Mistral 3.1 chat template for llama.cpp, including tools and support for multiple images. modify from vllm https://github.com/vllm-project/vllm/blob/main/examples/tool_chat_template_mistral.jinja

compose.yaml

services:
  llm:
    image: ghcr.io/ggml-org/llama.cpp:server-cuda-b5391
    ports:
      - "7071:8080"
    volumes:
      - ./models:/models
    env_file:
      - llm.env
    deploy:
      resources:
        reservations:
          devices:
            - driver: nvidia
              count: 1
              capabilities: [gpu]

llm.env

LLAMA_ARG_HOST=0.0.0.0

LLAMA_ARG_MODEL=/models/unsloth_Mistral-Small-3.1-24B-Instruct-2503-Q4_K_M.gguf
LLAMA_ARG_ALIAS=mistral-small3.1:24b-instruct-2503-q4_K_M

LLAMA_ARG_MMPROJ=/models/mmproj-unsloth_Mistral-Small-3.1-24B-Instruct-2503-f16.gguf
LLAMA_ARG_NO_MMPROJ_OFFLOAD=0

LLAMA_ARG_N_GPU_LAYERS=99
LLAMA_ARG_FLASH_ATTN=1
LLAMA_ARG_CTX_SIZE=8192
LLAMA_ARG_N_PARALLEL=1
LLAMA_ARG_CACHE_REUSE=256

LLAMA_ARG_CACHE_TYPE_K=q8_0
LLAMA_ARG_CACHE_TYPE_V=q8_0

LLAMA_ARG_JINJA=1
LLAMA_ARG_CHAT_TEMPLATE_FILE=/models/mistral.jinja

LLAMA_ARG_ENDPOINT_METRICS=1

mistral.jinja

{%- macro render_message_content(content) -%}
    {%- if content is string -%}
        {{ content }}
    {%- else -%}
        {%- for block in content -%}
            {%- if block["type"] == "text" -%}
                {{ block["text"] }}
            {%- elif block["type"] in ["image", "image_url"] -%}
                [IMG]
            {%- else -%}
                {{ raise_exception("Only text and image blocks are supported in message content!") }}
            {%- endif -%}
        {%- endfor -%}
    {%- endif -%}
{%- endmacro -%}

{%- macro validate_tool_call_id(id, context="") -%}
    {%- if not id is defined or id|length < 9 -%}
        {{ raise_exception("Tool call IDs should be alphanumeric strings with length >= 9! (" + context + ")") }}
    {%- endif -%}
    {{ id[-9:] }}
{%- endmacro -%}

{%- if messages[0]["role"] == "system" %}
    {%- if messages[0]["content"] is string %}
        {%- set system_message = messages[0]["content"] %}
    {%- else %}
        {%- set system_message = messages[0]["content"][0]["text"] %}
    {%- endif %}
    {%- set loop_messages = messages[1:] %}
{%- else %}
    {%- set loop_messages = messages %}
{%- endif %}
{%- if not tools is defined %}
    {%- set tools = none %}
{%- endif %}
{%- set user_messages = loop_messages | selectattr("role", "equalto", "user") | list %}

{%- set has_no_tool_calls = [] %}
{%- for message in loop_messages | rejectattr("role", "equalto", "tool") | rejectattr("role", "equalto", "tool_results") %}
    {%- if "tool_calls" not in message %}
        {%- do has_no_tool_calls.append(message) %}
    {%- endif %}
{%- endfor %}

{%- for message in has_no_tool_calls %}
    {%- if (message["role"] == "user") != (loop.index0 % 2 == 0) %}
        {{- raise_exception("After the optional system message, conversation roles must alternate user/assistant/user/assistant/...") }}
    {%- endif %}
{%- endfor %}

{{- bos_token }}
{%- for message in loop_messages %}
    {%- if message["role"] == "user" %}
        {%- if tools is not none and (message == user_messages[-1]) %}
            {{- "[AVAILABLE_TOOLS] [" }}
            {%- for tool in tools %}
                {%- set tool = tool.function %}
                {{- '{"type": "function", "function": {' }}
                {%- for key, val in tool.items() if key != "return" %}
                    {%- if val is string %}
                        {{- '"' + key + '": "' + val + '"' }}
                    {%- else %}
                        {{- '"' + key + '": ' + val|tojson }}
                    {%- endif %}
                    {%- if not loop.last %}
                        {{- ", " }}
                    {%- endif %}
                {%- endfor %}
                {{- "}}" }}
                {%- if not loop.last %}
                    {{- ", " }}
                {%- else %}
                    {{- "]" }}
                {%- endif %}
            {%- endfor %}
            {{- "[/AVAILABLE_TOOLS]" }}
        {%- endif %}
        
        {{- "[INST] " }}
        {%- if loop.last and system_message is defined %}
            {{- system_message + "\n\n" }}
        {%- endif %}
        {{- render_message_content(message["content"]) }}
        {{- "[/INST]" }}
        
    {%- elif message["role"] == "tool_calls" or message.tool_calls is defined %}
        {%- if message.tool_calls is defined %}
            {%- set tool_calls = message.tool_calls %}
        {%- else %}
            {%- set tool_calls = message.content %}
        {%- endif %}
        {{- "[TOOL_CALLS] [" }}
        {%- for tool_call in tool_calls %}
            {%- set out = tool_call.function|tojson %}
            {{- out[:-1] }}
            {{- ', "id": "' + validate_tool_call_id(tool_call.id, "1") + '"}' }}
            {%- if not loop.last %}
                {{- ", " }}
            {%- else %}
                {{- "]" + eos_token }}
            {%- endif %}
        {%- endfor %}
        
    {%- elif message["role"] == "assistant" %}
        {{- " " + render_message_content(message["content"]) + eos_token }}
        
    {%- elif message["role"] == "tool_results" or message["role"] == "tool" %}
        {%- if message.content is defined and message.content.content is defined %}
            {%- set content = message.content.content %}
        {%- else %}
            {%- set content = message.content %}
        {%- endif %}
        {{- '[TOOL_RESULTS] {"content": ' + content|string + ", " }}
        {{- '"call_id": "' + validate_tool_call_id(message.tool_call_id, "2") + '"}[/TOOL_RESULTS]' }}
        
    {%- else %}
        {{- raise_exception("Only user and assistant roles are supported, with the exception of an initial optional system message!") }}
    {%- endif %}
{%- endfor %}