Mila
Deep Neural Network Library
Loading...
Searching...
No Matches
Mila::Dnn::Quant::KvCache::KvCachePolicy Concept Referenceexport

Minimum contract for all KV cache management policies. More...

Concept definition

template<typename T>
concept KvCachePolicy = requires
{
{ T::kIsActive } -> std::convertible_to<bool>;
}
Minimum contract for all KV cache management policies.
Definition Policy.ixx:35

Detailed Description

Minimum contract for all KV cache management policies.

Requires only kIsActive. Active policies set kIsActive = true and provide algorithm-specific fields consumed by CudaGqaOp dispatch via if constexpr. Active policy refinements such as QuantKvPolicy impose additional field requirements in their own modules. A SlidingWindowPolicy satisfies KvCachePolicy without carrying dtype fields.

Template Parameters
TThe type to constrain against the policy contract.