LlamaGGUFNameResolver

Resolves network DSL module paths to GGUF tensor names for LLaMA-family models.

Module tree paths (from collectParamsWithPath): "MLP/token_embd" + "token_embd.weight" "MLP/blk.0/attn_norm" + "attn_norm.weight" "MLP/blk.0/attn" + "attn.q_proj.weight" "MLP/blk.0/ffn" + "ffn.gate_proj.weight" "MLP/output_norm" + "output_norm.weight" "MLP/output" + "output.weight"

GGUF tensor names: "token_embd.weight" "blk.0.attn_norm.weight" "blk.0.attn_q.weight" "blk.0.ffn_gate.weight" "output_norm.weight" "output.weight"

The resolver strips the MLP prefix, extracts the block prefix from the path, and applies GGUF-specific naming conventions.