diff --git a/ggml/src/ggml-rpc.cpp b/ggml/src/ggml-rpc.cpp
index 40ba5c82..e0f0fa05 100644
--- a/ggml/src/ggml-rpc.cpp
+++ b/ggml/src/ggml-rpc.cpp
@@ -466,7 +466,7 @@ GGML_CALL static void ggml_backend_rpc_buffer_init_tensor(ggml_backend_buffer_t
 
     // CUDA backend on the server pads everything to 512 due to CUDA limitations.
     // Due to bandwidth constraints, we only call the server init tensor functions if necessary.
-    // In particular, this is tensors with padding that needs to be cleared, so base tensors only and only misaligned.
+    // In particular, only quantized tensors need padding
     if (ggml_is_quantized(tensor->type) && (tensor->ne[0] % 512 != 0) && (tensor->view_src == nullptr)) {
         rpc_msg_init_tensor_req request;
 
@@ -628,7 +628,7 @@ GGML_CALL static size_t ggml_backend_rpc_get_max_size(ggml_backend_buffer_type_t
 GGML_CALL static size_t ggml_backend_rpc_buffer_type_get_alloc_size(ggml_backend_buffer_type_t buft, const ggml_tensor * tensor) {
 =======
     // See comments in init_tensor.
-    if (ggml_is_quantized(tensor->type) && (tensor->ne[0] % 512 != 0)) {
+    if (ggml_is_quantized(tensor->type) && (tensor->ne[0] % 512 != 0) && (tensor->view_src == nullptr)) {
         ggml_backend_rpc_buffer_type_context * buft_ctx = (ggml_backend_rpc_buffer_type_context *)buft->context;
         auto sock = get_socket(buft_ctx->endpoint);
 
@@ -1088,7 +1088,7 @@ bool rpc_server::init_tensor(const rpc_msg_init_tensor_req & request) {
         GGML_LOG_ERROR("Null buffer for tensor passed to init_tensor function\n");
     }
 
-    if(tensor->extra != nullptr) {
+    if (tensor->extra != nullptr) {
         // This pointer can either be passed around client/server, or probably better stored server-side and kept track of.
         // Currently unimplemented.
         GGML_LOG_ERROR("tensor->extra populated by the backend, this is currently unsupported.\n");