chat_template.jinja

{{ bos_token }}{% if messages[0].role == 'system' %}
{{ raise_exception('System message is not supported in gemma; merge the system prompt with the first user message') }}
{% endif %}
{% if not tools %}
{{ raise_exception('You should provide tools; this model was only trained for tool calling with reasoning') }}
{% endif %}
{% set first_message = True %}
{% for message in messages %}
  {% if first_message and tools %}
{% raw -%}
<start_of_turn>human
You are a function calling AI model. You are provided with function signatures within <tools></tools> XML tags.You may call one or more functions to assist with the user query. Don't make assumptions about what values to plug into functions.Here are the available tools:<tools> {% endraw %}{{ tools }}{% raw -%} </tools>Use the following pydantic model json schema for each tool call you will make: {'title': 'FunctionCall', 'type': 'object', 'properties': {'arguments': {'title': 'Arguments', 'type': 'object'}, 'name': {'title': 'Name', 'type': 'string'}}, 'required': ['arguments', 'name']}For each function call return a json object with function name and arguments within <tool_call></tool_call> XML tags as follows:
<tool_call>
{tool_call}
</tool_call>Also, before making a call to a function take the time to plan the function to take. Make that thinking process between <think>{your thought}</think>

{% endraw %}{{ message.content }}{% raw -%}
<eos_turn><eos>
<start_of_turn>model
{% endraw -%}
    {% set first_message = False %}
  {% endif %}
{% endfor %}