Mila
Deep Neural Network Library
Loading...
Searching...
No Matches
Dnn.Components.Gqa Module Reference

Exported Modules

module  Compute.DeviceTypeTraits
module  Dnn.TensorTypes
module  Dnn.ComponentType
module  Compute.DeviceId
module  Dnn.Tensor
module  Compute.Device
module  Dnn.Components.GqaConfig
module  Dnn.Quantization.KvCache.Policy
module  Compute.ExecutionContext
module  Compute.DeviceType
module  Dnn.ITensor
module  Serialization.Mode
module  Compute.IKvCacheLifecycle
module  Compute.MemoryResource
module  Compute.GqaState
module  Dnn.TensorDataType
module  Compute.OperationTraits
module  Compute.IKvInference
module  Compute.CpuMemoryResource
module  Dnn.Component
module  Dnn.TensorDataTypeTraits
module  Serialization.ModelArchive
module  Compute.ExecutionContextFactory

Classes

class  Mila::Dnn::GroupedQueryAttention< TDeviceType, TComputePrecision, TKvPolicy >
 Grouped-Query Attention module that accepts concatenated QKV input. More...

Files

file  Mila/Src/Dnn/Components/Attention/GQA/GroupedQueryAttention.ixx
 Grouped-Query Attention module (concatenated QKV input).