Fix -sot: mirror output_norm to split context for all devices

Ikawrakow's review: output norm tensors remain on buft_output.buft
(the per-device offload buft) instead of being on the split context.
With -sot, the output weight IS split across devices, but the norm
weights stay on a single device and never get mirrored.

Fix: in get_context_for_tensor(), when split_output_tensor is on,
route output_norm (and output_norm_bias, output.bias) to
ctx_output_split (= split_buft context). This puts them in
split_tensors so prepare_split_tensors(-1, ...) at tensor-load
time creates per-device mirrored copies. build_output() already
handles output_norm->extra correctly: line 2107 checks for it
and line 2115 uses per-device norm copies when available.

Author: Nexesenex

src/llama-load-tensors.cpp

@@ -447,6 +447,9 @@ ggml_context * create_tensors_helper::get_context_for_tensor(ggml_context * ctx,
         if (name == "output.weight" || name == "output_extra.weight") {
             return ctx;
         }
+        if (name.find("output_norm") == 0 || name.find(".output_norm") != std::string::npos || name == "output.bias") {
+            return ctx_output_split;
+        }
     }
     for (auto & o : overrides) {
         if (std::regex_search(name, o.first)) {