Mila
Deep Neural Network Library
Loading...
Searching...
No Matches
Mila::Dnn::Quant::Weight::PerGroupFp4< kGroupSize > Struct Template Referenceexport

Static Public Attributes

static constexpr bool kIsFp4E2M1 = true
static constexpr bool kIsQuantized = true
static constexpr bool kPerChannel = false
static constexpr int kQuantizationGroupSize = kGroupSize
static constexpr TensorDataType kScaleDtype = TensorDataType::FP32
static constexpr TensorDataType kStorageDtype = TensorDataType::UINT8

The documentation for this struct was generated from the following file:
  • Mila/Src/Dnn/Quantization/Weight/Policies.ixx