProfilePath()¶

Get the absolute path of the profile file (profile.pop)

Returns: The absolute path of profile.pop, or an empty string if not created.
Return type: str

getSummaryReport()¶

Get the summary report

Returns: The summary report string.
Return type: str

class popart.session.TrainingSession(fnModel, dataFlow, loss, optimizer, deviceInfo, inputShapeInfo=<popart_core.InputShapeInfo object>, patterns=None, userOptions=<popart_core.SessionOptions object>)¶

Bases: popart_core._TrainingSessionCore

Create a runtime class for executing an ONNX graph on a set of IPU hardware for training

A wrapper around the Session C++ class, renamed SessionCore in pybind, to enable more Pythonic use. See session.hpp for parameter descriptions.

Parameters

fnModel – ONNX model proto. Usually a loaded ONNX model, or from builder.getModelProto().
dataFlow – Configuration for the data feeds and fetches.
loss – A TensorId of the final scalar loss to use when training.
optimizer – The type of optimizer to use when training and it’s properties.
deviceInfo – DeviceInfo object specifying device type (IPU, IPUModel, CPU) and count.
inputShapeInfo – Information about the shapes of input and output tensors. Default: popart.InputShapeInfo().
patterns – Optimization patterns to apply. Default: None.
userOptions – Session options to apply. Default: popart.SessionOptions().

compileAndExport(filename)¶

Compiles the graph and exports it to the specified file.

This will form the snap::Graph and compile the polar::Executable before exporting the executable and metadata.

Parameters

filename – Where to save the executable and metadata. If
does not exist (it) –
will be created. (it) –

Raises

popart.OutOfMemoryException – If an out of memory event occurs
OSError – Thrown in the event of any file system related errors during the export

Return type

None

initAnchorArrays()¶

Create the anchor arrays to feed data back into Python with.

Returns: Dict of anchor names and their relevant np arrays.
Return type: Dict[str, numpy.array]

prepareDevice(loadEngine=True)¶

Prepare the network for execution.

This will create the snap::Graph and poplar::Engine, and set up poplar::Streams.

Parameters: loadEngine (bool) – Load the engine and connect the streams once the device is ready.
Raises: popart.OutOfMemoryException – If an out of memory event occurs
Return type: None

popart.session.makedirsAndCheckWritable(path)¶

2.2. Builder¶

class popart.builder.AiGraphcore(builder, version)¶

Bases: popart.builder.Opset

Return the builder interface for the given ai.graphcore version.

Raises: ValueError – Thrown if an invalid ai.graphcore opset version provided.

call(args, num_outputs, callee, debugName='')¶

Add a call operation to the model

This is a poplar extension, to expose manual code re-use to the builder

Parameters

args (List[int]) – List of tensor ids to feed as arguments.
num_outputs (int) – Number of output tensors from the called graph.
callee (popart.builder.Builder) – SubgraphBuilder for the graph to be called.
debugName (str) –

Keyword Arguments

debugName – A string to prepend to the name of the tensor. Default: “”.

Returns

Output tensor ids.

Return type

List[str]

class popart.builder.AiGraphcoreOpset1(builder, version)¶

Bases: popart.builder.AiGraphcore

Sub-class for backwards compatibility. Will forward all calls to AiGraphcore class.

class popart.builder.AiOnnx(builder, version)¶

Bases: popart.builder.Opset

Base class for the various AiOnnx builder interfaces. The most recent version of ONNX operators that require special treatment such as Loop, Scan, Logical_If etc. go here. While, older versions where the function signature differs are implemented on a corresponding subclass.

Parameters

builder – Parent class for access.
version – ai.Onnx opset version to use; 6 <= version <= 10. Default: 10.

logical_if(args, num_outputs, else_branch, then_branch, name='')¶

If conditional operation.

Parameters

args (List[str]) – List of tensor ids to feed as arguments.
num_outputs (int) – Number of output tensors from the if operator.
else_branch (popart.builder.Builder) – SubgraphBuilder for the graph to run if condition is false. Has num_outputs outputs: values you wish to live-out to the subgraph created by the if operation, other tensors will not be accessible to the wider graph. The number of outputs must match the number of outputs in the then_branch.
then_branch (popart.builder.Builder) – SubgraphBuilder for the graph to run if condition is true. Has num_outputs outputs: values you wish to be live-out to the enclosing scope. The number of outputs must match the number of outputs in the else_branch.
name (str) –

Keyword Arguments

name – A string to prepend to the name of the tensor. Default: “”.

Returns

Output tensor ids.

Return type

List[str]

loop(args, num_outputs, body, debugContext='')¶

Generic Looping construct op.

Parameters

args (List[str]) – List of tensor ids to feed as arguments.
num_outputs (int) – Number of output tensors from the loop operator.
body (popart.builder.Builder) – SubgraphBuilder for the graph to run in the loop.
debugContext (str) –

Keyword Arguments

debugContext – A string to prepend to the name of the tensor. Default: “”.

Returns

Output tensor ids.

Return type

List[str]

class popart.builder.AiOnnx10(builder, version)¶

Bases: popart.builder.AiOnnx9

Minimal builder interface for ai.onnx version 10. Once ai.onnx version 11 becomes the standard opset, this class must be updated to inherit from AiOnnx11, as described in T12084

class popart.builder.AiOnnx11(builder, version)¶

Bases: popart.builder.AiOnnx10

Minimal builder interface for ai.onnx version 11.

class popart.builder.AiOnnx6(builder, version)¶

Bases: popart.builder.AiOnnx

Minimal builder interface for ai.onnx version 6.

class popart.builder.AiOnnx7(builder, version)¶

Bases: popart.builder.AiOnnx6

Minimal builder interface for ai.onnx version 7.

class popart.builder.AiOnnx8(builder, version)¶

Bases: popart.builder.AiOnnx7

Minimal builder interface for ai.onnx version 8.

scan(args, num_outputs, body, num_scan_inputs, directions=[], debugContext='')¶

Scan-8 specific construct op.

Parameters

args (List[str]) – List of tensor ids to feed as arguments.
num_outputs (int) – Number of output tensors from the scan operator.
body (popart.builder.Builder) – SubgraphBuilder for the graph to run in the scan.
num_scan_inputs (int) – The number of scan_inputs
directions (List[int]) – A list of int which specifies the direction
the scan_input. 0 indicates forward direction and 1 (of) –
reverse direction. If not omitted (indicates) –
tensors (scan_input) –
be scanned in the forward direction. (will) –
debugContext (str) –

Keyword Arguments

debugContext – A string to prepend to the name of the tensor. Default: “”.

Returns

Output tensor ids.

Return type

List[str]

class popart.builder.AiOnnx9(builder, version)¶

Bases: popart.builder.AiOnnx8

Minimal builder interface for ai.onnx version 9.

scan(args, num_outputs, body, num_scan_inputs, scan_input_axes=[], scan_input_directions=[], scan_output_axes=[], scan_output_directions=[], debugContext='')¶

Generic Scan construct op.

Parameters

args (List[str]) – List of tensor ids to feed as arguments.
num_outputs (int) – Number of output tensors from the scan operator.
body (popart.builder.Builder) – SubgraphBuilder for the graph to run in the scan.
num_scan_inputs (int) – The number of scan_inputs
scan_input_axes (List[int]) – A list that specifies the axis to be scanned for the scan_input. If omitted, 0 will be used as the scan axis for every scan_input.
scan_input_directions (List[int]) – A list that specifies the direction to be scanned for the scan_input tensor. 0 indicates forward direction and 1 indicates reverse direction. If omitted, all scan_input tensors will be scanned in the forward direction.
scan_output_axes (List[int]) – A list that specifies the axis for the scan_output. The scan outputs are accumulated along the specified axis. If omitted, 0 will be used as the scan axis for every scan_output.
scan_output_directions (List[int]) – A list specifies whether the scan_output should be constructed by appending or prepending a new value in each iteration: 0 indicates appending and 1 indicates prepending. If omitted, all scan_output tensors will be produced by appending a value in each iteration.
debugContext (str) –

Keyword Arguments

debugContext – A string to prepend to the name of the tensor. Default: “”.

Returns

Output tensor ids.

Return type

List[str]

class popart.builder.AiOnnxMl(builder, version)¶

Bases: popart.builder.Opset

Return the builder interface for the given ai.onnx.ml version.

Raises: ValueError – Thrown if an invalid ai.onnx.ml opset version provided.

class popart.builder.Builder(modelProtoOrFilename=None, opsets=None, builderCore=None)¶

Bases: object

A wrapper around the Builder C++ class, renamed BuilderCore in pybind, to enable more Pythonic use. See builder.hpp for the class definition.

Parameters

modelProtoOrFilename – Model protobuf string or file path of saved ONNX model proto. Default: None.
opsets – Dict of opset versions. Default: None.
builderCore – _BuilderCore object if you want to create a subgraph builder using an existing buildercore object. Default: None.

aiOnnxOpsetVersion(version)¶

Parameters: version (int) –
Return type: None

createSubgraphBuilder()¶

Create a child builder to add ops to a subgraph using a call operation.

Returns: The child builder.
Return type: popart.builder.Builder

reshape_const(aiOnnx, args, shape, debugContext='')¶

Const version of the reshape op.

Parameters

aiOnnx (popart.builder.Opset) – Versioned aiOnnx opset, for example: aiOnnxOpset11.
args (List[str]) – List of tensor ids to feed as arguments.
shape (Iterable[int]) – Shape to reshape to, for example [3, 2, 4].
debugContext (str) –

Keyword Arguments

debugContext – String to use as a debug Context. Default: “”.

Returns

Output tensor ids.

Return type

List[int]

class popart.builder.Opset(builder, version)¶

Bases: object

Minimal base class for the opsets

Parameters

builder – An interface for a Builder, used for creating ONNX graphs.
version – Opset version to use for the given opset sub-class.

2.3. Tensor information¶

class popart.tensorinfo.TensorInfo(*args)¶

Bases: popart_core._TensorInfoCore

Python wrapper to TensorInfo to handle numpy types in constructor.

For example:: TensorInfo(dtype, shape) TensorInfo(numpy.ndarray)

Raises: TypeError – Raised if incorrect type is used to create a tensorinfo.

2.4. Writer¶

Framework independent functionality for driving PopART

class popart.writer.NetWriter(inNames, outNames, optimizer, dataFlow, inputShapeInfo)¶

Bases: object

Base class, to be inherited once per framework

Parameters

inNames – A list (in order) of all the inputs to the ONNX Model.
outNames – names of the outputs of the ONNX Model.
optimizer – An optimizer (ConstSGD, SGD, etc) or None if in inference mode.
anchors – Only relevant if in training mode: the names of tensors which must be computed and returned. If not in training mode, then outputs of forward are the (only) tensors to return.
dataFlow – Configuration for the data feeds and fetches.
inputShapeInfo – For every loss stream input and standard input: the shape, ONNX DataType and how to get data.

infer(inputsMap)¶: Perform batchesPerStep inference steps. This function only needs to be implemented by frameworks which will be used to verify PopART. See torchwriter.py for an example implementation.

saveModel(filename)¶: To be implemented once per framework: framework specific details of generating the ONNX model and writing it to file

train(inputsMap)¶: Perform batchesPerStep training steps. This function only needs to be implemented by frameworks which will be used to verify PopART. See torchwriter.py for an example implementation.

2.5. Builder¶

class popart_core._BuilderCore¶

addInitializedInputTensor(*args, **kwargs)¶

Overloaded function.

addInitializedInputTensor(self: popart_core._BuilderCore, initVal: numpy.ndarray, debugPrefix: popart_core.DebugContext = ‘’) -> str
addInitializedInputTensor(self: popart_core._BuilderCore, initVal: numpy.ndarray, debugContext: popart_core.DebugContext = ‘’) -> str

addInputTensor(*args, **kwargs)¶

Overloaded function.

addInputTensor(self: popart_core._BuilderCore, tensorInfo: popart_core._TensorInfoCore, debugPrefix: popart_core.DebugContext = ‘’) -> str
addInputTensor(self: popart_core._BuilderCore, tensorInfo: popart_core._TensorInfoCore, debugContext: popart_core.DebugContext = ‘’) -> str
addInputTensor(self: popart_core._BuilderCore, dataType: str, shape: List[int], debugPrefix: popart_core.DebugContext = ‘’) -> str
addInputTensor(self: popart_core._BuilderCore, dataType: str, shape: List[int], debugContext: popart_core.DebugContext = ‘’) -> str

addInputTensorFromParentGraph(self: popart_core._BuilderCore, tensorId: str) → None ¶

Add a new named input tensor to the model.

Parameter tensorId:: The identifier string of the input tensor. This identifier must already exist in the parent GraphProto’s name scope and must appear topologically before this sub-graph.

addNodeAttribute(*args, **kwargs)¶

Overloaded function.

addNodeAttribute(self: popart_core._BuilderCore, attributeName: str, attributeValue: int, nodeOutputNames: Set[str]) -> None
addNodeAttribute(self: popart_core._BuilderCore, attributeName: str, attributeValue: List[int], nodeOutputNames: Set[str]) -> None
addNodeAttribute(self: popart_core._BuilderCore, attributeName: str, attributeValue: float, nodeOutputNames: Set[str]) -> None
addNodeAttribute(self: popart_core._BuilderCore, attributeName: str, attributeValue: List[float], nodeOutputNames: Set[str]) -> None
addNodeAttribute(self: popart_core._BuilderCore, attributeName: str, attributeValue: str, nodeOutputNames: Set[str]) -> None
addNodeAttribute(self: popart_core._BuilderCore, attributeName: str, attributeValue: List[str], nodeOutputNames: Set[str]) -> None

addOutputTensor(self: popart_core._BuilderCore, outputName: str) → None ¶

addUntypedInputTensor(*args, **kwargs)¶

Overloaded function.

addUntypedInputTensor(self: popart_core._BuilderCore, debugPrefix: popart_core.DebugContext = ‘’) -> str

Add a new input tensor without a type or shape to the model.

Parameter debugContext:: Optional debug information.

Returns: The unique name of the input tensor.

addUntypedInputTensor(self: popart_core._BuilderCore, debugContext: popart_core.DebugContext = ‘’) -> str

Add a new input tensor without a type or shape to the model.

Parameter debugContext:: Optional debug information.

Returns: The unique name of the input tensor.

checkpointOutput(self: popart_core._BuilderCore, nodeOutputNames: List[str]) → List[str]¶

commGroup(self: popart_core._BuilderCore, type: int = 0, groupSize: int = 0) → AttributeContextManager¶

customOp(self: popart_core._BuilderCore, opName: str, opVersion: int, domain: str, inputs: list, attributes: dict, numOutputs: int = 1, name: str = '') → List[str]¶

excludePatterns(self: popart_core._BuilderCore, nodeOutputName: str, patternNames: List[str]) → None ¶

executionPhase(*args, **kwargs)¶

Overloaded function.

executionPhase(self: popart_core._BuilderCore, nodeOutputNames: str, value: int = 0) -> None
executionPhase(self: popart_core._BuilderCore, nodeOutputNames: Set[str], value: int = 0) -> None
executionPhase(self: popart_core._BuilderCore, value: int = 0) -> AttributeContextManager

getAllNodeAttributeNames(self: popart_core._BuilderCore, nodeOutputNames: Set[str]) → List[str]¶

Get all the attribute names from the ONNX node. This functions will throw an exception if it can’t find the unique node.

Parameter nodeOutputNames:: Names of the output tensors of the ONNX node used to find the node in the ONNX model.

getExecutionPhase(self: popart_core._BuilderCore) → int ¶

getFloatNodeAttribute(self: popart_core._BuilderCore, attributeName: str, nodeOutputNames: Set[str]) → float ¶

Get the float value of the attribute for the ONNX node. This functions will throw an exception if it can’t find the unique node or the attribute does not exist or it has not been set to the float type.

Parameter attributeName:: The name of the attribute to find.
Parameter nodeOutputNames:: Names of the output tensors of the ONNX node used to find the node in the ONNX model.

Returns: Value of the attribute.

getFloatVectorNodeAttribute(self: popart_core._BuilderCore, attributeName: str, nodeOutputNames: Set[str]) → List[float]¶

Get the ``std::vector``<float> value of the attribute for the ONNX node. This functions will throw an exception if it can’t find the unique node or the attribute does not exist.

Parameter attributeName:: The name of the attribute to find.
Parameter nodeOutputNames:: Names of the output tensors of the ONNX node used to find the node in the ONNX model.

Returns: Value of the attribute.

getInputTensorIds(self: popart_core._BuilderCore) → List[str]¶

getInt64NodeAttribute(self: popart_core._BuilderCore, attributeName: str, nodeOutputNames: Set[str]) → int ¶

Get the int64_t value of the attribute for the ONNX node. This functions will throw an exception if it can’t find the unique node or the attribute does not exist or it has not been set to the int64_t type.

Parameter attributeName:: The name of the attribute to find.
Parameter nodeOutputNames:: Names of the output tensors of the ONNX node used to find the node in the ONNX model.

Returns: Value of the attribute.

getInt64VectorNodeAttribute(self: popart_core._BuilderCore, attributeName: str, nodeOutputNames: Set[str]) → List[int]¶

Get the ``std::vector``<int64_t> value of the attribute for the ONNX node. This functions will throw an exception if it can’t find the unique node or the attribute does not exist or it has not been set to the ``std::vector``<int64_t> type.

Parameter attributeName:: The name of the attribute to find.
Parameter nodeOutputNames:: Names of the output tensors of the ONNX node used to find the node in the ONNX model.

Returns: Value of the attribute.

getModelProto(self: popart_core._BuilderCore) → bytes ¶

getNameScope(self: popart_core._BuilderCore, name: str = '') → str ¶

getOutputTensorIds(self: popart_core._BuilderCore) → List[str]¶

getPartialsType(self: popart_core._BuilderCore, nodeOutputName: str) → str ¶

Get the partials type for the given node.

Parameter nodeOutputName:: Name of the output tensor of the ONNX node.

getPipelineStage(self: popart_core._BuilderCore) → int ¶

getRecomputeOutputInBackwardPass(*args, **kwargs)¶

Overloaded function.

getRecomputeOutputInBackwardPass(self: popart_core._BuilderCore, nodeOutputName: str) -> bool
getRecomputeOutputInBackwardPass(self: popart_core._BuilderCore, nodeOutputNames: Set[str]) -> bool

getStringNodeAttribute(self: popart_core._BuilderCore, attributeName: str, nodeOutputNames: Set[str]) → str ¶

Get the std::string value of the attribute for the ONNX node. This functions will throw an exception if it can’t find the unique node or the attribute does not exist or it has not been set to the std::string type.

Parameter attributeName:: The name of the attribute to find.
Parameter nodeOutputNames:: Names of the output tensors of the ONNX node used to find the node in the ONNX model.

Returns: Value of the attribute.

getStringVectorNodeAttribute(self: popart_core._BuilderCore, attributeName: str, nodeOutputNames: Set[str]) → List[str]¶

Get the ``std::vector``<std::string> value of the attribute for the ONNX node. This functions will throw an exception if it can’t find the unique node or the attribute does not exist.

Parameter attributeName:: The name of the attribute to find.
Parameter nodeOutputNames:: Names of the output tensors of the ONNX node used to find the node in the ONNX model.

Returns: Value of the attribute.

getTensorDtypeString(self: popart_core._BuilderCore, id: str) → str ¶

getTensorShape(self: popart_core._BuilderCore, id: str) → List[int]¶

Return an ONNX graph tensor shape, from either the input, output, or value_info lists in the GraphProto.

Parameter id:: Tensor id.

Returns: A vector of tensor dimensions.

getTrainableTensorIds(self: popart_core._BuilderCore) → List[str]¶

getValueTensorIds(self: popart_core._BuilderCore) → List[str]¶

getVirtualGraph(*args, **kwargs)¶

Overloaded function.

getVirtualGraph(self: popart_core._BuilderCore) -> int
getVirtualGraph(self: popart_core._BuilderCore, nodeOutputNames: str) -> int

hasExecutionPhase(self: popart_core._BuilderCore) → bool ¶

hasPipelineStage(self: popart_core._BuilderCore) → bool ¶

hasVirtualGraph(self: popart_core._BuilderCore) → bool ¶

isInitializer(self: popart_core._BuilderCore, id: str) → bool ¶

Returns true if the ONNX tensor is in the initializer list of the GraphProto.

Parameter id:: Tensor id.

Returns: A boolean.

nameScope(self: popart_core._BuilderCore, name: str) → NameContextManager¶

nodeHasAttribute(self: popart_core._BuilderCore, attributeName: str, nodeOutputNames: Set[str]) → bool ¶

Check whether the ONNX node has an attribute set. This functions will throw an exception if it can’t find the unique node.

Parameter attributeName:: The name of the attribute to find.
Parameter nodeOutputNames:: Names of the output tensors of the ONNX node used to find the node in the ONNX model.

outlineAttributes(self: popart_core._BuilderCore, arg0: dict) → KeyValueContextManager¶

outputTensorLocation(*args, **kwargs)¶

Overloaded function.

outputTensorLocation(self: popart_core._BuilderCore, nodeOutputNames: str, value: popart_core.TensorLocation = <popart_core.TensorLocation object at 0x7f571cedf7d8>) -> None
outputTensorLocation(self: popart_core._BuilderCore, value: popart_core.TensorLocation = <popart_core.TensorLocation object at 0x7f571cedf810>) -> AttributeContextManager

pipelineStage(*args, **kwargs)¶

Overloaded function.

pipelineStage(self: popart_core._BuilderCore, nodeOutputNames: str, value: int = 0) -> None
pipelineStage(self: popart_core._BuilderCore, value: int) -> AttributeContextManager

recomputeOutput(*args, **kwargs)¶

Overloaded function.

recomputeOutput(self: popart_core._BuilderCore, nodeOutputNames: str, value: popart_core.RecomputeType = <RecomputeType.Undefined: 0>) -> None
recomputeOutput(self: popart_core._BuilderCore, value: popart_core.RecomputeType = <RecomputeType.Undefined: 0>) -> AttributeContextManager

recomputeOutputInBackwardPass(*args, **kwargs)¶

Overloaded function.

recomputeOutputInBackwardPass(self: popart_core._BuilderCore, nodeOutputName: str, value: popart_core.RecomputeType = <RecomputeType.Recompute: 2>) -> None
recomputeOutputInBackwardPass(self: popart_core._BuilderCore, nodeOutputNames: Set[str], value: popart_core.RecomputeType = <RecomputeType.Recompute: 2>) -> None

removeNodeAttribute(self: popart_core._BuilderCore, attributeName: str, nodeOutputNames: Set[str]) → None ¶

Remove an attribute from the ONNX node. This functions will throw an exception if it can’t find the unique node or the attribute does not exist.

Parameter attributeName:: The name of the attribute to find.
Parameter nodeOutputNames:: Names of the output tensors of the ONNX node used to find the node in the ONNX model.

saveInitializersExternally(self: popart_core._BuilderCore, ids: List[str], filename: str) → None ¶

Save tensor data externally.

The model data cannot exceed 2GB - the maximum size of a Protobuf message. To avoid this, for large models ONNX tensor data can be saved separately.

Parameter ids:: The names of tensors whose data is to be saved externally.
Parameter fn:: The name of a file containing the binary tensor data. This can be an absolute or relative path. If a relative path, when the ONNX model is saved, external tensor data will be written to a path relative to your current working directory.

saveModelProto(self: popart_core._BuilderCore, filename: str) → None ¶

Save the builder’s ONNX ModelProto into the builder and validate it.

Parameter fn:: The name of a file containing an ONNX model protobuf.

schedulePriority(self: popart_core._BuilderCore, value: float) → AttributeContextManager¶

setAvailableMemoryProportion(self: popart_core._BuilderCore, nodeOutputName: str, availableMemoryProportion: float) → None ¶

Set the available memory for the given node. Used on the convolution op.

Parameter nodeOutputName:: Name of the output tensor of the ONNX node.
Parameter availableMemoryProportion:: The available memory proportion 0 < x <= 1.

setGraphName(*args, **kwargs)¶

Overloaded function.

setGraphName(self: popart_core._BuilderCore, name: str) -> None

Specifies a graph name.

Parameter name:: String to name the graph.

setGraphName(self: popart_core._BuilderCore, name: str) -> None

setInplacePreferences(self: popart_core._BuilderCore, nodeOutputName: str, prefs: Dict[str, float]) → None ¶

setPartialsType(self: popart_core._BuilderCore, nodeOutputName: str, partialsType: str) → None ¶

Set the partials type for the given node. Used on the convolution op.

Parameter nodeOutputName:: Name of the output tensor of the ONNX node.
Parameter partialsType:: The type for the partials. Can be either FLOAT or HALF.

setSerializeMatMul(self: popart_core._BuilderCore, nodeOutputName: Set[str], mode: str, factor: int = 0, keep_precision: bool = False) → None ¶

virtualGraph(*args, **kwargs)¶

Overloaded function.

virtualGraph(self: popart_core._BuilderCore, nodeOutputNames: str, value: int = 0) -> None
virtualGraph(self: popart_core._BuilderCore, value: int) -> AttributeContextManager

2.5.1. AiGraphcoreOpset1¶

class popart_core.AiGraphcoreOpset1¶

abort(*args, **kwargs)¶

Overloaded function.

abort(self: popart_core.AiGraphcoreOpset1, args: List[str] = [], debugPrefix: popart_core.DebugContext = ‘’) -> None

Add abort operation to the model.

The operation can be conditional or unconditional.

Parameter args:: Optional input tensor to test condition

abort(self: popart_core.AiGraphcoreOpset1, args: List[str] = [], debugContext: popart_core.DebugContext = ‘’) -> None

Add abort operation to the model.

The operation can be conditional or unconditional.

Parameter args:: Optional input tensor to test condition

atan2(*args, **kwargs)¶

Overloaded function.

atan2(self: popart_core.AiGraphcoreOpset1, args: List[str], debugPrefix: popart_core.DebugContext = ‘’) -> str

Add an atan2 operation to the model.

Returns the element-wise angle theta as a tensor, -pi < theta <= pi, such that for two input tensors x and y and given r != 0, x = r cos theta, and y = r sin theta, element-wise.

In the case of x > 0, theta = arctan(y/x).

Parameter args:: Vector of input tensor ids: [y, x].
Parameter name:: Optional identifier for operation.

Returns: The name of the result tensor containing element wise theta values.

atan2(self: popart_core.AiGraphcoreOpset1, args: List[str], debugContext: popart_core.DebugContext = ‘’) -> str

Add an atan2 operation to the model.

Returns the element-wise angle theta as a tensor, -pi < theta <= pi, such that for two input tensors x and y and given r != 0, x = r cos theta, and y = r sin theta, element-wise.

In the case of x > 0, theta = arctan(y/x).

Parameter args:: Vector of input tensor ids: [y, x].
Parameter name:: Optional identifier for operation.

Returns: The name of the result tensor containing element wise theta values.

bitwiseand(self: popart_core.AiGraphcoreOpset1, args: List[str] = [], debugContext: popart_core.DebugContext = '') → str ¶

Add a bitwise AND operation to the model.

The operation computes the bitwise AND of given two integer tensors.

Parameter args:: Two broadcastable input tensors of type integer.

Returns: The name of the result tensor.

bitwisenot(self: popart_core.AiGraphcoreOpset1, args: List[str] = [], debugContext: popart_core.DebugContext = '') → str ¶

Add a bitwise NOT operation to the model.

The operation computes the bitwise NOT of a given integer tensor.

Parameter args:: Input tensor of type integer.

Returns: The name of the result tensor.

bitwiseor(self: popart_core.AiGraphcoreOpset1, args: List[str] = [], debugContext: popart_core.DebugContext = '') → str ¶

Add a bitwise OR operation to the model.

The operation computes the bitwise OR of given two integer tensors.

Parameter args:: Two broadcastable input tensors of type integer.

Returns: The name of the result tensor.

bitwisexnor(self: popart_core.AiGraphcoreOpset1, args: List[str] = [], debugContext: popart_core.DebugContext = '') → str ¶

Add a bitwise XNOR operation to the model.

The operation computes the bitwise XNOR of given two integer tensors.

Parameter args:: Two broadcastable input tensors of type integer.

Returns: The name of the result tensor.

bitwisexor(self: popart_core.AiGraphcoreOpset1, args: List[str] = [], debugContext: popart_core.DebugContext = '') → str ¶

Add a bitwise XOR operation to the model.

The operation computes the bitwise XOR of given two integer tensors.

Parameter args:: Two broadcastable input tensors of type integer.

Returns: The name of the result tensor.

call(*args, **kwargs)¶

Overloaded function.

call(self: popart_core.AiGraphcoreOpset1, args: List[str], num_outputs: int, callee: popart::Builder, debugPrefix: popart_core.DebugContext = ‘’) -> List[str]

Add a call operation to the model

This is a Poplar extension, to expose manual code re-use to the builder.

Parameter args:: Vector of input tensor ids.
Parameter callee:: The subgraph to call into.
Parameter debugContext:: Optional debug context.

Returns: A vector of tensors; the subgraph outputs.

call(self: popart_core.AiGraphcoreOpset1, args: List[str], num_outputs: int, callee: popart::Builder, debugContext: popart_core.DebugContext = ‘’) -> List[str]

Add a call operation to the model

This is a Poplar extension, to expose manual code re-use to the builder.

Parameter args:: Vector of input tensor ids.
Parameter callee:: The subgraph to call into.
Parameter debugContext:: Optional debug context.

Returns: A vector of tensors; the subgraph outputs.

ctcbeamsearchdecoder(self: popart_core.AiGraphcoreOpset1, args: List[str], blank: int = 0, beam_width: int = 100, top_paths: int = 1, debug_context: popart_core.DebugContext = '') → List[str]¶

ctcloss(self: popart_core.AiGraphcoreOpset1, args: List[str], reduction: popart_core.ReductionType = <ReductionType.Mean: 1>, blank: int = 0, outDataType: str = 'UNDEFINED', debugContext: popart_core.DebugContext = '') → str ¶

Add a connectionist temporal classification (CTC) loss operation to the model.

With T being maximum input length, N being batch size, C being number of classes, S being a maximum target length, this op calculates the CTC loss for a logarithmised probabilities tensor with shape [T, N, C], a class target tensor with shape [N, S], an input lengths tensor [N] and a target lengths tensor [N].

Note that C includes a blank class (default=0). The probabilities tensor is padded as required. Target sequences are also padded and are populated with values less than equal to C, not including the blank class, up to their respective target lengths. Note that target lengths cannot exceed input lengths.

Parameter args:: [log_probs,targets,input_lengths,target_lengths]
Parameter reduction:: Type of reduction to perform on the individual losses
Parameter blank:: The integer representing the blank class.
Parameter debugContext:: Optional debug context

Returns: The name of the result tensor

depthtospace(*args, **kwargs)¶

Overloaded function.

depthtospace(self: popart_core.AiGraphcoreOpset1, args: List[str], blocksize: int, mode: str, debugPrefix: popart_core.DebugContext = ‘’) -> str

Add the DepthToSpace to the model. (This allows DepthToSpace_11 to be targeted from earlier opsets.)

The purpose of Depth to Space, also known as pixel shuffling, is to rearrange data from the depth (channels) dimension into the spacial (width and height) dimensions. It is an efficient means of learning upsampling alongside mixing convolution with bilinear interpolation and using transpose convolution.

https://github.com/onnx/onnx/blob/master/docs/Operators.md#DepthToSpac e

Parameter args:: Vector containing single tensor input id.
Parameter blocksize:: Indicates the scale factor: if the input is [N, C, H, W] and the blocksize is B, the output will be [N, C/(B*B), H*B, W*B].
Parameter mode:: Specifies how the data is rearranged: * “DCR”: depth-column-row order * “CRD”: column-row-depth order
Parameter debugContext:: Optional debug context.

Returns: A tensor which is a rearrangement of the input tensor.

depthtospace(self: popart_core.AiGraphcoreOpset1, args: List[str], blocksize: int, mode: str, debugContext: popart_core.DebugContext = ‘’) -> str

Add the DepthToSpace to the model. (This allows DepthToSpace_11 to be targeted from earlier opsets.)

The purpose of Depth to Space, also known as pixel shuffling, is to rearrange data from the depth (channels) dimension into the spacial (width and height) dimensions. It is an efficient means of learning upsampling alongside mixing convolution with bilinear interpolation and using transpose convolution.

https://github.com/onnx/onnx/blob/master/docs/Operators.md#DepthToSpac e

Parameter args:: Vector containing single tensor input id.
Parameter blocksize:: Indicates the scale factor: if the input is [N, C, H, W] and the blocksize is B, the output will be [N, C/(B*B), H*B, W*B].
Parameter mode:: Specifies how the data is rearranged: * “DCR”: depth-column-row order * “CRD”: column-row-depth order
Parameter debugContext:: Optional debug context.

Returns: A tensor which is a rearrangement of the input tensor.

detach(*args, **kwargs)¶

Overloaded function.

detach(self: popart_core.AiGraphcoreOpset1, args: List[str], debugPrefix: popart_core.DebugContext = ‘’) -> str

Add a detach operation to the model.

Parameter args:: Vector of input tensor ids.
Parameter debugContext:: Optional debug context.

Returns: The name of the result tensor.

detach(self: popart_core.AiGraphcoreOpset1, args: List[str], debugContext: popart_core.DebugContext = ‘’) -> str

Add a detach operation to the model.

Parameter args:: Vector of input tensor ids.
Parameter debugContext:: Optional debug context.

Returns: The name of the result tensor.

dynamicadd(*args, **kwargs)¶

Overloaded function.

dynamicadd(self: popart_core.AiGraphcoreOpset1, args: List[str], axes: List[int], sizes: List[int], debugPrefix: popart_core.DebugContext = ‘’) -> str

Add a dynamic add operation to the model.

Creates a copy of tensor with slice added at offset. For example:

out = tensor, out[offset] += slice

Parameter args:: Vector of input tensor ids: [tensor, offset, slice].
Parameter axes:: Axes along which to add.
Parameter sizes:: Size of the slice in each axis.
Parameter debugContext:: Optional debug context.

Returns: The name of the result tensor.

dynamicadd(self: popart_core.AiGraphcoreOpset1, args: List[str], axes: List[int], sizes: List[int], debugContext: popart_core.DebugContext = ‘’) -> str

Add a dynamic add operation to the model.

Creates a copy of tensor with slice added at offset. For example:

out = tensor, out[offset] += slice

Parameter args:: Vector of input tensor ids: [tensor, offset, slice].
Parameter axes:: Axes along which to add.
Parameter sizes:: Size of the slice in each axis.
Parameter debugContext:: Optional debug context.

Returns: The name of the result tensor.

dynamicslice(*args, **kwargs)¶

Overloaded function.

dynamicslice(self: popart_core.AiGraphcoreOpset1, args: List[str], axes: List[int], sizes: List[int], noOverlap: int = 0, debugPrefix: popart_core.DebugContext = ‘’) -> str

Add a dynamic slice operation to the model.

Creates a new slice tensor. For example:

slice = tensor[offset]

Parameter args:: Vector of input tensor ids: [tensor, offset].
Parameter axes:: Axes along which to slice.
Parameter sizes:: Size of the slice in each axis.
Parameter debugContext:: Optional debug context.

Returns: The name of the result tensor.

dynamicslice(self: popart_core.AiGraphcoreOpset1, args: List[str], axes: List[int], sizes: List[int], noOverlap: int = 0, debugContext: popart_core.DebugContext = ‘’) -> str

Add a dynamic slice operation to the model.

Creates a new slice tensor. For example:

slice = tensor[offset]

Parameter args:: Vector of input tensor ids: [tensor, offset].
Parameter axes:: Axes along which to slice.
Parameter sizes:: Size of the slice in each axis.
Parameter debugContext:: Optional debug context.

Returns: The name of the result tensor.

dynamicupdate(*args, **kwargs)¶

Overloaded function.

dynamicupdate(self: popart_core.AiGraphcoreOpset1, args: List[str], axes: List[int], sizes: List[int], noOverlap: int = 0, debugPrefix: popart_core.DebugContext = ‘’) -> str

Add a dynamic update operation to the model.

Creates a copy of a tensor with a slice inserted at offset. For example:

out = tensor, out[offset] = slice

Parameter args:: Vector of input tensor ids: [tensor, offset, slice].
Parameter axes:: Axes along which to update.
Parameter sizes:: Size of the slice in each axis.
Parameter debugContext:: Optional debug context.

Returns: The name of the result tensor.

dynamicupdate(self: popart_core.AiGraphcoreOpset1, args: List[str], axes: List[int], sizes: List[int], noOverlap: int = 0, debugContext: popart_core.DebugContext = ‘’) -> str

Add a dynamic update operation to the model.

Creates a copy of a tensor with a slice inserted at offset. For example:

out = tensor, out[offset] = slice

Parameter args:: Vector of input tensor ids: [tensor, offset, slice].
Parameter axes:: Axes along which to update.
Parameter sizes:: Size of the slice in each axis.
Parameter debugContext:: Optional debug context.

Returns: The name of the result tensor.

dynamiczero(*args, **kwargs)¶

Overloaded function.

dynamiczero(self: popart_core.AiGraphcoreOpset1, args: List[str], axes: List[int], sizes: List[int], debugPrefix: popart_core.DebugContext = ‘’) -> str

Add a dynamic zero operation to the model.

Creates a copy of tensor with a slice at offset set to zero. For example:

out = tensor, out[offset] = 0.0

Parameter args:: Vector of input tensor ids [tensor, offset].
Parameter axes:: Axes along which to erase.
Parameter sizes:: Size of the slice in each axis.
Parameter debugContext:: Optional debug context.

Returns: The name of the result tensor.

dynamiczero(self: popart_core.AiGraphcoreOpset1, args: List[str], axes: List[int], sizes: List[int], debugContext: popart_core.DebugContext = ‘’) -> str

Add a dynamic zero operation to the model.

Creates a copy of tensor with a slice at offset set to zero. For example:

out = tensor, out[offset] = 0.0

Parameter args:: Vector of input tensor ids [tensor, offset].
Parameter axes:: Axes along which to erase.
Parameter sizes:: Size of the slice in each axis.
Parameter debugContext:: Optional debug context.

Returns: The name of the result tensor.

expm1(*args, **kwargs)¶

Overloaded function.

expm1(self: popart_core.AiGraphcoreOpset1, args: List[str], debugPrefix: popart_core.DebugContext = ‘’) -> str

Add expm1 operation to the model. It computes exp(x) - 1. Calculates the element-wise exponential of the input tensor and subtracts one.

Parameter args:: Vector of input tensor ids.
Parameter name:: Optional identifier for operation.

Returns: The name of the result tensor.

expm1(self: popart_core.AiGraphcoreOpset1, args: List[str], debugContext: popart_core.DebugContext = ‘’) -> str

Add expm1 operation to the model. It computes exp(x) - 1. Calculates the element-wise exponential of the input tensor and subtracts one.

Parameter args:: Vector of input tensor ids.
Parameter name:: Optional identifier for operation.

Returns: The name of the result tensor.

fmod(*args, **kwargs)¶

Overloaded function.

fmod(self: popart_core.AiGraphcoreOpset1, args: List[str], debugContext: popart_core.DebugContext = ‘’) -> str

Add fmod operation to the model.

This is equivalent to C’s fmod function. The result has the same sign as the dividend.

Parameter args:: Input tensors.

Returns: Computes the element-wise remainder of division. The remainder has the same sign as the dividend.

fmod(self: popart_core.AiGraphcoreOpset1, args: List[str], debugPrefix: popart_core.DebugContext = ‘’) -> str

Add fmod operation to the model.

This is equivalent to C’s fmod function. The result has the same sign as the dividend.

Parameter args:: Input tensors.

Returns: Computes the element-wise remainder of division. The remainder has the same sign as the dividend.

gelu(*args, **kwargs)¶

Overloaded function.

gelu(self: popart_core.AiGraphcoreOpset1, args: List[str], debugPrefix: popart_core.DebugContext = ‘’) -> str

Add a GELU operation to the model.

This is a Poplar extension.

Parameter args:: Vector of input tensor ids.
Parameter debugContext:: Optional debug context.

Returns: The name of the result tensor.

gelu(self: popart_core.AiGraphcoreOpset1, args: List[str], debugContext: popart_core.DebugContext = ‘’) -> str

Add a GELU operation to the model.

This is a Poplar extension.

Parameter args:: Vector of input tensor ids.
Parameter debugContext:: Optional debug context.

Returns: The name of the result tensor.

groupnormalization(*args, **kwargs)¶

Overloaded function.

groupnormalization(self: popart_core.AiGraphcoreOpset1, args: List[str], num_groups: int, epsilon: float = 9.999999747378752e-06, debugPrefix: popart_core.DebugContext = ‘’) -> List[str]

Add a group normalization operation to the model.

This is a Poplar extension.

The group will be created from a strided input.

Parameter args:: A vector of input tensors: [x, scale, bias].
Parameter num_groups:: The number of groups to separate the channels into.
Parameter epsilon:: The epsilon value to use to avoid division by zero.
Parameter debugContext:: Optional debug context.

Returns: [y, mean, var].
Return type: A vector of tensors

groupnormalization(self: popart_core.AiGraphcoreOpset1, args: List[str], num_groups: int, epsilon: float = 9.999999747378752e-06, debugContext: popart_core.DebugContext = ‘’) -> List[str]

Add a group normalization operation to the model.

This is a Poplar extension.

The group will be created from a strided input.

Parameter args:: A vector of input tensors: [x, scale, bias].
Parameter num_groups:: The number of groups to separate the channels into.
Parameter epsilon:: The epsilon value to use to avoid division by zero.
Parameter debugContext:: Optional debug context.

Returns: [y, mean, var].
Return type: A vector of tensors

identityloss(*args, **kwargs)¶

Overloaded function.

identityloss(self: popart_core.AiGraphcoreOpset1, args: List[str], reduction: popart_core.ReductionType = <ReductionType.Mean: 1>, debugPrefix: popart_core.DebugContext = ‘’) -> str

Add an identity loss operation to the model.

Calculates the loss using the identity operator.

Parameter args:: Vector of input tensor ids.
Parameter reduction:: Type of reduction to perform on the individual losses.
Parameter debugContext:: Optional debug context.

Returns: The name of the result tensor

identityloss(self: popart_core.AiGraphcoreOpset1, args: List[str], reduction: popart_core.ReductionType = <ReductionType.Mean: 1>, debugContext: popart_core.DebugContext = ‘’) -> str

Add an identity loss operation to the model.

Calculates the loss using the identity operator.

Parameter args:: Vector of input tensor ids.
Parameter reduction:: Type of reduction to perform on the individual losses.
Parameter debugContext:: Optional debug context.

Returns: The name of the result tensor

init(*args, **kwargs)¶

Overloaded function.

init(self: popart_core.AiGraphcoreOpset1, shape: List[int], data_type: int, init_type: int, batch_axis: int, debugPrefix: popart_core.DebugContext = ‘’) -> str

Add an init operation to the model.

Parameter shape:: Shape of the tensor to initialise.
Parameter data_type:: Data type to initialise tensor with.
Parameter init_type:: Mode of tensor initialisations.
Parameter batch_axis:: Axis relative to batch size.
Parameter debugContext:: Optional debug context.

Returns: The name of the result tensor.

init(self: popart_core.AiGraphcoreOpset1, shape: List[int], data_type: int, init_type: int, batch_axis: int, debugContext: popart_core.DebugContext = ‘’) -> str

Add an init operation to the model.

Parameter shape:: Shape of the tensor to initialise.
Parameter data_type:: Data type to initialise tensor with.
Parameter init_type:: Mode of tensor initialisations.
Parameter batch_axis:: Axis relative to batch size.
Parameter debugContext:: Optional debug context.

Returns: The name of the result tensor.

init(self: popart_core.AiGraphcoreOpset1, shape: List[int], data_type: int, init_type: int, debugPrefix: popart_core.DebugContext = ‘’) -> str

Add an init operation to the model.

Parameter shape:: Shape of the tensor to initialise.
Parameter data_type:: Data type to initialise tensor with.
Parameter init_type:: Mode of tensor initialisations.
Parameter debugContext:: Optional debug context.

Returns: The name of the result tensor.

init(self: popart_core.AiGraphcoreOpset1, shape: List[int], data_type: int, init_type: int, debugContext: popart_core.DebugContext = ‘’) -> str

Add an init operation to the model.

Parameter shape:: Shape of the tensor to initialise.
Parameter data_type:: Data type to initialise tensor with.
Parameter init_type:: Mode of tensor initialisations.
Parameter debugContext:: Optional debug context.

Returns: The name of the result tensor.

l1loss(*args, **kwargs)¶

Overloaded function.

l1loss(self: popart_core.AiGraphcoreOpset1, args: List[str], lambda: float, reduction: popart_core.ReductionType = <ReductionType.Mean: 1>, debugPrefix: popart_core.DebugContext = ‘’) -> str

Add an l1 loss operation to the model.

Calculates the mean absolute error between each element in the input with a zero target.

Parameter args:: Vector of input tensor ids.
Parameter lambda:: Scale factor of L1 loss.
Parameter reduction:: Type of reduction to perform on the individual losses.
Parameter debugContext:: Optional debug context.

Returns: The name of the result tensor.

l1loss(self: popart_core.AiGraphcoreOpset1, args: List[str], lambda: float, reduction: popart_core.ReductionType = <ReductionType.Mean: 1>, debugContext: popart_core.DebugContext = ‘’) -> str

Add an l1 loss operation to the model.

Calculates the mean absolute error between each element in the input with a zero target.

Parameter args:: Vector of input tensor ids.
Parameter lambda:: Scale factor of L1 loss.
Parameter reduction:: Type of reduction to perform on the individual losses.
Parameter debugContext:: Optional debug context.

Returns: The name of the result tensor.

log1p(*args, **kwargs)¶

Overloaded function.

log1p(self: popart_core.AiGraphcoreOpset1, args: List[str], debugPrefix: popart_core.DebugContext = ‘’) -> str

Add log1p operation to the model. It computes log(x + 1). This calculates the element-wise logarithm of the input tensor plus one.

Parameter args:: Vector of input tensor ids.
Parameter name:: Optional identifier for operation.

Returns: The name of the result tensor.

log1p(self: popart_core.AiGraphcoreOpset1, args: List[str], debugContext: popart_core.DebugContext = ‘’) -> str

Add log1p operation to the model. It computes log(x + 1). This calculates the element-wise logarithm of the input tensor plus one.

Parameter args:: Vector of input tensor ids.
Parameter name:: Optional identifier for operation.

Returns: The name of the result tensor.

lstm(*args, **kwargs)¶

Overloaded function.

lstm(self: popart_core.AiGraphcoreOpset1, args: List[str], outputFullSequence: int = 1, debugPrefix: popart_core.DebugContext = ‘’) -> List[str]
lstm(self: popart_core.AiGraphcoreOpset1, args: List[str], outputFullSequence: int = 1, debugContext: popart_core.DebugContext = ‘’) -> List[str]

multiconv(*args, **kwargs)¶

Overloaded function.

multiconv(self: popart_core.AiGraphcoreOpset1, args: List[List[str]], dilations: List[List[int]] = [], inDilations: List[List[int]] = [], pads: List[List[int]] = [], outPads: List[List[int]] = [], strides: List[List[int]] = [], availableMemoryProportions: List[float] = [], partialsTypes: List[str] = [], planType: Optional[str] = None, perConvReservedTiles: Optional[int] = None, cycleBackOff: Optional[float] = None, debugPrefix: popart_core.DebugContext = ‘’) -> List[str]

Add a multi-convolution to the model.

Using this multi-convolution API ensures that the convolutions are executed in parallel on the device.

Functionally, a multi-convolution is equivalent to a series of single convolutions. Using this multi-convolution API is always equivalent to calling the single-convolution API (conv) once for each argument.

For example, calling:

A0 = conv({X0, W0, B0}) A1 = conv({X1, W1})

Is functionally equivalent to calling:

{A0, A1} = multiconv({{X0, W0, B0}, {X1, Q1}).

It is possible that any two convolutions cannot be executed in parallel due to topological constraints. For example, the following:

B = conv({A, W0}); C = B + A D = conv({C, W1});

Cannot be converted to:

{B, D} = multiconv({{A, W0}, {C, W1}}).

Note that it is not possible to create such a cycle by adding a multi- convolution with this API.

Calls to multiconv() are mapped to poplar::poplin::multiconv::convolution().

Parameter tensors:: List of [DataId, WeightId, BiasId (optional)] for each convolution.
Parameter dilations:: The dilations attributes for each convolution.
Parameter inDilations:: The input dilations attributes for each convolution.
Parameter pads:: The pads for each convolution.
Parameter outPads:: The output padding for each convolution.
Parameter strides:: The strides for each convolution.
Parameter availableMemoryProportions:: The available memory proportions per conv, each [0, 1).
Parameter partialsTypes:: The partials type per convolution.
Parameter planType:: Run convolutions in parallel or series.
Parameter perConvReservedTiles:: Tiles to reserve per convolution when planning.
Parameter cycleBackOff:: Cycle back-off proportion, [0, 1).
Parameter debugContext:: Optional debug context.

All input vectors must be either empty, or equal in length to the number of convolutions. Note that groups for each convolution are automatically inferred from the shapes of the data and weight inputs.

Returns: The TensorId of the output tensor from each convolution.

multiconv(self: popart_core.AiGraphcoreOpset1, args: List[List[str]], dilations: List[List[int]] = [], inDilations: List[List[int]] = [], pads: List[List[int]] = [], outPads: List[List[int]] = [], strides: List[List[int]] = [], availableMemoryProportions: List[float] = [], partialsTypes: List[str] = [], planType: Optional[str] = None, perConvReservedTiles: Optional[int] = None, cycleBackOff: Optional[float] = None, debugContext: popart_core.DebugContext = ‘’) -> List[str]

Add a multi-convolution to the model.

Using this multi-convolution API ensures that the convolutions are executed in parallel on the device.

Functionally, a multi-convolution is equivalent to a series of single convolutions. Using this multi-convolution API is always equivalent to calling the single-convolution API (conv) once for each argument.

For example, calling:

A0 = conv({X0, W0, B0}) A1 = conv({X1, W1})

Is functionally equivalent to calling:

{A0, A1} = multiconv({{X0, W0, B0}, {X1, Q1}).

It is possible that any two convolutions cannot be executed in parallel due to topological constraints. For example, the following:

B = conv({A, W0}); C = B + A D = conv({C, W1});

Cannot be converted to:

{B, D} = multiconv({{A, W0}, {C, W1}}).

Note that it is not possible to create such a cycle by adding a multi- convolution with this API.

Calls to multiconv() are mapped to poplar::poplin::multiconv::convolution().

Parameter tensors:: List of [DataId, WeightId, BiasId (optional)] for each convolution.
Parameter dilations:: The dilations attributes for each convolution.
Parameter inDilations:: The input dilations attributes for each convolution.
Parameter pads:: The pads for each convolution.
Parameter outPads:: The output padding for each convolution.
Parameter strides:: The strides for each convolution.
Parameter availableMemoryProportions:: The available memory proportions per conv, each [0, 1).
Parameter partialsTypes:: The partials type per convolution.
Parameter planType:: Run convolutions in parallel or series.
Parameter perConvReservedTiles:: Tiles to reserve per convolution when planning.
Parameter cycleBackOff:: Cycle back-off proportion, [0, 1).
Parameter debugContext:: Optional debug context.

All input vectors must be either empty, or equal in length to the number of convolutions. Note that groups for each convolution are automatically inferred from the shapes of the data and weight inputs.

Returns: The TensorId of the output tensor from each convolution.

nllloss(*args, **kwargs)¶

Overloaded function.

nllloss(self: popart_core.AiGraphcoreOpset1, args: List[str], reduction: popart_core.ReductionType = <ReductionType.Mean: 1>, ignoreIndex: Optional[int] = None, inputIsLogProbability: bool = False, debugPrefix: popart_core.DebugContext = ‘’) -> str

Add a negative log-likelihood loss operation to the model.

Calculates the nll loss given a probability tensor over classes, and a target tensor containing class labels.

Parameter args:: Vector of input tensor ids: probability and tensor.
Parameter reduction:: Type of reduction to perform on the individual losses.
Parameter ignoreIndex:: Optional class index to ignore in loss calculation.
Parameter inputIsLogProbability:: Specifies if the input tensor contains log-probabilities or raw probabilities (false, default).
Parameter debugContext:: Optional debug context.

Returns: The name of the result tensor.

nllloss(self: popart_core.AiGraphcoreOpset1, args: List[str], reduction: popart_core.ReductionType = <ReductionType.Mean: 1>, ignoreIndex: Optional[int] = None, inputIsLogProbability: bool = False, debugContext: popart_core.DebugContext = ‘’) -> str

Add a negative log-likelihood loss operation to the model.

Calculates the nll loss given a probability tensor over classes, and a target tensor containing class labels.

Parameter args:: Vector of input tensor ids: probability and tensor.
Parameter reduction:: Type of reduction to perform on the individual losses.
Parameter ignoreIndex:: Optional class index to ignore in loss calculation.
Parameter inputIsLogProbability:: Specifies if the input tensor contains log-probabilities or raw probabilities (false, default).
Parameter debugContext:: Optional debug context.

Returns: The name of the result tensor.

nop(*args, **kwargs)¶

Overloaded function.

nop(self: popart_core.AiGraphcoreOpset1, args: List[str], debugPrefix: popart_core.DebugContext = ‘’) -> str

Add a no-op operation to the model.

Parameter args:: Vector of input tensor ids.
Parameter debugContext:: Optional debug context.

Returns: The name of the result tensor.

nop(self: popart_core.AiGraphcoreOpset1, args: List[str], debugContext: popart_core.DebugContext = ‘’) -> str

Add a no-op operation to the model.

Parameter args:: Vector of input tensor ids.
Parameter debugContext:: Optional debug context.

Returns: The name of the result tensor.

printtensor(*args, **kwargs)¶

Overloaded function.

printtensor(self: popart_core.AiGraphcoreOpset1, args: List[str], print_gradient: int = 1, debugPrefix: popart_core.DebugContext = ‘’, title: str = ‘’) -> str

Add a print tensor operation to the model.

This is a Poplar extension.

Parameter args:: Vector of tensor ids to print.
Parameter print_gradient:: $Parameter debugContext:

Optional debug context.

Parameter title:: $Returns:

The name of the result tensor.

printtensor(self: popart_core.AiGraphcoreOpset1, args: List[str], print_gradient: int = 1, debugContext: popart_core.DebugContext = ‘’, title: str = ‘’) -> str

Add a print tensor operation to the model.

This is a Poplar extension.

Parameter args:: Vector of tensor ids to print.
Parameter print_gradient:: $Parameter debugContext:

Optional debug context.

Parameter title:: $Returns:

The name of the result tensor.

reducemedian(self: popart_core.AiGraphcoreOpset1, args: List[str], axes: Optional[List[int]] = None, keepdims: int = 1, debugContext: popart_core.DebugContext = '') → List[str]¶

remainder(*args, **kwargs)¶

Overloaded function.

remainder(self: popart_core.AiGraphcoreOpset1, args: List[str], debugPrefix: popart_core.DebugContext = ‘’) -> str

Add remainder operation to the model.

This is equivalent to Python’s modulo operator %. The result has the same sign as the divisor.

Parameter args:: Input tensors.

Returns: Computes the element-wise remainder of division. The remainder has the same sign as the divisor.

remainder(self: popart_core.AiGraphcoreOpset1, args: List[str], debugContext: popart_core.DebugContext = ‘’) -> str

Add remainder operation to the model.

This is equivalent to Python’s modulo operator %. The result has the same sign as the divisor.

Parameter args:: Input tensors.

Returns: Computes the element-wise remainder of division. The remainder has the same sign as the divisor.

replicatedallreduce(*args, **kwargs)¶

Overloaded function.

replicatedallreduce(self: popart_core.AiGraphcoreOpset1, args: List[str], commGroup: Optional[List[int]] = None, debugPrefix: popart_core.DebugContext = ‘’) -> str

Add a replicated all-reduce operation to the model.

This is a Poplar extension, to expose manual code re-use to the builder.

Parameter args:: Vector of input tensor ids to reduce across.
Parameter commGroup:: GCL CommGroup parameter.
Parameter debugContext:: Optional debug context.

Returns: The name of the result tensor.

replicatedallreduce(self: popart_core.AiGraphcoreOpset1, args: List[str], commGroup: Optional[List[int]] = None, debugContext: popart_core.DebugContext = ‘’) -> str

Add a replicated all-reduce operation to the model.

This is a Poplar extension, to expose manual code re-use to the builder.

Parameter args:: Vector of input tensor ids to reduce across.
Parameter commGroup:: GCL CommGroup parameter.
Parameter debugContext:: Optional debug context.

Returns: The name of the result tensor.

reshape(*args, **kwargs)¶

Overloaded function.

reshape(self: popart_core.AiGraphcoreOpset1, args: str, shape: List[int], debugPrefix: popart_core.DebugContext = ‘’) -> str

Add reshape operation to the model. Reshape the input tensor. This reshape takes the shape to reshape into as an attribute instead of a tensor input as the ONNX reshape op.

Parameter arg:: Vector with single input tensor id.
Parameter shape:: The shape of the output Tensor. The output Tensor must contain the same number of elements as the input Tensor.
Parameter name:: Optional identifier for operation.

Returns: The name of the result tensor.

reshape(self: popart_core.AiGraphcoreOpset1, args: str, shape: List[int], debugContext: popart_core.DebugContext = ‘’) -> str

Add reshape operation to the model. Reshape the input tensor. This reshape takes the shape to reshape into as an attribute instead of a tensor input as the ONNX reshape op.

Parameter arg:: Vector with single input tensor id.
Parameter shape:: The shape of the output Tensor. The output Tensor must contain the same number of elements as the input Tensor.
Parameter name:: Optional identifier for operation.

Returns: The name of the result tensor.

reverse(self: popart_core.AiGraphcoreOpset1, args: List[str], dimensions: List[int], debugContext: popart_core.DebugContext = '') → str ¶

Add a reverse operator to the model.

Reverse, or ‘flip’, the tensor along the specified dimensions

Parameter args:: Input tensors.
Parameter dimensions:: Dimensions along which to reverse the tensor. If this is empty then this is equivalent to the identity operator

Returns: The name of the result tensor.

round(*args, **kwargs)¶

Overloaded function.

round(self: popart_core.AiGraphcoreOpset1, args: List[str], debugPrefix: popart_core.DebugContext = ‘’) -> str

Add a Round operation to the model. (This allows Round_11 to be targeted from earlier opsets.)

https://github.com/onnx/onnx/blob/master/docs/Operators.md#Round

Parameter args:: Vector of input tensor ids.
Parameter debugContext:: Optional debug context.

Returns: The normalized output tensor ids.

round(self: popart_core.AiGraphcoreOpset1, args: List[str], debugContext: popart_core.DebugContext = ‘’) -> str

Add a Round operation to the model. (This allows Round_11 to be targeted from earlier opsets.)

https://github.com/onnx/onnx/blob/master/docs/Operators.md#Round

Parameter args:: Vector of input tensor ids.
Parameter debugContext:: Optional debug context.

Returns: The normalized output tensor ids.

scale(*args, **kwargs)¶

Overloaded function.

scale(self: popart_core.AiGraphcoreOpset1, args: List[str], scale: float, debugPrefix: popart_core.DebugContext = ‘’) -> str

Add a scale operation to the model.

This is a Poplar extension.

Parameter args:: Vector of input tensor ids.
Parameter scale:: The scale to apply.
Parameter debugContext:: Optional debug context.

Returns: The name of the result tensor.

scale(self: popart_core.AiGraphcoreOpset1, args: List[str], scale: float, debugContext: popart_core.DebugContext = ‘’) -> str

Add a scale operation to the model.

This is a Poplar extension.

Parameter args:: Vector of input tensor ids.
Parameter scale:: The scale to apply.
Parameter debugContext:: Optional debug context.

Returns: The name of the result tensor.

scaledadd(*args, **kwargs)¶

Overloaded function.

scaledadd(self: popart_core.AiGraphcoreOpset1, args: List[str], scale0: float = 1.0, scale1: float = 1.0, debugPrefix: popart_core.DebugContext = ‘’) -> str

Add a scaled add operation to the model.

X = scale0 * T0 + scale1 * T1

Parameter args:: Vector of input tensor ids: [T0, T1, scale0, scale1].
Parameter scale0:: The scale to apply (if no scale0 tensor is supplied).
Parameter scale1:: The scale to apply (if no scale1 tensor is supplied).
Parameter debugContext:: Optional debug context.

Returns: The name of the result tensor.

scaledadd(self: popart_core.AiGraphcoreOpset1, args: List[str], scale0: float = 1.0, scale1: float = 1.0, debugContext: popart_core.DebugContext = ‘’) -> str

Add a scaled add operation to the model.

X = scale0 * T0 + scale1 * T1

Parameter args:: Vector of input tensor ids: [T0, T1, scale0, scale1].
Parameter scale0:: The scale to apply (if no scale0 tensor is supplied).
Parameter scale1:: The scale to apply (if no scale1 tensor is supplied).
Parameter debugContext:: Optional debug context.

Returns: The name of the result tensor.

scatterreduce(self: popart_core.AiGraphcoreOpset1, args: List[str] = [], axis_size: int, axis: int = -1, reduction: popart_core.ScatterReduction = <ScatterReduction.Sum: 0>, debugContext: popart_core.DebugContext = '') → str ¶

sequenceslice(self: popart_core.AiGraphcoreOpset1, args: List[str], zeroUnused: int = 0, debugContext: popart_core.DebugContext = '') → str ¶

Slice a 2D tensor based on offsets specified by a tensor.

The outermost dimension is sliced; tOut[tOutOffset:tOutOffset+tN][…] = tIn[tInOffset:tInOffset+tN][…] for each entry in tN/tInOffset/tOutOffset; entries after the first tN==0 may be ignored. Unreferenced elements of tOut are zeroed if zeroUnused is set. The same output element should not be written by multiple inputs.

tIn and tOut must have rank greater than or equal to 2. The outer dimension is sliced; the product of the inner dimensions must match. tInOffset, tOutOffset and tN must be 1d and the same size. param [source, destination, N, sourceOffset, destinationOffset]

Parameter zeroUnused:: Whether to zero unreferenced tOut elements.
Parameter debugContext:: Optional debug context.

shapeddropout(*args, **kwargs)¶

Overloaded function.

shapeddropout(self: popart_core.AiGraphcoreOpset1, args: List[str], shape: List[int], ratio: float = 0.5, debugPrefix: popart_core.DebugContext = ‘’) -> str

Add a shaped dropout operation to the model.

Applies a shaped dropout to the input tensor. This operator requires a shape parameter that is used to define the shape of the dropout mask so that strongly correlated features in the input tensor can be preserved. The provided shape must be broadcastable to the input tensor. Note that this operation targets the poprand library function of the same name.

Parameter args:: Vector of input tensor ids.
Parameter shape:: Shape of dropout mask. Must be broadcastable to the input.
Parameter ratio:: Probability of dropping an input feature (default = 0.5).
Parameter name:: Optional identifier for operation.

Returns: The name of the result tensor.

shapeddropout(self: popart_core.AiGraphcoreOpset1, args: List[str], shape: List[int], ratio: float = 0.5, debugContext: popart_core.DebugContext = ‘’) -> str

Add a shaped dropout operation to the model.

Applies a shaped dropout to the input tensor. This operator requires a shape parameter that is used to define the shape of the dropout mask so that strongly correlated features in the input tensor can be preserved. The provided shape must be broadcastable to the input tensor. Note that this operation targets the poprand library function of the same name.

Parameter args:: Vector of input tensor ids.
Parameter shape:: Shape of dropout mask. Must be broadcastable to the input.
Parameter ratio:: Probability of dropping an input feature (default = 0.5).
Parameter name:: Optional identifier for operation.

Returns: The name of the result tensor.

subsample(*args, **kwargs)¶

Overloaded function.

subsample(self: popart_core.AiGraphcoreOpset1, args: List[str], strides: List[int], debugPrefix: popart_core.DebugContext = ‘’) -> str

Add a sub-sample operation to the model.

This is a Poplar extension.

If multiple tensors are provided that strides will applied to them all.

Parameter args:: Vector of tensor ids to sub-sample.
Parameter strides:: The strides to use.
Parameter debugContext:: Optional debug context.

Returns: The name of the result tensor.

subsample(self: popart_core.AiGraphcoreOpset1, args: List[str], strides: List[int], debugContext: popart_core.DebugContext = ‘’) -> str

Add a sub-sample operation to the model.

This is a Poplar extension.

If multiple tensors are provided that strides will applied to them all.

Parameter args:: Vector of tensor ids to sub-sample.
Parameter strides:: The strides to use.
Parameter debugContext:: Optional debug context.

Returns: The name of the result tensor.

2.6. Session¶

class popart_core._InferenceSessionCore¶

compileAndExport(self: popart_core._InferenceSessionCore, filename: str, err: popart_core.OutOfMemoryError) → None ¶

getCycleCount(self: popart_core._InferenceSessionCore, id: str = '') → int ¶: Copy the cycle count tensor to host from the device.

getExecutionReport(self: popart_core._InferenceSessionCore, useCbor: bool = False, resetProfile: bool = True) → bytes ¶

Retrieve the execution report from the poplar::Engine.

The options which were given to the constructor will influence the information in the report. By default a JSON format report is produced.

This may only be called after the prepareDevice() call has been made.

Parameter useCbor:: Produce a CBOR formatted report.
Parameter resetProfile:: Resets the execution profile.

Returns: A string containing the execution report.

getGraphReport(self: popart_core._InferenceSessionCore, useCbor: bool = False) → bytes ¶

Retrieve the graph report from the poplar::Engine.

The options which were given to the constructor will influence the information in the report. By default a JSON format report is produced.

This may only be called after the prepareDevice() call has been made.

Parameter useCbor:: Produce a CBOR formatted report.

Returns: A string containing the graph (compilation) report.

getInfo(self: popart_core._InferenceSessionCore, arg0: str) → popart_core._TensorInfoCore¶: Get the TensorInfo on a Tensor.

getRNGState(self: popart_core._InferenceSessionCore) → List[int]¶

getReport(self: popart_core._InferenceSessionCore) → pva::Report¶

Retrieve the graph report from the poplar::Engine.

The options which were given to the constructor will influence the information in the report.

This may only be called after the prepareDevice() call has been made.

Returns: A the libpva report containing the graph report.

getSerializedGraph(self: popart_core._InferenceSessionCore) → bytes ¶

getSummaryReport(self: popart_core._InferenceSessionCore, resetProfile: bool = True) → str ¶

Retrieve the summary from from the poplar::Engine.

The options which were given to the constructor will influence the information in the report.

This may only be called after the prepareDevice() call has been made.

Parameter resetProfile:: Resets the execution profile.

Returns: A string containing the report.

loadEngineAndConnectStreams(self: popart_core._InferenceSessionCore) → None ¶

Load the engine on the device and connect the streams

This will set up the poplar::Streams.

Note: This call is optional. The engine will implicitly be loaded on the device when required.

loadExecutable(self: popart_core._InferenceSessionCore, filename: str) → None ¶

Load the poplar::Executable and the PopART metadata from the given file. The file must have been created with compileAndExport()

Parameter filename:: Name of the file to load the executable from.

modelToHost(self: popart_core._InferenceSessionCore, arg0: str) → None ¶

Write current model to ONNX file.

Parameter fn:: Path to file. Can be absolute or relative. If you plan to run your program in multiple processes simultaneously, you should avoid possible race conditions by writing to different files, for example by using temporary files.

prepareDevice(self: popart_core._InferenceSessionCore, loadEngine: bool = True, err: popart_core.OutOfMemoryError) → None ¶

Prepare the network for execution.

This will create the snap::Graph and poplar::Engine.

Parameter loadEngine:: Load the engine and connect the streams once the device is ready.

resetHostWeights(self: popart_core._InferenceSessionCore, modelProtoOrFilename: str, ignoreWeightsInModelWithoutCorrespondingHostWeight: bool = False) → None ¶

Reset the weights with the weights in an ONNX model that differs from the current model only in weights. This only updates the weights on the host; the user still needs to call weightsFromHost() after this to update the weights on the device.

Parameter model:: Either an ONNX model protobuf, or the name of a file containing an ONNX model protobuf.
Parameter ignoreWeightsInModelWithoutCorrespondingHostWeight:: If true, do not error if there are initializers in the ONNX model with no corresponding initializer tensor in the session’s IR.

run(self: popart_core._InferenceSessionCore, stepio: popart_core.IStepIO, debugName: str = '') → None ¶

Perform one step.

Read input data from address in stepIO.in. Write the output data to addresses in stepIO.out.

Parameter stepIO:: Input and output data.
Parameter debugName:: Debug string to identify this run in logs.

setRNGState(self: popart_core._InferenceSessionCore, rngValue: List[int]) → None ¶

setRandomSeed(self: popart_core._InferenceSessionCore, seedValue: int) → None ¶

Sets the random number generator seed on all tiles of the device. This ensures deterministic behaviour of random operations in the graph.

Parameter The:: seed value.

updateExternallySavedTensorLocations(self: popart_core._InferenceSessionCore, arg0: str, arg1: str) → None ¶

Update the tensor locations of the tensors in the Session’s ONNX model. The new file will be created at this point, and written to when the ONNX model is saved with a subsequent call to modelToHost.

Parameter fromLocation:: All externally saved tensors with location fromLocation will have their location updated to toLocation.
Parameter toLocation:: The updated location. Must not already exist.

weightsFromHost(self: popart_core._InferenceSessionCore) → None ¶: Write weights from host to the device.

writeWeights(self: popart_core._InferenceSessionCore, arg0: popart_core.IWeightsIO) → None ¶

Write the weights. Must call weightsFromHost() after this.

The weight data is written to the addresses in weightsIo.out.

class popart_core._TrainingSessionCore¶

compileAndExport(self: popart_core._TrainingSessionCore, filename: str, err: popart_core.OutOfMemoryError) → None ¶

connectStreamToCallback(self: popart_core._TrainingSessionCore, arg0: str, arg1: Callable[[capsule], None], arg2: int) → None ¶: Connect Poplar stream callbacks. In conjunction with getGradAndVarStreamIds the streams can be used to copy gradients to the host to perform collective operations after which the variables can be streamed back after they have been updated to the device. p index referes to the replica index when using replicated graphs.

getCycleCount(self: popart_core._TrainingSessionCore, id: str = '') → int ¶: Copy the cycle count tensor to host from the device.

getExecutionReport(self: popart_core._TrainingSessionCore, useCbor: bool = False, resetProfile: bool = True) → bytes ¶

Retrieve the execution report from the poplar::Engine.

The options which were given to the constructor will influence the information in the report. By default a JSON format report is produced.

This may only be called after the prepareDevice() call has been made.

Parameter useCbor:: Produce a CBOR formatted report.
Parameter resetProfile:: Resets the execution profile.

Returns: A string containing the execution report.

getGraphReport(self: popart_core._TrainingSessionCore, useCbor: bool = False) → bytes ¶

Retrieve the graph report from the poplar::Engine.

The options which were given to the constructor will influence the information in the report. By default a JSON format report is produced.

This may only be called after the prepareDevice() call has been made.

Parameter useCbor:: Produce a CBOR formatted report.

Returns: A string containing the graph (compilation) report.

getHostReduceStreamIds(self: popart_core._TrainingSessionCore) → List[str]¶: Access the stream IDs for variables that are involved in host side reductions on the host. Only populated if hostAllReduce is enabled in the SessionOptions

getInfo(self: popart_core._TrainingSessionCore, arg0: str) → popart_core._TensorInfoCore¶: Get the TensorInfo on a Tensor.

getIr(self: popart_core._TrainingSessionCore) → popart::Ir¶

getRNGState(self: popart_core._TrainingSessionCore) → List[int]¶

getReport(self: popart_core._TrainingSessionCore) → pva::Report¶

Retrieve the graph report from the poplar::Engine.

The options which were given to the constructor will influence the information in the report.

This may only be called after the prepareDevice() call has been made.

Returns: A the libpva report containing the graph report.

getSerializedGraph(self: popart_core._TrainingSessionCore) → bytes ¶

Retrieve the serialized graph from the poplar::Engine.

A JSON format report is produced.

This may only be called after the prepareDevice() call has been made.

Returns: A string containing the serialized graph.

getSummaryReport(self: popart_core._TrainingSessionCore, resetProfile: bool = True) → str ¶

Retrieve the summary from from the poplar::Engine.

The options which were given to the constructor will influence the information in the report.

This may only be called after the prepareDevice() call has been made.

Parameter resetProfile:: Resets the execution profile.

Returns: A string containing the report.

loadEngineAndConnectStreams(self: popart_core._TrainingSessionCore) → None ¶

Load the engine on the device and connect the streams

This will set up the poplar::Streams.

Note: This call is optional. The engine will implicitly be loaded on the device when required.

loadExecutable(self: popart_core._InferenceSessionCore, filename: str) → None ¶

Load the poplar::Executable and the PopART metadata from the given file. The file must have been created with compileAndExport()

Parameter filename:: Name of the file to load the executable from.

modelToHost(self: popart_core._TrainingSessionCore, arg0: str) → None ¶

Write current model to ONNX file.

Parameter fn:: Path to file. Can be absolute or relative. If you plan to run your program in multiple processes simultaneously, you should avoid possible race conditions by writing to different files, for example by using temporary files.

prepareDevice(self: popart_core._TrainingSessionCore, loadEngine: bool = True, err: popart_core.OutOfMemoryError) → None ¶

Prepare the network for execution.

This will create the snap::Graph and poplar::Engine.

Parameter loadEngine:: Load the engine and connect the streams once the device is ready.

readWeights(self: popart_core._TrainingSessionCore, arg0: popart_core.IWeightsIO) → None ¶

Read the weights. Must have called weightsToHost() first.

The weight data is written to the addresses in weightsIo.out.

resetHostWeights(self: popart_core._TrainingSessionCore, modelProtoOrFilename: str, ignoreWeightsInModelWithoutCorrespondingHostWeight: bool = False) → None ¶

Reset the weights with the weights in an ONNX model that differs from the current model only in weights. This only updates the weights on the host; the user still needs to call weightsFromHost() after this to update the weights on the device.

Parameter model:: Either an ONNX model protobuf, or the name of a file containing an ONNX model protobuf.
Parameter ignoreWeightsInModelWithoutCorrespondingHostWeight:: If true, do not error if there are initializers in the ONNX model with no corresponding initializer tensor in the session’s IR.

run(self: popart_core._TrainingSessionCore, stepio: popart_core.IStepIO, debugName: str = '') → None ¶

Perform one step.

Read input data from address in stepIO.in. Write the output data to addresses in stepIO.out.

Parameter stepIO:: Input and output data.
Parameter debugName:: Debug string to identify this run in logs.

setRNGState(self: popart_core._TrainingSessionCore, rngValue: List[int]) → None ¶

setRandomSeed(self: popart_core._TrainingSessionCore, seedValue: int) → None ¶

Sets the random number generator seed on all tiles of the device. This ensures deterministic behaviour of random operations in the graph.

Parameter The:: seed value.

updateExternallySavedTensorLocations(self: popart_core._TrainingSessionCore, arg0: str, arg1: str) → None ¶

Update the tensor locations of the tensors in the Session’s ONNX model. The new file will be created at this point, and written to when the ONNX model is saved with a subsequent call to modelToHost.

Parameter fromLocation:: All externally saved tensors with location fromLocation will have their location updated to toLocation.
Parameter toLocation:: The updated location. Must not already exist.

updateOptimizerFromHost(self: popart_core._TrainingSessionCore, arg0: popart_core.Optimizer) → None ¶

Update the optimizer and the associated hyperparameters but not the optimizer state tensors.

NOTE: The optimizer parameter has to be compatible with the optimizer passed to the constructor. For example, you cannot call this function with an SDG1 optimizer if you created the session with an SDG0 optimizer. The reason for this is that it is not possible to change the IR after it has been constructed.

Parameter optimizer:: A pointer to a popart::Optimizer.

weightsFromHost(self: popart_core._TrainingSessionCore) → None ¶: Write weights from host to the device.

weightsToHost(self: popart_core._TrainingSessionCore) → None ¶: Copy the weights to host from the device.

writeWeights(self: popart_core._TrainingSessionCore, arg0: popart_core.IWeightsIO) → None ¶

Write the weights. Must call weightsFromHost() after this.

The weight data is written to the addresses in weightsIo.out.

2.7. Patterns¶

class popart.Patterns¶

enablePattern(self: popart_core.Patterns, arg0: str, arg1: bool) → popart_core.Patterns¶

enableRuntimeAsserts(self: popart_core.Patterns, arg0: bool) → popart_core.Patterns¶

isPatternEnabled(self: popart_core.Patterns, arg0: str) → bool ¶

2.8. Session Options¶

class popart.AccumulateOuterFragmentSettings¶

property excludedVirtualGraphs¶: A setting to explicitly tell PopART to avoid to try and parallelise the given virtual graph ids. This setting is experimental and may change.

property schedule¶: Tell PopART how you would like to schedule the accumulate outer fragment. This setting is experimental and may change.

class popart.AccumulateOuterFragmentSchedule¶

Enum type that determines how the operations in the accumulate outer fragment will be scheduled accross virtual graphs (only relevant to pipelined modes).

Members:

Scheduler : Don’t add additional constraints and let the scheduler work it out.

Serial : Add constraints that ensure ops are executed in virtual graph ID

order.

OverlapCycleOptimized : Try and parallelise ops with different virtual graph IDs as much as

possible.

OverlapMemoryOptimized : Try and parallelise ops with different virtual graph IDs but avoid

certain steps that are costly in terms of memory usage.

class popart.AutodiffSettings¶

class popart.AutodiffStitchStrategy¶

Members:

RecomputeMinimal

RecomputeAllNonInputs

AddFwdOutputs

SafeAddFwdOutputs

class popart.BatchSerializationSettings¶

A structure containing batch serialization settings.

property batchSchedule¶: Experimental value that changes how operations are scheduled.

property concatOnExecutionPhaseChange¶: Break batch serialization chains when the execution phase changes (by concatenating the compute batches to the local batch).

property concatOnPipelineStageChange¶: Break batch serialization chains when the pipeline stage changes (by concatenating the compute batches to the local batch).

property concatOnVirtualGraphChange¶: Break batch serialization chains when the virtual graph changes (by concatenating the compute batches to the local batch).

property factor¶: The number of compute batches to split operations into.

property method¶: Experimental value to control how batch serialization is applied.

property transformContext¶: Experimental value to control when batch serialization is applied.

class popart.BatchSerializationBatchSchedule¶

Enum type that describes how to change the batch serialisation subgraph schedule before outlining. NOTE: This setting is experimental and may change.

Members:

Scheduler : Don’t encourage any particular scheduling for ops within batch

subgraphs (leave it to the scheduler) but tell the scheduler to schedule subgraphs in sequence.

Isomorphic : Encourage all ops within batch subgraphs to be scheduled identically

and for each subgraph to be scheduled in sequence (good for outlineability).

OverlapOnIo : Attempt to put the RemoteLoad for batch N+1 right after the compute

phase of batch N.

OverlapOnCompute : Attempt to put the RemoteLoad for batch N+1 right before the compute

phase of batch N.

class popart.DotCheck¶

Enum type used to identify at which stages of IR construction to export .dot files.

Members:

Fwd0 : Generate graph after construction of the forward pass.

Fwd1 : Generate graph after running pre-aliasing patterns.

Bwd0 : Generate graph after backwards construction.

PreAlias : Generate graph after all transformations, patterns, except the

aliasing.

Final : Generate graph after running aliasing patterns (the final IR).

property name¶

class popart.ExecutionPhaseSchedule¶

Enum type to specify the order of processing optimizer operations for different weights of the same execution phase.

The steps for phased execution consists of: - Copy to IO tiles if necessary (1) - Run collective operations if necessary (2) - Load optimizer state (3) - Update optimizer state (4) - Apply optimizer (5) - Store updated tensor if necessary (6)

Members:

Interleaving : Process above steps for one weight at a time (for example: 123456,

123456, 123456). The scheduler may interleave these steps.

Batch : Process above steps for all weights together, in a way that maximises

overlap potential between compute and exchange (for example: 333, 111, 222, 444, 555, 666).

BatchClusteredIO : Process above steps for all weights together, in a way that maximises

overlap potential between compute and exchange, and maximise stream copy merges by keeping RemoteLoad/RemoteStore operations clustered (for example: 333, 111, 222, 444, 555, 666).

class popart.TensorLocationSettings¶

property location¶: The default tensor location for this tensor type.

property minElementsForOffChip¶: A minimum number of elements below which offloading won’t be considered.

property minElementsForReplicatedTensorSharding¶: A minimum number of elements below which replicated tensor sharding (RTS) won’t be considered.

class popart.ReplicatedTensorSharding¶

Enum type to specify whether to shard tensors over replicas.

Members:

Off : Don’t shard tensors over replicas.

On : Do shard tensors over replicas.

class popart.TensorLocation¶

class popart.TensorStorage¶

Enum type that determines where a tensor is stored.

Members:

OnChip : Store the tensor in on-chip memory.

OffChip : Store the tensor in streaming memory.

class popart.TileSet¶

Enum type to specify a set of tiles.

Members:

Compute : The set of tiles designated for compute operations.

IO : The set of tiles designated for IO operations.

class popart.Instrumentation¶

Members:

Outer : Outer loop instrumentation, graph over all IPUs.

Inner : Inner loop instrumentation, graph per IPU.

class popart.MergeVarUpdateType¶

Enum type used to specify which VarUpdateOp ops to merge.

Members:

Off : Do not merge VarUpdateOp ops.

All : Merge all VarUpdateOp ops into as few groups as possible. This is a

good choice when memory is not a constraint.

AutoTight : Merge into groups, so that VarUpdateOp ops process tensors of exactly

mergeVarUpdateMemThreshold in size.

AutoLoose : Merge into groups while attempting not to increase maximum variable

liveness, and also not slice tensor variables so they they will need to be processed by different VarUpdateOp ops.

class popart.SyntheticDataMode¶

Members:

Off : Use real data.

Zeros : Input tensors are initialised to all zeros.

RandomNormal : Input tensors are initialised with distribution ~N(0,1).

class popart.RecomputationType¶

Enum type to specify which ops to recompute in the backwards pass when doing auto-recomputation.

Members:

NoRecompute : No ops are recomputed.

Standard : Algorithm to pick checkpoints to try and minimise max liveness.

NormOnly : Only Norm ops (+ non-linearities, if following) are recomputed.

Pipeline : Recompute all forward pipeline stages.

class popart.VirtualGraphMode¶

Members:

Off : Virtual graphs are not enabled.

Manual : User must set the virtualGraph attribute on all ops.

Auto : Use autoVirtualGraph transform.

ExecutionPhases : Virtual graphs are tied to execution phases.

class popart.SubgraphCopyingStrategy¶

Members:

OnEnterAndExit : Copy all inputs before the start of the subgraph, copy all outputs

after all ops in the subgraph. With this strategy subgraphs will always map to a single Poplar function.

JustInTime : Copy inputs just before they are consumed and copy outputs as soon as

they are produced. With this strategy subgraphs may be lowered into multiple Poplar functions.

class popart.ExecutionPhaseSettings¶

property activationIOSchedule¶: The execution phase IO schedule for activation and gradient tensors.

property phases¶: Number of ExecutionPhases for the whole model

property stages¶

Parallel streaming memory, default for 1 IPU / replica 2: PingPong between 2 IPUs, default for >= 2 IPUs / replica

Type: Number of overlapping stages 1

property weightIOSchedule¶: The execution phase IO schedule for weight tensors.

class popart_core.SessionOptions¶

property accumulateOuterFragmentSettings¶: Configuration setting for operations in the accumulate outer fragment.

property accumulationAndReplicationReductionType¶: Specify how gradients are reduced when using gradient accumulation. Note, this option has been deprecated in favour of accumulationAndReplicationReductionType, and will be removed in a future release.

property accumulationFactor¶: Specify the number of micro-batches to accumulate before applying the varUpdate.

property accumulatorTensorLocationSettings¶: Tensor location for gradient accumulator tensors.

property activationTensorLocationSettings¶: Tensor location settings for activation/gradient tensors.

property aliasZeroCopy¶: Enable zero-copy for subgraphs.

property autoRecomputation¶: Enable recomputation of operations in the graph in the backwards pass to reduce model size at the cost of computation cycles.

property batchSerializationSettings¶: Configuration setting for batch serialization.

property cachePath¶: Folder to save the poplar::Executable to.

property compileEngine¶: If false, the backend will build the Poplar graph but not compile it into an Engine. In this case, no execution can be performed, and nothing can be transferred to the device. API calls which retrieve information from the graph building stage, such as tile mapping introspection, can still be used.

property constantWeights¶: An optimization for an inference session to have constant weights, true by default. Set this option to false if you are going to want to change the weights with a call to Session::resetHostWeights after the session has been prepared. This option has no effect on a training session

property customCodeletCompileFlags¶: Compile flags for the custom codelets. For example -g to generate debug info.

property customCodelets¶: List of codelets (with filetype) to be added to the Poplar graph. See the Poplar documentation for more information.

property decomposeGradSum¶: Replaces single sums of partial gradients with a tree of additions. This can reduce max liveness at the cost of extra cycles. A typical use case for this would be if a large weight tensor is used as an input to many operations.

property disableGradAccumulationTensorStreams¶: If true, the weight gradient tensors are not saved off the device when devicex.weightsFromHost() is called. Note: this option is overridden if #syntheticDataMode is not #SyntheticDataMode::Off.

property dotChecks¶: When to write .dot files during Ir construction.

property dotOpNames¶: Include the Op name in the .dot file (the Op type is always exported).

property enableDistributedReplicatedGraphs¶: Enable training with Poplar replicated graphs across multiple PopART instances.

property enableEngineCaching¶: Enable Poplar executable caching.

property enableFloatingPointChecks¶: Throw an exception when floating point errors occur.

property enableFullyConnectedPass¶: Enable the global #fullyConnectedPass option for matmuls.

property enableGradientAccumulation¶: Enable gradient accumulation.

property enableGroupedMatmuls¶: Enable/disable the grouping of matmuls that are the same shape.

property enableLoadAndOffloadRNGState¶: Allows to load/offload device RNG state from host.

property enableNonStableSoftmax¶: By default, we use the stable softmax Poplar function. The input tensor to softmax, _x_, is preprocessed by subtracting max(_x_) from each element before computing the exponentials, ensuring numerical stability. If you are sure the inputs to your softmax operations are small enough to not cause overflow when computing the exponential, you can enable the non-stable version instead, to increase the speed.

property enableOutlining¶: Identify and extract repeated parts of computational graph into subgraphs.

property enableOutliningCopyCostPruning¶: When true the cost of copying of cached sections should be included in the outlining cost model.

property enablePipelining¶: Enable pipelining of virtual graphs

property enableReplicatedGraphs¶: Enable replication of graphs.

property enableStableNorm¶: If true, computes the mean first and subtracts the activations from it before computing the variance. The implementation with this flag set to true is slower than when set to false. The stable version requires the first order moment to be estimated and applied to the sample set before the second order central moment is calculated.

property enableStochasticRounding¶: Enable stochastic rounding.

property executionPhaseSettings¶: Configuration settings for execution phases.

property explicitRecomputation¶: Enable explicit recomputation.

property exportPoplarComputationGraph¶: Export Poplar computation graph.

property exportPoplarVertexGraph¶: Export Poplar vertex graph.

property finalDotOp¶: See #firstDotOp.

property firstDotOp¶: The ops to write to the .dot file will be a continuous interval of the schedule, controlled by firstDotOp and finalDotOp. In particular, it will be [min(0, firstDotOp), max(N ops in Ir, finalDotOp)).

property globalReplicaOffset¶: The first replica index that this PopART instance is running.

property globalReplicationFactor¶: The total number of replicas in a multi instance replicated graph training session (this should be left as the default value (1) if distributed replicated graphs are disabled). This value includes local replication.

property groupHostSync¶: Allows to group the streams from host at the beginning and the streams to host at the end, this trades off sum-liveness efficiency for cycle efficiency.

property hostAllReduce¶: Perform AllReduce operation on the host. Only useful for training session.

property hostAllReduceRemoteBuffer¶: Enable the use of poplar::RemoteBuffers for hostAllReduce operations.

property hostWeightUpdate¶: Perform weight update on the host. Only useful for training session.

property instrumentWithHardwareCycleCounter¶: Add instrumentation to your program to count the number of device cycles (of a single tile, on a single IPU) that your main program takes to execute. Expect this to have a small detrimental impact on performance.

property kahnTieBreaker¶: The initial scheduling is done with Kahn’s algorithm. When several Ops are free to be scheduled, this controls which method is used.

property logDir¶: A directory for log traces to be written into.

property mergeVarUpdate¶: Enable merging of VarUpdates into groups of VarUpdates, by flattening and concatenating variable tensors and updating tensors.

property mergeVarUpdateMemThreshold¶: The #MergeVarUpdateType::AutoLoose and #MergeVarUpdateType::AutoTight VarUpdateOp merging algorithms have a threshold on the total memory of variable tensors to merge for updating. Defined as total memory in bytes.

property optimizerStateTensorLocationSettings¶: Tensor location for optimizer state tensors.

property opxAliasChecking¶: Run Opx checks to verify IR tensor aliasing information corresponds to lowered Poplar tensor aliasing.

property opxModifyChecking¶: Run Opx checks to verify IR tensor modification information corresponds to lowered Poplar tensor modifications.

property outlineSequenceBreakCost¶: The penalty applied to outlining potential sub-graphs if the sub-graph to be created breaks up a sequence of operations that are more efficient (for example for overlapping compute and exchange) when outlined together. Default value is set to ~10 * Op::getHighSubgraphValue().

property outlineThreshold¶: The incremental value that a sub-graph requires, relative to its nested sub-graphs (if any), to be eligible for outlining. A high threshold results in fewer sub-graphs being outlined, a negative value results in all being outlined. The gross value of a sub-graph is the sum of its constituent Ops’ Op::getSubgraphValue() values. To disable outlining, it is better to set enableOutlining to false than to set this value to infinity. The default value of 1.0f results in all high value operations such as convolution being cached, but standalone low Value operations such as Relu will not be.

property partialsTypeMatMuls¶: Set the partials type globally for matmuls. Can be overridden individually with Builder.setPartialsType(). Valid values are "float" and "half". By default, this is not set, so no global partials type is imposed.

property rearrangeAnchorsOnHost¶: Before anchor tensors are streamed from device to host, they are not necessarily arranged in memory as required when they are to be copied from host stream to host. This can be done on the device or on the host. Done on host by default to save memory, but often at the expense of cycles, especially for larger anchor tensors.

property replicatedGraphCount¶: If enableReplicatedGraphs is true, replicatedGraphCount will set the number of model replications. For example, if your model uses 1 IPU, a replicatedGraphCount of 2 will use 2 IPUs. If your model is pipelined across 4 IPUs, a replicatedGraphCount of 4 will use 16 IPUs total. Therefore, the number of IPUs you request must be a multiple of replicatedGraphCount. If the training is done across multiple instances then the replicatedGraphCount is the number of replicas for this instance.

property separateCallOpPdfs¶: When generating PDFs of IR graphs, create separate PDFs for each subgraph.

property serializedPoprithmsAnnealGraphsDir¶: PopART uses Poprithms for scheduling PopART graphs. The Poprithms graphs created for scheduling can be optionally serialised (written to file). The string below specified the directory to serialize Poprithms graphs to. If it is empty, then the graphs will not be serialised. The names of serialization files will be poprithms_shift_graph_i.json for the lowest non-existing values of i. The directory must already exist, PopART will not create it.

property serializedPoprithmsShiftGraphsDir¶: PopART uses Poprithms for scheduling PopART graphs. The Poprithms graphs created for scheduling can be optionally serialised (written to file). The string below specified the directory to serialize Poprithms graphs to. If it is empty, then the graphs will not be serialised. The names of serialization files will be poprithms_shift_graph_i.json for the lowest non-existing values of i. The directory must already exist, PopART will not create it.

property strictOpVersions¶: Strict op version checks will throw an error if the exact version of an op required for the models opset is not supported. Turning this check off will cause PopART to fall back to the latest implementation of the op that is supported. Warning, turning off these checks may cause undefined behaviour.

property subgraphCopyingStrategy¶: This setting determines how copies for inputs and outputs for subgraphs are lowered. By setting this value to JustInTime you may save memory at the cost of fragmenting subgraphs into multiple Poplar functions. This may be particularly useful when a number of weight updates are outlined in one subgraph, as it may prevent multiple weight tensors from being live at the same time inside the subgraph.

property swapLimitScheduler¶: The maximum number of improving steps allowed by the scheduling algorithm before a solution must be returned.

property syntheticDataMode¶

disable data transfer to/from the host. Set to #SyntheticDataMode::Off to use real data.

Type: Use synthetic data

property timeLimitScheduler¶: The maximum allowed time that can be spent searching for a good graph schedule before a solution must be returned.

property weightTensorLocationSettings¶: Tensor location for weight tensors.

2.9. Optimizers¶

class popart_core.Optimizer¶

getLossScalingVal(self: popart_core.Optimizer) → float ¶

class popart_core.SGD¶

Stochastic Gradient Descent (%SGD) optimizer.

Akin to any optimizer implementation, this class is responsible for updating each weight tensor ($w$) in the model using the gradient ($g$) of the loss function with respect to the weight as calculated during the backwards pass.

The %SGD optimizer has the following state for each weight:

velocity ($v$)

The %SGD optimizer has the following hyper parameters:

learning rate ($text{lr}$) * momentum ($text{mm}$) * *weight

decay* ($text{wd}$) * dampening ($text{dm}$) * velocity scaling ($text{vs}$) * loss scaling ($text{ls}$) * clip norm settings

The values of these parameters can be shared between all weights but some can be overridden with weight-specific values (see SGD::insertSpecific). Hyper parameters are captured using OptimizerValue objects and therefore can be either a constant value or a non-constant value that can be adjusted by the user.

In the following we will describe how this optimizer updates a weight using a gradient. In the context of this description the gradient is is the value of the gradient after any gradient accumulation has been performed and after the application of a loss scaling factor to the gradient has been corrected for.

When the optimizer needs to update a weight, $w$, using a gradient, $g$, it first updates the optimizer state as follows:

f[ v’ := v * text{mm} + (1 - text{dm}) * (g + text{wd} * w) text{ . } f]

Following the update of the optimizer state the optimizer uses said state to update the weight:

f[ w’ := w - text{lr} * v’ text{ . } f]

In addition to the above, the velocity scaling hyper parameter is a scaling factor that can provide improved numerical stability by ensuring the values stored in the optimizer state, $v$, are scaled by this value. When using this parameter PopART will automatically deal with the artificially scaled velocity value during the weight update and other hyper parameters do not need to be adjusted).

In addition, the loss scaling hyper parameter is similar in nature to the velocity scaling parameter. It is a scaling value that is applied to the loss gradient at the start of the the backwards pass and, at the end of the backwards pass, this scaling is reversed by multiplying the gradients for each weight with the inverse of the loss scaling value prior to updating the optimizer state. Using loss scaling can also improve numerical stability in some cases.

Finally, it is possible to add clip norm settings for this optimizer. These clip norms compute the L2 norm for a group of weights and adds a scalar term to the weight update that effectively divides it by the norm (or a constant value that is provided as part of the clip norm, which ever is greater).

dampenings(self: popart_core.SGD) → popart_core.OptimizerValueMap¶

insertSpecific(self: popart_core.SGD, arg0: str, arg1: dict) → None ¶

learningRates(self: popart_core.SGD) → popart_core.OptimizerValueMap¶

momentums(self: popart_core.SGD) → popart_core.OptimizerValueMap¶

velocityScalings(self: popart_core.SGD) → popart_core.OptimizerValueMap¶

weightDecays(self: popart_core.SGD) → popart_core.OptimizerValueMap¶

class popart_core.ConstSGD¶

Stochastic Gradient Descent (SGD) optimizer with constant learning rate, weight decay, loss scaling and clip norm settings (and default values for momentum, dampening or velocity scaling).

NOTE: See SGD for detailed meaning for these parameters.

NOTE: This class exists for backwards compatibility with the Python API and may be removed at some point in the future.

class popart_core.Adam¶

AdamW, Lamb and AdaMax optimizer implementation.

Akin to any optimizer implementation, this class is responsible for updating each weight tensor ($w$) in the model using the gradient ($g$) of the loss function with respect to the weight as calculated during the backwards pass.

The optimizer has the following state for each weight:

first-order momentum ($m$) * second-order momentum ($v$) * *time

step* ($t$)

The optimizer has the following hyper parameters:

learning rate ($text{lr}$) * weight decay ($text{wd}$) *

beta1 ($beta_1$) * beta2 ($beta_2$) * epsilon ($epsilon$) * loss scaling ($text{ls}$) * maximum weight norm ($text{mwn}$)

The values of these parameters can be shared between all weights but some can be overridden with weight-specific values (see Adam::insertSpecific). Hyper parameters are captured using OptimizerValue objects and therefore can be either a constant value or a non-constant value that can be adjusted by the user.

The values of #AdamMode and #WeightDecayMode passed to the constructor determines how weights are updated (see below).

In the following we will describe how this optimizer updates a weight using a gradient. In the context of this description the gradient is is the value of the gradient after any gradient accumulation has been performed and after the application of a loss scaling factor to the gradient has been corrected for.

When the optimizer needs to update a weight, $w$, using a gradient, $g$, it first computes a term $g_text{tmp}$, which is effectively is $g$ with L2 regularization applied if the #WeightDecayMode is set to WeightDecayMode::L2Regularization this, as follows:

f[ g_text{tmp} := left{begin{aligned} g & text{ ; (Decay) } \ (g + text{wd} * w) & text{ ; (L2Regularization) ; . } \ end{aligned}right.\ f]

Secondly, the optimizer updates the optimizer state as follows:

f[ m’ &:= beta_1 * m + (1 - beta_1) * g_text{tmp} \ v’ &:= left{begin{aligned} beta_2 * v + (1 - beta_2) * g_text{tmp}^2 & text{ ; (Adam/AdamNoBias) } \ beta_2 * v + (1 - beta_2) * g_text{tmp}^2 & text{ ; (Lamb/LambNoBias) } \ text{max}(beta_2 * v, |g_\text{tmp}|) & text{ ; (AdaMax) } \ end{aligned}right.\ t’ &:= t + 1 \ f]

Next, it computes the following terms:

f[ m_text{tmp} &:= left{begin{aligned} m’ & text{ ; (AdamNoBias/LambNoBias) } \ frac{m’}{(1 - beta_1^{t’})} & text{ ; (Adam/Lamb/AdaMax) } \ end{aligned}right.\ v_text{tmp} &:= left{begin{aligned} v’ & text{ ; (AdamNoBias/LambNoBias) } \ frac{v’}{(1 - beta_2^{t’})} & text{ ; (Adam/Lamb/AdaMax) } \ end{aligned}right.\ u_text{tmp} &:= left{begin{aligned} frac{m_text{tmp}}{(sqrt{v_text{tmp}} + epsilon)} + text{wd} * w &text{ ; (Decay) } \ frac{m_text{tmp}}{(sqrt{v_text{tmp}} + epsilon)} &text{ ; (L2Regularization) } \ end{aligned}right. f]

Finally, the optimizer updates the weight as follows:

f[ w’ := left{begin{aligned} w - text{lr} * u_text{tmp} &text{ ; (Adam/AdamNoBias/AdaMax) } \ w - biggl(frac{text{min}(lVert{w}rVert, text{mwn})}{lVert{u_text{tmp}}rVert}biggr) * text{lr} * u_text{tmp} &text{ ; (Lamb/LambNoBias) } \ end{aligned}right. f]

In addition to the above, the loss scaling hyper parameter is similar in nature to the velocity scaling parameter. It is a scaling value that is applied to the loss gradient at the start of the the backwards pass and, at the end of the backwards pass, this scaling is reversed by multiplying the gradients for each weight with the inverse of the loss scaling value prior to updating the optimizer state. Using loss scaling can also improve numerical stability in some cases.

NOTE: The maximum weight norm is referred to as $phi$ in [You et al., 2020](https://arxiv.org/abs/1904.00962).

beta1s(self: popart_core.Adam) → popart_core.OptimizerValueMap¶

beta2s(self: popart_core.Adam) → popart_core.OptimizerValueMap¶

epss(self: popart_core.Adam) → popart_core.OptimizerValueMap¶

insertSpecific(self: popart_core.Adam, arg0: str, arg1: dict) → None ¶

learningRates(self: popart_core.Adam) → popart_core.OptimizerValueMap¶

maxWeightNorms(self: popart_core.Adam) → popart_core.OptimizerValueMap¶

weightDecays(self: popart_core.Adam) → popart_core.OptimizerValueMap¶