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
@@ -44,7 +44,7 @@ llm_build_t5_enc::llm_build_t5_enc(const llama_model & model, const llm_graph_pa
            ggml_tensor * kq_b = build_pos_bias(pos_bucket_enc, attn_rel_b);

            cur = build_attn(inp_attn,
-                    model.layers[il].wo_enc, nullptr,
+                    model.layers[il].wo_enc, nullptr, nullptr,
                    Qcur, Kcur, Vcur, kq_b, nullptr, nullptr, 1.0f, il);
            cb(cur, "kqv_out", il);
        }