LlamaGgufTensorNames

Standard GGUF tensor naming for LLaMA-family models.

Uses the canonical blk.N.* naming scheme from llama.cpp GGUF files. This is also the default scheme for Mistral, Qwen, and other LLaMA derivatives.

Functions

open override fun layerAttnK(layer: Int): String

open override fun layerAttnNorm(layer: Int): String

open override fun layerAttnO(layer: Int): String

open override fun layerAttnQ(layer: Int): String

open override fun layerAttnV(layer: Int): String

open override fun layerFfnDown(layer: Int): String

open override fun layerFfnGate(layer: Int): String

open override fun layerFfnNorm(layer: Int): String

open override fun layerFfnUp(layer: Int): String

open override fun outputNorm(): String

open override fun outputWeight(): String

open override fun tokenEmbedding(): String