Display name of model at top of web ui

llamafile/server/flagz.cpp

@@ -18,6 +18,7 @@
 #include "client.h"
 #include "llama.cpp/llama.h"
 #include "llamafile/llamafile.h"
+#include "llamafile/string.h"
 #include "llamafile/server/json.h"
 
 namespace lf {
@@ -27,6 +28,7 @@ bool
 Client::flagz()
 {
     jt::Json json;
+    json["model"] = stripext(basename(FLAG_model));
     json["prompt"] = FLAG_prompt;
     json["no_display_prompt"] = FLAG_no_display_prompt;
     json["nologo"] = FLAG_nologo;

llamafile/server/worker.cpp

@@ -24,6 +24,7 @@
 #include "llamafile/server/tokenbucket.h"
 #include "llamafile/threadlocal.h"
 #include "llamafile/trust.h"
+#include <cosmo.h>
 #include <cassert>
 #include <exception>
 #include <pthread.h>
@@ -109,7 +110,11 @@ void
 Worker::handle()
 {
     if ((client_.fd_ = server_->accept(&client_.client_ip_)) == -1) {
-        SLOG("accept returned %m");
+        if (IsWindows() && errno == ENOTSOCK) {
+            // Server::shutdown() calls close() on the listening socket
+        } else {
+            SLOG("accept returned %m");
+        }
         return;
     }
 

llamafile/server/www/chatbot.js

@@ -21,6 +21,7 @@ const DEFAULT_SYSTEM_PROMPT =
       "detailed, and polite answers to the human's questions.";
 
 const DEFAULT_FLAGZ = {
+  "model": null,
   "prompt": null,
   "no_display_prompt": false,
   "frequency_penalty": 0,
@@ -342,6 +343,12 @@ function getSystemPrompt() {
   return prompt;
 }
 
+function updateModelInfo() {
+  if (flagz.model) {
+    document.getElementById("model").textContent = flagz.model;
+  }
+}
+
 function startChat(history) {
   chatHistory = history;
   chatMessages.innerHTML = "";
@@ -356,6 +363,7 @@ function startChat(history) {
 
 async function chatbot() {
   flagz = await fetchFlagz();
+  updateModelInfo();
   startChat([{ role: "system", content: getSystemPrompt() }]);
   sendButton.addEventListener("click", sendMessage);
   stopButton.addEventListener("click", stopMessage);

llamafile/server/www/index.html

@@ -8,7 +8,10 @@
 
 <div class="chat-container">
   <div class="chat-header">
-    <h1><img src="chatbot.png" alt="[logo]" width="64" height="64"> llamafile</h1>
+    <h1>
+      <img src="chatbot.png" alt="[llamafile]" title="llamafile" width="64" height="64">
+      <span id="model">llamafile</span>
+    </h1>
   </div>
   <div class="chat-messages" id="chat-messages">
     <div class="message system">

llamafile/string.cpp

@@ -102,6 +102,14 @@ std::string basename(const std::string_view &path) {
     }
 }
 
+std::string stripext(const std::string &path) {
+    size_t i = path.size();
+    while (i--)
+        if (path[i] == '.')
+            return path.substr(0, i);
+    return path;
+}
+
 std::string_view extname(const std::string_view &path) {
     size_t i = path.size();
     while (i--)

llamafile/string.h

@@ -33,6 +33,7 @@ std::string format(const char *, ...) __attribute__((format(printf, 1, 2)));
 std::string iso8601(struct timespec);
 std::string join(const std::vector<std::string> &, const std::string_view &);
 std::string resolve(const std::string_view &, const std::string_view &);
+std::string stripext(const std::string &);
 std::string tolower(const std::string_view &);
 std::string_view extname(const std::string_view &);
 void append_wchar(std::string *, wchar_t);