Fp16 Moe Plugin#
-
class Fp16MoePlugin : public nvinfer1::IPluginV3, public nvinfer1::IPluginV3OneCore, public nvinfer1::IPluginV3OneBuild, public nvinfer1::IPluginV3OneRuntime#
TensorRT V3 plugin for the portable FP16 MoE CuTeDSL kernels.
Public Functions
- Fp16MoePlugin(
- std::string const &name,
- int32_t numExperts,
- int32_t topK,
- int32_t hiddenSize,
- int32_t moeInterSize,
- int32_t activationType,
- int32_t normTopkProb,
- int32_t maxRoutedRows,
- Fp16MoePlugin(
- std::string const &name,
- nvinfer1::PluginFieldCollection const *fields,
-
Fp16MoePlugin() = delete#
-
Fp16MoePlugin(Fp16MoePlugin const&) = delete#
-
~Fp16MoePlugin() noexcept override = default#
- nvinfer1::IPluginCapability *getCapabilityInterface(
- nvinfer1::PluginCapabilityType type,
-
nvinfer1::IPluginV3 *clone() noexcept override#
-
char const *getPluginName() const noexcept override#
-
char const *getPluginVersion() const noexcept override#
-
char const *getPluginNamespace() const noexcept override#
-
void setPluginNamespace(char const *pluginNamespace) noexcept#
-
int32_t getNbOutputs() const noexcept override#
- int32_t getOutputDataTypes(
- nvinfer1::DataType *outputTypes,
- int32_t nbOutputs,
- nvinfer1::DataType const *inputTypes,
- int32_t nbInputs,
- 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,
- bool supportsFormatCombination(
- int32_t pos,
- nvinfer1::DynamicPluginTensorDesc const *inOut,
- int32_t nbInputs,
- int32_t nbOutputs,
- int32_t configurePlugin(
- nvinfer1::DynamicPluginTensorDesc const *inputs,
- int32_t nbInputs,
- nvinfer1::DynamicPluginTensorDesc const *outputs,
- int32_t nbOutputs,
- size_t getWorkspaceSize(
- nvinfer1::DynamicPluginTensorDesc const *inputs,
- int32_t nbInputs,
- nvinfer1::DynamicPluginTensorDesc const *outputs,
- int32_t nbOutputs,
- int32_t enqueue(
- nvinfer1::PluginTensorDesc const *inputDesc,
- nvinfer1::PluginTensorDesc const *outputDesc,
- void const *const *inputs,
- void *const *outputs,
- void *workspace,
- cudaStream_t stream,
- int32_t onShapeChange(
- nvinfer1::PluginTensorDesc const *inputs,
- int32_t nbInputs,
- nvinfer1::PluginTensorDesc const *outputs,
- int32_t nbOutputs,
- nvinfer1::IPluginV3 *attachToContext(
- nvinfer1::IPluginResourceContext *context,
- nvinfer1::PluginFieldCollection const *getFieldsToSerialize(
-
class Fp16MoePluginCreator : public nvinfer1::IPluginCreatorV3One#
Creator for Fp16MoePlugin version 1.
Public Functions
-
Fp16MoePluginCreator()#
-
~Fp16MoePluginCreator() override = default#
-
char const *getPluginName() const noexcept override#
-
char const *getPluginVersion() const noexcept override#
- nvinfer1::PluginFieldCollection const *getFieldNames(
-
char const *getPluginNamespace() const noexcept override#
-
void setPluginNamespace(char const *pluginNamespace) noexcept#
- nvinfer1::IPluginV3 *createPlugin(
- char const *name,
- nvinfer1::PluginFieldCollection const *fields,
- nvinfer1::TensorRTPhase phase,
-
Fp16MoePluginCreator()#