fix: Only allocate attention cache for attention layers (not non-recurrent)

https://github.com/ggml-org/llama.cpp/issues/nemotron-nano-15409
Branch: gabe-l-hart/nvidia-nemotron-nano-15409

Signed-off-by: Gabe Goodhart <[email protected]>

Author: gabe-l-hart

src/llama-model.cpp

@@ -18316,6 +18316,23 @@ llama_memory_i * llama_model::create_memory(const llama_memory_params & params,
                             std::max((uint32_t) 1, cparams.n_seq_max),
                             cparams.n_seq_max);
                 } else if (llm_arch_is_hybrid(arch)) {
+
+                    // The main difference between hybrid architectures is the
+                    // layer filters, so pick the right one here
+                    llama_memory_hybrid::layer_filter_cb filter_attn = nullptr;
+                    llama_memory_hybrid::layer_filter_cb filter_recr = nullptr;
+                    if (arch == LLM_ARCH_FALCON_H1) {
+                        filter_attn = [&](int32_t) { return true; };
+                        filter_recr = [&](int32_t) { return true; };
+                    } else if (arch == LLM_ARCH_NEMOTRONH) {
+                        filter_attn = [&](int32_t il) {
+                            return !hparams.is_recurrent(il) && hparams.n_ff(il) == 0;
+                        };
+                        filter_recr = [&](int32_t il) {
+                            return hparams.is_recurrent(il) && hparams.n_ff(il) == 0;
+                        };
+                    }
+
                     const auto padding = llama_kv_cache::get_padding(cparams);
 
                     cparams.n_ctx = GGML_PAD(cparams.n_ctx, padding);
@@ -18335,8 +18352,8 @@ llama_memory_i * llama_model::create_memory(const llama_memory_params & params,
                         /* n_seq_max         */ cparams.n_seq_max,
                         /* offload           */ cparams.offload_kqv,
                         /* unified           */ cparams.kv_unified,
-                        /* filter_attn       */ (arch == LLM_ARCH_FALCON_H1) ? [&](int32_t) { return true; } : (llama_memory_hybrid::layer_filter_cb)nullptr,
-                        /* filter_recr       */ (arch == LLM_ARCH_FALCON_H1) ? [&](int32_t) { return true; } : (llama_memory_hybrid::layer_filter_cb)nullptr);
+                        /* filter_attn       */ std::move(filter_attn),
+                        /* filter_recr       */ std::move(filter_recr));
                 } else {
                     const auto padding = llama_kv_cache::get_padding(cparams);