Nvfp4 A16 Moe Plugin#

class Nvfp4A16MoePlugin : public nvinfer1::IPluginV3, public nvinfer1::IPluginV3OneCore, public nvinfer1::IPluginV3OneBuild, public nvinfer1::IPluginV3OneRuntime#

TensorRT V3 plugin for FP16/BF16 activations and Marlin-packed NVFP4 expert weights.

max_routed_rows is the capacity of the padded Marlin routing arrays, not only the number of selected slots. A runtime shape with T tokens requires at most T * top_k + num_experts * (block_size - 1) rows, where block_size is 8 for S == 1 and 32 otherwise.

Public Functions

Nvfp4A16MoePlugin(
std::string const &name,
int32_t numExperts,
int32_t topK,
int32_t hiddenSize,
int32_t moeInterSize,
int32_t activationType,
int32_t nGroup,
int32_t topkGroup,
int32_t normTopkProb,
float routedScalingFactor,
int32_t routingMode,
int32_t maxRoutedRows
)#
Nvfp4A16MoePlugin(
std::string const &name,
nvinfer1::PluginFieldCollection const *fc
)#
Nvfp4A16MoePlugin() = delete#
Nvfp4A16MoePlugin(Nvfp4A16MoePlugin const&) = delete#
Nvfp4A16MoePlugin &operator=(Nvfp4A16MoePlugin const&) = delete#
~Nvfp4A16MoePlugin() noexcept override#
nvinfer1::IPluginCapability *getCapabilityInterface(
nvinfer1::PluginCapabilityType type
) noexcept override#
nvinfer1::IPluginV3 *clone() noexcept override#
char const *getPluginName() const noexcept override#
char const *getPluginVersion() const noexcept override#
char const *getPluginNamespace() const noexcept override#
int32_t getNbOutputs() const noexcept override#
int32_t getOutputDataTypes(
nvinfer1::DataType *outputTypes,
int32_t nbOutputs,
nvinfer1::DataType const *inputTypes,
int32_t nbInputs
) const noexcept override#
int32_t getOutputShapes(
nvinfer1::DimsExprs const *inputs,
int32_t nbInputs,
nvinfer1::DimsExprs const *shapeInputs,
int32_t nbShapeInputs,
nvinfer1::DimsExprs *outputs,
int32_t nbOutputs,
nvinfer1::IExprBuilder &exprBuilder
) noexcept override#
bool supportsFormatCombination(
int32_t pos,
nvinfer1::DynamicPluginTensorDesc const *inOut,
int32_t nbInputs,
int32_t nbOutputs
) noexcept override#
int32_t configurePlugin(
nvinfer1::DynamicPluginTensorDesc const *in,
int32_t nbInputs,
nvinfer1::DynamicPluginTensorDesc const *out,
int32_t nbOutputs
) noexcept override#
size_t getWorkspaceSize(
nvinfer1::DynamicPluginTensorDesc const *inputs,
int32_t nbInputs,
nvinfer1::DynamicPluginTensorDesc const *outputs,
int32_t nbOutputs
) const noexcept override#
int32_t enqueue(
nvinfer1::PluginTensorDesc const *inputDesc,
nvinfer1::PluginTensorDesc const *outputDesc,
void const *const *inputs,
void *const *outputs,
void *workspace,
cudaStream_t stream
) noexcept override#
int32_t onShapeChange(
nvinfer1::PluginTensorDesc const *in,
int32_t nbInputs,
nvinfer1::PluginTensorDesc const *out,
int32_t nbOutputs
) noexcept override#
nvinfer1::IPluginV3 *attachToContext(
nvinfer1::IPluginResourceContext *context
) noexcept override#
nvinfer1::PluginFieldCollection const *getFieldsToSerialize(
) noexcept override#
void setPluginNamespace(char const *pluginNamespace) noexcept#
class Nvfp4A16MoePluginCreator : public nvinfer1::IPluginCreatorV3One#

Creator for Nvfp4A16MoePlugin.

Public Functions

Nvfp4A16MoePluginCreator()#
~Nvfp4A16MoePluginCreator() override = default#
char const *getPluginName() const noexcept override#
char const *getPluginVersion() const noexcept override#
nvinfer1::PluginFieldCollection const *getFieldNames(
) noexcept override#
char const *getPluginNamespace() const noexcept override#
void setPluginNamespace(char const *pluginNamespace) noexcept#
nvinfer1::IPluginV3 *createPlugin(
char const *name,
nvinfer1::PluginFieldCollection const *fc,
nvinfer1::TensorRTPhase phase
) noexcept override#