master
model : refactor QKV into common build_qkv and create_tensor_qkv helpers (#21245)
model : refactor bias tensor variable names (#22079)
refactor : llama-model.cpp (#16252)
model : support NVFP4 tensors for Gemma4 (#21971)
graph : remove redundant GDN state transposes (#20443)
model : wire up Nemotron-H tensors for NVFP4 support (#20561)
models : deduplicate delta-net graphs for Qwen family (#19597)
model: using single llm_build per arch (#21970)
models : move the token embedding norms to the first layer (#20943)