auto: update

Author: github-actions[bot]

README.md

@@ -1,6 +1,6 @@
 # llama.cpp - feature development info
 
-Auto-generated on 2026-06-23 05:54:34 UTC
+Auto-generated on 2026-06-23 08:58:30 UTC
 
 **Repo:** https://github.com/ggml-org/llama.cpp
 
@@ -27,6 +27,10 @@ graph BT
     C_23ee879["<a href='https://github.com/ggml-org/llama.cpp/pull/24923'>#24923</a><br/><span style='font-size:smaller;color:gray'>opencl: q8_0 gemv precision improvement (#24923)</span>"]
     C_23ee879 --> C_73618f2
     C_73618f2["<a href='https://github.com/ggml-org/llama.cpp/pull/24176'>#24176</a><br/><span style='font-size:smaller;color:gray'>server: improve user message detection and create checkpoints at every user message (#24176)</span>"]
+    C_73618f2 --> C_035cd8f
+    C_035cd8f["<a href='https://github.com/ggml-org/llama.cpp/pull/24930'>#24930</a><br/><span style='font-size:smaller;color:gray'>codeowners: add yomaytk to ggml-webgpu (#24930)</span>"]
+    C_035cd8f --> C_7c90850
+    C_7c90850["<a href='https://github.com/ggml-org/llama.cpp/pull/24811'>#24811</a><br/><span style='font-size:smaller;color:gray'>ggml-webgpu: improve MTP inference by using mat-vec path for small batches (#24811)</span>"]
     COMMON -->|dev-metal| C_bec3083
     C_bec3083["<a href='https://github.com/ggml-org/llama.cpp/pull/24021'>#24021</a><br/><span style='font-size:smaller;color:gray'>metal : per-op source split + parallel compile (#24021)</span>"]
 ```

branches.txt

@@ -3,7 +3,9 @@ Repo: https://github.com/ggml-org/llama.cpp
 Common ancestor: f8cc15f163e784c58fe13aee58ebc03055bb0c40
 Branches analyzed: 2
 
---- master (tip: 73618f27a801c0b8614ceaf3547d3c2a99baae14, commits from ancestor: 7) ---
+--- master (tip: 7c908502ea0868e6ae913f79ba84ba844a5b386a, commits from ancestor: 9) ---
+7c90850 ggml-webgpu: improve MTP inference by using mat-vec path for small batches (#24811)
+035cd8f codeowners: add yomaytk to ggml-webgpu (#24930)
 73618f2 server: improve user message detection and create checkpoints at every user message (#24176)
 23ee879 opencl: q8_0 gemv precision improvement (#24923)
 dec5ca5 server : Add id to tool call responses api (#24882)