Merge pull request ggml-org#9 from anon998/stopping-strings

Fix stopping strings.

examples/server/server.cpp

@@ -20,6 +20,33 @@ static size_t common_part(const std::vector<llama_token> & a, const std::vector<
   return i;
 }
 
+enum stop_type {
+    STOP_FULL,
+    STOP_PARTIAL,
+};
+
+bool ends_with(const std::string &str, const std::string &suffix)
+{
+    return str.size() >= suffix.size() &&
+           0 == str.compare(str.size() - suffix.size(), suffix.size(), suffix);
+}
+
+size_t find_partial_stop_string(const std::string &stop, const std::string &text)
+{
+    if (!text.empty()) {
+        const char text_last_char = text.back();
+        for (int64_t char_index = stop.size() - 1; char_index >= 0; char_index--) {
+            if (stop[char_index] == text_last_char) {
+                const std::string current_partial = stop.substr(0, char_index + 1);
+                if (ends_with(text, current_partial)) {
+                    return text.size() - char_index - 1;
+                }
+            }
+        }
+    }
+    return std::string::npos;
+}
+
 struct llama_server_context
 {
   bool stream = false;
@@ -248,6 +275,31 @@ struct llama_server_context
     return result;
   }
 
+  size_t findStoppingStrings(const std::string &text, const size_t last_token_size,
+                             const stop_type type)
+  {
+    size_t stop_pos = std::string::npos;
+    for (const std::string &word : params.antiprompt) {
+        size_t pos;
+        if (type == STOP_FULL) {
+            const size_t tmp = word.size() + last_token_size;
+            const size_t from_pos = text.size() > tmp ? text.size() - tmp : 0;
+            pos = text.find(word, from_pos);
+        } else {
+            pos = find_partial_stop_string(word, text);
+        }
+        if (pos != std::string::npos &&
+            (stop_pos == std::string::npos || pos < stop_pos)) {
+            if (type == STOP_FULL) {
+                stopping_word = word;
+                has_next_token = false;
+            }
+            stop_pos = pos;
+        }
+    }
+    return stop_pos;
+  }
+
   std::string doCompletion()
   {
     llama_token token = nextToken();
@@ -272,16 +324,6 @@ struct llama_server_context
               stopping_word.c_str());
     }
 
-    for (const std::string& word : params.antiprompt) {
-      size_t i = generated_text.find(word, generated_text.size() - (word.size() + token_text.size()));
-      if (i != std::string::npos) {
-        generated_text.erase(generated_text.begin() + i, generated_text.end());
-        stopping_word = word;
-        has_next_token = false;
-        break;
-      }
-    }
-
     return token_text;
   }
 
@@ -711,7 +753,14 @@ int main(int argc, char **argv)
 
       if (!llama.stream) {
           while (llama.has_next_token) {
-              llama.doCompletion();
+              const std::string token_text = llama.doCompletion();
+              const size_t stop_pos = llama.findStoppingStrings(
+                  llama.generated_text, token_text.size(), STOP_FULL);
+
+              if (stop_pos != std::string::npos) {
+                  llama.generated_text.erase(llama.generated_text.begin() + stop_pos,
+                                             llama.generated_text.end());
+              }
           }
 
           json data = {{"content", llama.generated_text},
@@ -724,7 +773,7 @@ int main(int argc, char **argv)
 
           llama_print_timings(llama.ctx);
 
-          return res.set_content(
+          res.set_content(
               data.dump(llama.json_indent, ' ', false, json::error_handler_t::replace),
               "application/json");
       } else {
@@ -733,7 +782,7 @@ int main(int argc, char **argv)
               int32_t multibyte_pending = 0;
 
               while (llama.has_next_token) {
-                  std::string token_text = llama.doCompletion();
+                  const std::string token_text = llama.doCompletion();
 
                   if (multibyte_pending > 0) {
                       multibyte_pending -= token_text.size();
@@ -761,8 +810,22 @@ int main(int argc, char **argv)
                       continue;
                   }
 
-                  const size_t pos = std::min(sent_count, llama.generated_text.size());
-                  std::string to_send = llama.generated_text.substr(pos);
+                  size_t pos = std::min(sent_count, llama.generated_text.size());
+
+                  const char *str_test = llama.generated_text.c_str() + pos;
+                  size_t stop_pos =
+                      llama.findStoppingStrings(str_test, token_text.size(), STOP_FULL);
+                  if (stop_pos != std::string::npos) {
+                      llama.generated_text.erase(
+                          llama.generated_text.begin() + pos + stop_pos,
+                          llama.generated_text.end());
+                      pos = std::min(sent_count, llama.generated_text.size());
+                  } else {
+                      stop_pos = llama.findStoppingStrings(str_test, token_text.size(),
+                                                           STOP_PARTIAL);
+                  }
+
+                  std::string to_send = llama.generated_text.substr(pos, stop_pos);
                   sent_count += to_send.size();
 
                   json data;
@@ -808,7 +871,6 @@ int main(int argc, char **argv)
       }
   });
 
-
   svr.Post("/tokenize", [&llama](const Request &req, Response &res)
             {
               json body = json::parse(req.body);