Support DeepSeek-R1 Qwen (#3431)

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

gpt4all-backend/deps/llama.cpp-mainline

@@ -1 +1 @@
-Subproject commit 58a55efc4ae5dd3bc12887d47981faa7136027af
+Subproject commit b06658d366abe3cea92f4e868db72165531a74da

gpt4all-chat/CHANGELOG.md

@@ -6,6 +6,9 @@ The format is based on [Keep a Changelog](https://keepachangelog.com/en/1.1.0/).
 
 ## [Unreleased]
 
+### Added
+- Support DeepSeek-R1 Qwen models ([#3431](https://github.com/nomic-ai/gpt4all/pull/3431))
+
 ### Fixed
 - Fix regression while using localdocs with server API ([#3410](https://github.com/nomic-ai/gpt4all/pull/3410))
 - Don't show system messages in server chat view ([#3411](https://github.com/nomic-ai/gpt4all/pull/3411))

gpt4all-chat/src/chatllm.cpp

@@ -64,6 +64,14 @@ namespace ranges = std::ranges;
 //#define DEBUG
 //#define DEBUG_MODEL_LOADING
 
+static std::string jinjaGetStringArg(const jinja2::ValuesMap &args, const std::string &name)
+{
+    auto arg = args.find(name);
+    if (arg == args.end() || !arg->second.isString())
+        throw std::runtime_error(fmt::format("'{}' argument to raise_exception() must be a string", name));
+    return arg->second.asString();
+}
+
 // NOTE: not threadsafe
 static jinja2::TemplateEnv *jinjaEnv()
 {
@@ -75,27 +83,34 @@ static jinja2::TemplateEnv *jinjaEnv()
         settings.lstripBlocks = true;
         env.AddGlobal("raise_exception", jinja2::UserCallable(
             /*callable*/ [](auto &params) -> jinja2::Value {
-                auto messageArg = params.args.find("message");
+                auto message = jinjaGetStringArg(params.args, "message");
-                if (messageArg == params.args.end() || !messageArg->second.isString())
+                throw std::runtime_error(fmt::format("Jinja template error: {}", message));
-                    throw std::runtime_error("'message' argument to raise_exception() must be a string");
-                throw std::runtime_error(fmt::format("Jinja template error: {}", messageArg->second.asString()));
             },
             /*argsInfo*/ { jinja2::ArgInfo("message", /*isMandatory*/ true) }
         ));
         env.AddGlobal("strftime_now", jinja2::UserCallable(
             /*callable*/ [](auto &params) -> jinja2::Value {
                 using Clock = std::chrono::system_clock;
-                auto formatArg = params.args.find("format");
+                auto format = jinjaGetStringArg(params.args, "format");
-                if (formatArg == params.args.end() || !formatArg->second.isString())
-                    throw std::runtime_error("'format' argument to strftime_now() must be a string");
                 time_t nowUnix = Clock::to_time_t(Clock::now());
                 auto localDate = *std::localtime(&nowUnix);
                 std::ostringstream ss;
-                ss << std::put_time(&localDate, formatArg->second.asString().c_str());
+                ss << std::put_time(&localDate, format.c_str());
                 return ss.str();
             },
             /*argsInfo*/ { jinja2::ArgInfo("format", /*isMandatory*/ true) }
         ));
+        env.AddGlobal("regex_replace", jinja2::UserCallable(
+            /*callable*/ [](auto &params) -> jinja2::Value {
+                auto str     = jinjaGetStringArg(params.args, "str"    );
+                auto pattern = jinjaGetStringArg(params.args, "pattern");
+                auto repl    = jinjaGetStringArg(params.args, "repl"   );
+                return std::regex_replace(str, std::regex(pattern), repl);
+            },
+            /*argsInfo*/ { jinja2::ArgInfo("str",     /*isMandatory*/ true),
+                           jinja2::ArgInfo("pattern", /*isMandatory*/ true),
+                           jinja2::ArgInfo("repl",    /*isMandatory*/ true) }
+        ));
     }
     return &*environment;
 }

gpt4all-chat/src/jinja_replacements.cpp

@@ -52,6 +52,30 @@ const std::unordered_map<std::string_view, std::string_view> CHAT_TEMPLATE_SUBST
     {{- '<|assistant|>\n' }}
 {%- else %}
     {{- eos_token }}
+{%- endif %})TEMPLATE",
+    },
+    // DeepSeek-R1-Distill-Qwen-7B-Q4_0.gguf
+    {
+        // original
+        R"TEMPLATE({% if not add_generation_prompt is defined %}{% set add_generation_prompt = false %}{% endif %}{% set ns = namespace(is_first=false, is_tool=false, is_output_first=true, system_prompt='') %}{%- for message in messages %}{%- if message['role'] == 'system' %}{% set ns.system_prompt = message['content'] %}{%- endif %}{%- endfor %}{{bos_token}}{{ns.system_prompt}}{%- for message in messages %}{%- if message['role'] == 'user' %}{%- set ns.is_tool = false -%}{{'<｜User｜>' + message['content']}}{%- endif %}{%- if message['role'] == 'assistant' and message['content'] is none %}{%- set ns.is_tool = false -%}{%- for tool in message['tool_calls']%}{%- if not ns.is_first %}{{'<｜Assistant｜><｜tool▁calls▁begin｜><｜tool▁call▁begin｜>' + tool['type'] + '<｜tool▁sep｜>' + tool['function']['name'] + '\n' + '```json' + '\n' + tool['function']['arguments'] + '\n' + '```' + '<｜tool▁call▁end｜>'}}{%- set ns.is_first = true -%}{%- else %}{{'\n' + '<｜tool▁call▁begin｜>' + tool['type'] + '<｜tool▁sep｜>' + tool['function']['name'] + '\n' + '```json' + '\n' + tool['function']['arguments'] + '\n' + '```' + '<｜tool▁call▁end｜>'}}{{'<｜tool▁calls▁end｜><｜end▁of▁sentence｜>'}}{%- endif %}{%- endfor %}{%- endif %}{%- if message['role'] == 'assistant' and message['content'] is not none %}{%- if ns.is_tool %}{{'<｜tool▁outputs▁end｜>' + message['content'] + '<｜end▁of▁sentence｜>'}}{%- set ns.is_tool = false -%}{%- else %}{% set content = message['content'] %}{% if '</think>' in content %}{% set content = content.split('</think>')[-1] %}{% endif %}{{'<｜Assistant｜>' + content + '<｜end▁of▁sentence｜>'}}{%- endif %}{%- endif %}{%- if message['role'] == 'tool' %}{%- set ns.is_tool = true -%}{%- if ns.is_output_first %}{{'<｜tool▁outputs▁begin｜><｜tool▁output▁begin｜>' + message['content'] + '<｜tool▁output▁end｜>'}}{%- set ns.is_output_first = false %}{%- else %}{{'\n<｜tool▁output▁begin｜>' + message['content'] + '<｜tool▁output▁end｜>'}}{%- endif %}{%- endif %}{%- endfor -%}{% if ns.is_tool %}{{'<｜tool▁outputs▁end｜>'}}{% endif %}{% if add_generation_prompt and not ns.is_tool %}{{'<｜Assistant｜>'}}{% endif %})TEMPLATE",
+        // replacement
+        R"TEMPLATE({%- if not add_generation_prompt is defined %}
+    {%- set add_generation_prompt = false %}
+{%- endif %}
+{%- if messages[0]['role'] == 'system' %}
+    {{- messages[0]['content'] }}
+{%- endif %}
+{%- for message in messages %}
+    {%- if message['role'] == 'user' %}
+        {{- '<｜User｜>' + message['content'] }}
+    {%- endif %}
+    {%- if message['role'] == 'assistant' %}
+        {%- set content = message['content'] | regex_replace('^[\\s\\S]*</think>', '') %}
+        {{- '<｜Assistant｜>' + content + '<｜end▁of▁sentence｜>' }}
+    {%- endif %}
+{%- endfor -%}
+{%- if add_generation_prompt %}
+    {{- '<｜Assistant｜>' }}
 {%- endif %})TEMPLATE",
     },
     // gemma-2-9b-it-Q4_0.gguf (nomic-ai/gpt4all#3282)