DataType

Enums · AiDotNet.Onnx.Protobuf

No summary documentation available yet.

Fields

Field Summary
DataType Bfloat16 Non-IEEE floating-point format based on IEEE754 single-precision floating-point number truncated to 16 bits.
DataType Bool bool
DataType Complex128 complex with float64 real and imaginary components
DataType Complex64 complex with float32 real and imaginary components
DataType Double
DataType Float Basic types.
DataType Float16 IEEE754 half-precision floating-point format (16 bits wide).
DataType Float8E4M3Fn Non-IEEE floating-point format based on papers FP8 Formats for Deep Learning, https://arxiv.org/abs/2209.05433, 8-bit Numerical Formats For Deep Neural Networks, https://arxiv.org/pdf/2206.02915.pdf.
DataType Float8E4M3Fnuz float 8, mostly used for coefficients, supports nan, not inf, no negative zero
DataType Float8E5M2 follows IEEE 754, supports nan, inf, mostly used for gradients
DataType Float8E5M2Fnuz follows IEEE 754, supports nan, not inf, mostly used for gradients, no negative zero
DataType Int16 int16_t
DataType Int32 int32_t
DataType Int4 Signed integer in range [-8, 7], using two’s-complement representation
DataType Int64 int64_t
DataType Int8 int8_t
DataType String string
DataType Uint16 uint16_t
DataType Uint32
DataType Uint4 4-bit data-types
DataType Uint64
DataType Uint8 uint8_t
DataType Undefined
int value__