ollama/llm/llama.cpp/ggml_patch/0001-add-detokenize-endpoint.patch

From 032ef7ff2423f5117bb59d42fb71be9cebf0a2de Mon Sep 17 00:00:00 2001
From: Bruce MacDonald <brucewmacdonald@gmail.com>
Date: Mon, 28 Aug 2023 18:08:12 -0400
Subject: [PATCH] add detokenize endpoint

---
 examples/server/server.cpp | 21 +++++++++++++++++++++
 1 file changed, 21 insertions(+)

diff --git a/examples/server/server.cpp b/examples/server/server.cpp
index 9966045..5014691 100644
--- a/examples/server/server.cpp
+++ b/examples/server/server.cpp
@@ -1075,6 +1075,12 @@ static json format_tokenizer_response(const std::vector<llama_token> &tokens)
         {"tokens", tokens}};
 }
 
+static json format_detokenized_response(std::string content)
+{
+    return json{
+        {"content", content}};
+}
+
 static void parse_options_completion(const json &body, llama_server_context &llama)
 {
     gpt_params default_params;
@@ -1361,6 +1367,21 @@ int main(int argc, char **argv)
         const json data = format_tokenizer_response(tokens);
         return res.set_content(data.dump(), "application/json"); });
 
+    svr.Post("/detokenize", [&llama](const Request &req, Response &res)
+             {
+        auto lock = llama.lock();
+
+        const json body = json::parse(req.body);
+        std::string content;
+        if (body.count("tokens") != 0)
+        {
+            const std::vector<llama_token> tokens = body["tokens"];
+            content = tokens_to_str(llama.ctx, tokens.cbegin(), tokens.cend());
+        }
+
+        const json data = format_detokenized_response(content);
+        return res.set_content(data.dump(), "application/json"); });
+
     svr.Post("/embedding", [&llama](const Request &req, Response &res)
              {
         auto lock = llama.lock();
-- 
2.39.2 (Apple Git-143)