Fix BA views as well

src/models/llm_build_qwen3next.cpp

@@ -317,14 +317,13 @@ ggml_tensor * llm_build_qwen3next::build_qwen3next_linear_attn_layer(llm_graph_i
         num_v_heads / num_k_heads   // alpha size
     };
 
-    ggml_tensor * b =
-        ggml_view_4d(ctx0, mixed_ba_reshaped, split_sizes_ba[0], num_k_heads, n_tokens, n_seqs,
-                     split_sizes_ba[0] * sizeof(float), mixed_ba_reshaped->nb[1], mixed_ba_reshaped->nb[2], 0);
+    ggml_tensor * b = ggml_view_4d(ctx0, mixed_ba_reshaped, split_sizes_ba[0], num_k_heads, n_tokens, n_seqs,
+                        mixed_ba_reshaped->nb[1], mixed_ba_reshaped->nb[2], mixed_ba_reshaped->nb[3], 0);
     cb(b, "b", il);
 
     ggml_tensor * a = ggml_view_4d(ctx0, mixed_ba_reshaped, split_sizes_ba[1], num_k_heads, n_tokens, n_seqs,
-                                   split_sizes_ba[1] * sizeof(float), mixed_ba_reshaped->nb[1],
-                                   mixed_ba_reshaped->nb[2], split_sizes_ba[0] * sizeof(float));
+                        mixed_ba_reshaped->nb[1], mixed_ba_reshaped->nb[2], mixed_ba_reshaped->nb[3], 
+                        split_sizes_ba[0] * ggml_element_size(mixed_ba_reshaped));
     cb(a, "a", il);
 
     // Reshape b and a to merge head dimensions: [batch, seq_len, num_k_heads, num_v_heads/num_k_heads] -> [batch, seq_len, num_v_heads]