server : support slot save/restore with media inputs (#26640)

* server : save serialized image chunks at the end of the llama state * server : support multimodal slot state save/restore with packed payload * server : refine image slot state serialization * server : support media slot state and centralize media validation * server : remove unnecessary comment * server : remove defensive media checks and move the chunk type check to validate()
2026-08-12 12:20:28 +02:00
parent 4dd127584b
commit 5d9e5ac30e
7 changed files with 593 additions and 78 deletions
@@ -883,6 +883,7 @@ extern "C" {
               const llama_token * tokens,
                          size_t   n_token_count);

+    // If tokens_out is NULL, only the token count is reported through n_token_count_out and no state is loaded
    LLAMA_API size_t llama_state_seq_load_file(
            struct llama_context * ctx,
                      const char * filepath,