chat: cut v3.3.1 release

Signed-off-by: Jared Van Bortel <jared@nomic.ai>

gpt4all-chat/CHANGELOG.md

@@ -4,6 +4,12 @@ All notable changes to this project will be documented in this file.
 
 The format is based on [Keep a Changelog](https://keepachangelog.com/en/1.1.0/).
 
+## [3.3.1] - 2024-09-26
+
+### Fixed
+- Fix a crash when attempting to continue a chat loaded from disk ([#2995](https://github.com/nomic-ai/gpt4all/pull/2995))
+- Fix the local server rejecting min\_p/top\_p less than 1 ([#2996](https://github.com/nomic-ai/gpt4all/pull/2996))
+
 ## [3.3.0] - 2024-09-20
 
 ### Added
@@ -119,6 +125,7 @@ The format is based on [Keep a Changelog](https://keepachangelog.com/en/1.1.0/).
 - Fix several Vulkan resource management issues ([#2694](https://github.com/nomic-ai/gpt4all/pull/2694))
 - Fix crash/hang when some models stop generating, by showing special tokens ([#2701](https://github.com/nomic-ai/gpt4all/pull/2701))
 
+[3.3.1]: https://github.com/nomic-ai/gpt4all/compare/v3.3.0...v3.3.1
 [3.3.0]: https://github.com/nomic-ai/gpt4all/compare/v3.2.1...v3.3.0
 [3.2.1]: https://github.com/nomic-ai/gpt4all/compare/v3.2.0...v3.2.1
 [3.2.0]: https://github.com/nomic-ai/gpt4all/compare/v3.1.1...v3.2.0

gpt4all-chat/CMakeLists.txt

@@ -2,7 +2,7 @@ cmake_minimum_required(VERSION 3.25)  # for try_compile SOURCE_FROM_VAR
 
 set(APP_VERSION_MAJOR 3)
 set(APP_VERSION_MINOR 3)
-set(APP_VERSION_PATCH 0)
+set(APP_VERSION_PATCH 1)
 set(APP_VERSION_BASE "${APP_VERSION_MAJOR}.${APP_VERSION_MINOR}.${APP_VERSION_PATCH}")
 set(APP_VERSION "${APP_VERSION_BASE}")
 

gpt4all-chat/metadata/latestnews.md

@@ -1,10 +1,9 @@
 ## Latest News
 
-Version 3.2.1 has now been released which fixes an issue with poor quality responses on NVIDIA GPUs in 3.2.0. The new 3.2 minor version brings:
+GPT4All v3.3.0 was released on September 23rd. Changes include:
 
-* **Official Language Translations**: Translations for Simplified Chinese, Traditional Chinese, Italian, Portuguese, Romanian, and Spanish.<br/>
-  Go to Settings > Language and Locale to change the application language.
-* **Context Window Improvements**: Significantly faster context recalculation when context runs out
-* **Bugfixes**: Models no longer stop generating when they run out of context
-
-Also, Qwen2-1.5B-Instruct was recently added to the model list, which has good Chinese support.
+* **UI Improvements**: The minimum window size now adapts to the font size. A few labels and links have been fixed. The Embeddings Device selection of "Auto"/"Application default" works again. The window icon is now set on Linux. The antenna icon now displays when the API server is listening.
+* **Single Instance**: Only one instance of GPT4All can be opened at a time. This is now enforced.
+* **Greedy Sampling**: Set temperature to zero to enable greedy sampling.
+* **API Server Changes**: The built-in API server now responds correctly to both legacy completions, and chats with message history. Also, it now uses the system prompt configured in the UI.
+* **Translation Improvements**: The Italian, Romanian, and Traditional Chinese translations have been updated.

gpt4all-chat/metadata/release.json

@@ -203,5 +203,10 @@
     "version": "3.2.1",
     "notes": "<b>&mdash; Fixes &mdash;</b>\n* Fix a potential Vulkan crash on application exit on some Linux systems\n* Fix a bad CUDA build option that led to gibberish on newer NVIDIA GPUs\n",
     "contributors": "* Jared Van Bortel (Nomic AI)"
+  },
+  {
+    "version": "3.3.0",
+    "notes": "* **UI Improvements**: The minimum window size now adapts to the font size. A few labels and links have been fixed. The Embeddings Device selection of \"Auto\"/\"Application default\" works again. The window icon is now set on Linux. The antenna icon now displays when the API server is listening.\n* **Single Instance**: Only one instance of GPT4All can be opened at a time. This is now enforced.\n* **Greedy Sampling**: Set temperature to zero to enable greedy sampling.\n* **API Server Changes**: The built-in API server now responds correctly to both legacy completions, and chats with message history. Also, it now uses the system prompt configured in the UI.\n* **Translation Improvements**: The Italian, Romanian, and Traditional Chinese translations have been updated.\n",
+    "contributors": "* Jared Van Bortel (Nomic AI)\n* Adam Treat (Nomic AI)\n* 3Simplex (`@3Simplex`)\n* Riccardo Giovanetti (`@Harvester62`)\n* Victor Emanuel (`@SINAPSA-IC`)\n* Dominik (`@cosmic-snow`)\n* Shiranui (`@supersonictw`)"
   }
 ]

gpt4all-chat/src/chatllm.cpp

@@ -1314,8 +1314,16 @@ void ChatLLM::processRestoreStateFromText()
         auto &response = *it++;
         Q_ASSERT(response.first != "Prompt: ");
 
-        m_llModelInfo.model->prompt(prompt.second.toStdString(), promptTemplate.toStdString(), promptFunc, nullptr,
-                                    /*allowContextShift*/ true, m_ctx, false, response.second.toUtf8().constData());
+        // FIXME(jared): this doesn't work well with the "regenerate" button since we are not incrementing
+        //               m_promptTokens or m_promptResponseTokens
+        m_llModelInfo.model->prompt(
+            prompt.second.toStdString(), promptTemplate.toStdString(),
+            promptFunc, /*responseFunc*/ [](auto &&...) { return true; },
+            /*allowContextShift*/ true,
+            m_ctx,
+            /*special*/ false,
+            response.second.toUtf8().constData()
+        );
     }
 
     if (!m_stopGenerating) {

gpt4all-chat/src/server.cpp

@@ -193,11 +193,11 @@ protected:
         if (!value.isNull())
             this->temperature = float(value.toDouble());
 
-        value = reqValue("top_p", Number, /*min*/ 0, /*max*/ 1);
+        value = reqValue("top_p", Number, false, /*min*/ 0, /*max*/ 1);
         if (!value.isNull())
             this->top_p = float(value.toDouble());
 
-        value = reqValue("min_p", Number, /*min*/ 0, /*max*/ 1);
+        value = reqValue("min_p", Number, false, /*min*/ 0, /*max*/ 1);
         if (!value.isNull())
             this->min_p = float(value.toDouble());