unsloth

/

Devstral-Small-2507-GGUF

like 49

Follow

Unsloth AI 7.81k

@squid2 We can investigate but Mistral's official gguf doesnt support tool-calling so unsure how youre getting results with it.

We tested every scenario for tool calling and it worked perfectly so unsure why you're not having the same experience

Where are you using this btw? llama.cpp, lmstudio?

I'm using llama.cpp (with llama-server). Even though I didn't see any mentions of tools in the chat template for Mistral's official Devstral GGUF, tool calling worked perfectly out of the box with it. Here are the version details and the command I used to launch llama-server.

sultan@Sultans-MacBook-Pro llm_scripts % llama-server --version
version: 5840 (75c91de6)
built with Apple clang version 17.0.0 (clang-1700.0.13.3) for arm64-apple-darwin24.4.0

sultan@Sultans-MacBook-Pro llm_scripts % cat devstral_unsloth.sh
llama-server -hf unsloth/Devstral-Small-2507-GGUF:IQ4_XS --temp 0.15 -fa -ctk q8_0 -ctv q8_0 -c 262144 -np 2 --jinja

sultan@Sultans-MacBook-Pro llm_scripts % cat devstral.sh 
llama-server -hf mistralai/Devstral-Small-2507_gguf:Q4_K_M --temp 0.15 -fa -ctk q8_0 -ctv q8_0 -c 262144 -np 2 --jinja

It looks like vllm, and a parser I'm working on, was using the existing format of [TOOL_CALLS][ <valid JSON of tool calls> ] where the new template uses {{- "[TOOL_CALLS]" + tool.function.name + "[ARGS]" + arguments }}

https://github.com/vllm-project/vllm/blob/e2de455c349df8385b18fe447beb6325dcb6af9c/vllm/entrypoints/openai/tool_parsers/mistral_tool_parser.py#L78

The Ollama template for 2505 was also the following:

{{- else if .ToolCalls }}[TOOL_CALLS][
{{- range .ToolCalls }}{"name": "{{ .Function.Name }}", "arguments": {{ .Function.Arguments }}}
{{- end }}]

On the latest build of llama.cpp, the tool calling chain breaks after several calls, just as described earlier. If the new tool call template really is the intended one - oddly, it lacks closing tags for [TOOL_CALLS] and [ARGS] - then the root cause is likely that llama.cpp still relies on the older format internally:

common/chat.cpp

1810     // Mistral Nemo (w/ tools)
1811     if (src.find("[TOOL_CALLS]") != std::string::npos) {
1812         return common_chat_params_init_mistral_nemo(tmpl, params);
1813     }  

I haven’t really dug into this, but here's my quick-fix Jinja template that keeps the legacy tool calls format. It works in llama.cpp, but if the new tool call template is the intended approach, this workaround could hurt model performance.

You can provide it to llama.cpp using the --chat-template-file argument or LLAMA_ARG_CHAT_TEMPLATE_FILE environment variable.

{%- set default_system_message = 'You are Devstral, a helpful agentic model trained by Mistral AI and using the OpenHands scaffold. You can interact with a computer to solve tasks.\n\n<ROLE>\nYour primary role is to assist users by executing commands, modifying code, and solving technical problems effectively. You should be thorough, methodical, and prioritize quality over speed.\n* If the user asks a question, like \"why is X happening\", don\'t try to fix the problem. Just give an answer to the question.\n</ROLE>\n\n<EFFICIENCY>\n* Each action you take is somewhat expensive. Wherever possible, combine multiple actions into a single action, e.g. combine multiple bash commands into one, using sed and grep to edit/view multiple files at once.\n* When exploring the codebase, use efficient tools like find, grep, and git commands with appropriate filters to minimize unnecessary operations.\n</EFFICIENCY>\n\n<FILE_SYSTEM_GUIDELINES>\n* When a user provides a file path, do NOT assume it\'s relative to the current working directory. First explore the file system to locate the file before working on it.\n* If asked to edit a file, edit the file directly, rather than creating a new file with a different filename.\n* For global search-and-replace operations, consider using `sed` instead of opening file editors multiple times.\n</FILE_SYSTEM_GUIDELINES>\n\n<CODE_QUALITY>\n* Write clean, efficient code with minimal comments. Avoid redundancy in comments: Do not repeat information that can be easily inferred from the code itself.\n* When implementing solutions, focus on making the minimal changes needed to solve the problem.\n* Before implementing any changes, first thoroughly understand the codebase through exploration.\n* If you are adding a lot of code to a function or file, consider splitting the function or file into smaller pieces when appropriate.\n</CODE_QUALITY>\n\n<VERSION_CONTROL>\n* When configuring git credentials, use \"openhands\" as the user.name and \"[email protected]\" as the user.email by default, unless explicitly instructed otherwise.\n* Exercise caution with git operations. Do NOT make potentially dangerous changes (e.g., pushing to main, deleting repositories) unless explicitly asked to do so.\n* When committing changes, use `git status` to see all modified files, and stage all files necessary for the commit. Use `git commit -a` whenever possible.\n* Do NOT commit files that typically shouldn\'t go into version control (e.g., node_modules/, .env files, build directories, cache files, large binaries) unless explicitly instructed by the user.\n* If unsure about committing certain files, check for the presence of .gitignore files or ask the user for clarification.\n</VERSION_CONTROL>\n\n<PULL_REQUESTS>\n* When creating pull requests, create only ONE per session/issue unless explicitly instructed otherwise.\n* When working with an existing PR, update it with new commits rather than creating additional PRs for the same issue.\n* When updating a PR, preserve the original PR title and purpose, updating description only when necessary.\n</PULL_REQUESTS>\n\n<PROBLEM_SOLVING_WORKFLOW>\n1. EXPLORATION: Thoroughly explore relevant files and understand the context before proposing solutions\n2. ANALYSIS: Consider multiple approaches and select the most promising one\n3. TESTING:\n   * For bug fixes: Create tests to verify issues before implementing fixes\n   * For new features: Consider test-driven development when appropriate\n   * If the repository lacks testing infrastructure and implementing tests would require extensive setup, consult with the user before investing time in building testing infrastructure\n   * If the environment is not set up to run tests, consult with the user first before investing time to install all dependencies\n4. IMPLEMENTATION: Make focused, minimal changes to address the problem\n5. VERIFICATION: If the environment is set up to run tests, test your implementation thoroughly, including edge cases. If the environment is not set up to run tests, consult with the user first before investing time to run tests.\n</PROBLEM_SOLVING_WORKFLOW>\n\n<SECURITY>\n* Only use GITHUB_TOKEN and other credentials in ways the user has explicitly requested and would expect.\n* Use APIs to work with GitHub or other platforms, unless the user asks otherwise or your task requires browsing.\n</SECURITY>\n\n<ENVIRONMENT_SETUP>\n* When user asks you to run an application, don\'t stop if the application is not installed. Instead, please install the application and run the command again.\n* If you encounter missing dependencies:\n  1. First, look around in the repository for existing dependency files (requirements.txt, pyproject.toml, package.json, Gemfile, etc.)\n  2. If dependency files exist, use them to install all dependencies at once (e.g., `pip install -r requirements.txt`, `npm install`, etc.)\n  3. Only install individual packages directly if no dependency files are found or if only specific packages are needed\n* Similarly, if you encounter missing dependencies for essential tools requested by the user, install them when possible.\n</ENVIRONMENT_SETUP>\n\n<TROUBLESHOOTING>\n* If you\'ve made repeated attempts to solve a problem but tests still fail or the user reports it\'s still broken:\n  1. Step back and reflect on 5-7 different possible sources of the problem\n  2. Assess the likelihood of each possible cause\n  3. Methodically address the most likely causes, starting with the highest probability\n  4. Document your reasoning process\n* When you run into any major issue while executing a plan from the user, please don\'t try to directly work around it. Instead, propose a new plan and confirm with the user before proceeding.\n</TROUBLESHOOTING>' %}

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

{#- This block checks for alternating user/assistant messages, skipping tool calling messages #}
{%- set ns = namespace() %}
{%- set ns.index = 0 %}
{%- for message in loop_messages %}
    {%- if not (message.role == "tool" or message.role == "tool_results" or (message.tool_calls is defined and message.tool_calls is not none)) %}
        {%- if (message["role"] == "user") != (ns.index % 2 == 0) %}
            {{- raise_exception("After the optional system message, conversation roles must alternate user/assistant/user/assistant/...") }}
        {%- endif %}
        {%- set ns.index = ns.index + 1 %}
    {%- endif %}
{%- endfor %}

{{- bos_token }}
{%- if system_message is defined %}
{{- '[SYSTEM_PROMPT]' + system_message + '[/SYSTEM_PROMPT]' }}
{%- endif %}


{%- if tools is defined and tools is not none and tools|length > 0 %}

    {%- set has_tools = true %}
    {%- set tools_description = "[AVAILABLE_TOOLS]" + (tools | tojson) + "[/AVAILABLE_TOOLS]" %}

    {{- tools_description }}

{%- endif %}

{%- for message in loop_messages %}
    {%- if message["role"] == "user" %}
    {%- if message["content"] is string %}
            {{- "[INST]" + message["content"] + "[/INST]" }}
        {%- endif %}
    {%- elif (message.tool_calls is defined and message.tool_calls is not none) %}
        {{- "[TOOL_CALLS][" }}
        {%- for tool_call in message.tool_calls %}
            {%- set out = tool_call.function|tojson %}
            {{- out[:-1] }}
            {{- ', "id": "' + tool_call.id + '"}' }}
            {%- if not loop.last %}
                {{- ", " }}
            {%- else %}
                {{- "]" + eos_token }}
            {%- endif %}
        {%- endfor %}
    {%- elif message["role"] == "assistant" %}
        {{- 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": "' + message.tool_call_id + '"}[/TOOL_RESULTS]' }}
    {%- else %}
        {{- raise_exception("Only user and assistant roles are supported, with the exception of an initial optional system message!") }}
    {%- endif %}
{%- endfor %}

Same here.

Much apologies on the delay, but I rechecked using mistral-common and the GGUF - all tests and tool calling seem to be exactly equivalent - it is in fact possible llama.cpp themselves have not used the latest Devstral tool calling approach, since it is different.

@danielhanchen I am using the UD Q6_K_XL variant, and tool calling under Kilo code (Roocode fork) is also failing. It's outputting the following text:

Let's start by examining the TrainingPathCertificationsForm.php file to understand how the certification type is saved during editing:read_file>


app/Modules/TrainingPaths/Details/TrainingPathCertificationsForm.php


</read_file>

I am using the --jinja template through llama.cpp. Any ideas why this is happening?

@danielhanchen Interestingly, tool calling works fine if I remove --jinja. So something seems broken there!