EMLXAxon.Qwen3.Model.State (emlx_axon v0.3.0)

Copy Markdown View Source

Loaded model weights and config.

Summary

Types

t()

Types

t()

@type t() :: %EMLXAxon.Qwen3.Model.State{
  config: map(),
  embed_tokens: Nx.Tensor.t(),
  layers: [map()],
  lm_head: Nx.Tensor.t(),
  norm: Nx.Tensor.t()
}