QuantizedTensor<T>

Models & Types · AiDotNet.Tensors.LinearAlgebra

No summary documentation available yet.

Properties

Property Summary
QuantizationBits Bits { get; }
QuantizationScale Scale { get; }
QuantizationScheme Scheme { get; }
byte[] Payload { get; }
int Length { get; }
int[] Shape { get; }

Methods

Method Summary
QuantizedTensor<PackedInt4> FromFloatInt4(Tensor<float> source, int groupSize, QuantizationScheme scheme)
QuantizedTensor<sbyte> FromFloatInt8(Tensor<float> source, int groupSize, QuantizationScheme scheme)
Tensor<float> Dequantize()