model : support NVFP4 tensors for Gemma4 (#21971)

* support nvfp4 tensors for Gemma4 * add wo_s to build_attn * add wo_s to build_attn * fix glm4
2026-04-16 16:51:47 +02:00
parent b572d1ecd6
commit f772f6e434
105 changed files with 149 additions and 148 deletions
@@ -67,7 +67,7 @@ llm_build_openai_moe_iswa::llm_build_openai_moe_iswa(const llama_model & model,
            cb(Vcur, "Vcur", il);

            cur = build_attn(inp_attn,
-                    model.layers[il].wo, model.layers[il].bo,
+                    model.layers[il].wo, model.layers[il].bo, model.layers[il].wo_s,
                    Qcur, Kcur, Vcur, nullptr, model.layers[il].attn_sinks, nullptr, 1.0f/sqrtf(float(n_rot)), il);

            cb(cur, "attn_out", il);