llama : fix op mul check with command-r-plus (ggml-org#10476)

slaren · web-flow · commit dc39012cbaf8 · 2024-11-24T16:10:26.000+01:00
diff --git a/src/llama.cpp b/src/llama.cpp
@@ -7181,12 +7181,12 @@ static bool weight_buft_supported(const llama_hparams & hparams, ggml_tensor * w
             } break;
         case GGML_OP_ADD:
             {
-                ggml_tensor * a = ggml_new_tensor_2d(ctx, GGML_TYPE_F32, w->ne[0], 512);
+                ggml_tensor * a = ggml_new_tensor_4d(ctx, GGML_TYPE_F32, w->ne[0], w->ne[1], w->ne[2], w->ne[3]);
                 op_tensor = ggml_add(ctx, a, w);
             } break;
         case GGML_OP_MUL:
             {
-                ggml_tensor * a = ggml_new_tensor_2d(ctx, GGML_TYPE_F32, w->ne[0], 512);
+                ggml_tensor * a = ggml_new_tensor_4d(ctx, GGML_TYPE_F32, w->ne[0], w->ne[1], w->ne[2], w->ne[3]);
                 op_tensor = ggml_mul(ctx, a, w);
             } break;
         case GGML_OP_DIV:
