Eval bug: smolvlm templates are not parsed correctly

[cchadowitz]

Name and Version

$ ./llama-server --version
ggml_metal_device_init: tensor API disabled for pre-M5 and pre-A19 devices
ggml_metal_library_init: using embedded metal library
ggml_metal_library_init: loaded in 7.150 sec
ggml_metal_device_init: GPU name: Apple M2 Pro
ggml_metal_device_init: GPU family: MTLGPUFamilyApple8 (1008)
ggml_metal_device_init: GPU family: MTLGPUFamilyCommon3 (3003)
ggml_metal_device_init: GPU family: MTLGPUFamilyMetal3 (5001)
ggml_metal_device_init: simdgroup reduction = true
ggml_metal_device_init: simdgroup matrix mul. = true
ggml_metal_device_init: has unified memory = true
ggml_metal_device_init: has bfloat = true
ggml_metal_device_init: has tensor = false
ggml_metal_device_init: use residency sets = true
ggml_metal_device_init: use shared buffers = true
ggml_metal_device_init: recommendedMaxWorkingSetSize = 22906.50 MB
register_backend: registered backend Metal (1 devices)
register_device: registered device Metal (Apple M2 Pro)
register_backend: registered backend BLAS (1 devices)
register_device: registered device BLAS (Accelerate)
register_backend: registered backend CPU (1 devices)
register_device: registered device CPU (Apple M2 Pro)
version: 7169 (7cba58b)
built with Apple clang version 17.0.0 (clang-1700.4.4.1) for arm64-apple-darwin24.6.0

Operating systems

Mac

GGML backends

Metal

Hardware

Macbook Pro 16" 2023

Models

ggml-org/SmolVLM-500M-Instruct

Problem description & steps to reproduce

When I run llama-server with the above model, it uses the (new) default of enabling jinja (this became default in #17524). This results in it falling back to chatml which leads to worse output.

First Bad Commit

#17524

Relevant log output

Failed to generate tool call example: Value is not callable: null at row 1, column 72:
<|im_start|>{% for message in messages %}{{message['role'] | capitalize}}{% if message['content'][0]['type'] == 'image' %}{{':'}}{% else %}{{': '}}{% endif %}{% for line in message['content'] %}{% if line['type'] == 'text' %}{{line['text']}}{% elif line['type'] == 'image' %}{{ '<image>' }}{% endif %}{% endfor %}<end_of_utterance>
                                                                       ^
{% endfor %}{% if add_generation_prompt %}{{ 'Assistant:' }}{% endif %}
 at row 1, column 42:
<|im_start|>{% for message in messages %}{{message['role'] | capitalize}}{% if message['content'][0]['type'] == 'image' %}{{':'}}{% else %}{{': '}}{% endif %}{% for line in message['content'] %}{% if line['type'] == 'text' %}{{line['text']}}{% elif line['type'] == 'image' %}{{ '<image>' }}{% endif %}{% endfor %}<end_of_utterance>
                                         ^
{% endfor %}{% if add_generation_prompt %}{{ 'Assistant:' }}{% endif %}
 at row 1, column 42:
<|im_start|>{% for message in messages %}{{message['role'] | capitalize}}{% if message['content'][0]['type'] == 'image' %}{{':'}}{% else %}{{': '}}{% endif %}{% for line in message['content'] %}{% if line['type'] == 'text' %}{{line['text']}}{% elif line['type'] == 'image' %}{{ '<image>' }}{% endif %}{% endfor %}<end_of_utterance>
                                         ^
{% endfor %}{% if add_generation_prompt %}{{ 'Assistant:' }}{% endif %}
 at row 1, column 13:
<|im_start|>{% for message in messages %}{{message['role'] | capitalize}}{% if message['content'][0]['type'] == 'image' %}{{':'}}{% else %}{{': '}}{% endif %}{% for line in message['content'] %}{% if line['type'] == 'text' %}{{line['text']}}{% elif line['type'] == 'image' %}{{ '<image>' }}{% endif %}{% endfor %}<end_of_utterance>
            ^
{% endfor %}{% if add_generation_prompt %}{{ 'Assistant:' }}{% endif %}
 at row 1, column 1:
<|im_start|>{% for message in messages %}{{message['role'] | capitalize}}{% if message['content'][0]['type'] == 'image' %}{{':'}}{% else %}{{': '}}{% endif %}{% for line in message['content'] %}{% if line['type'] == 'text' %}{{line['text']}}{% elif line['type'] == 'image' %}{{ '<image>' }}{% endif %}{% endfor %}<end_of_utterance>
^
{% endfor %}{% if add_generation_prompt %}{{ 'Assistant:' }}{% endif %}

srv    load_model: load_model: Chat template parsing error: Value is not callable: null at row 1, column 72:
<|im_start|>{% for message in messages %}{{message['role'] | capitalize}}{% if message['content'][0]['type'] == 'image' %}{{':'}}{% else %}{{': '}}{% endif %}{% for line in message['content'] %}{% if line['type'] == 'text' %}{{line['text']}}{% elif line['type'] == 'image' %}{{ '<image>' }}{% endif %}{% endfor %}<end_of_utterance>
                                                                       ^
{% endfor %}{% if add_generation_prompt %}{{ 'Assistant:' }}{% endif %}
 at row 1, column 42:
<|im_start|>{% for message in messages %}{{message['role'] | capitalize}}{% if message['content'][0]['type'] == 'image' %}{{':'}}{% else %}{{': '}}{% endif %}{% for line in message['content'] %}{% if line['type'] == 'text' %}{{line['text']}}{% elif line['type'] == 'image' %}{{ '<image>' }}{% endif %}{% endfor %}<end_of_utterance>
                                         ^
{% endfor %}{% if add_generation_prompt %}{{ 'Assistant:' }}{% endif %}
 at row 1, column 42:
<|im_start|>{% for message in messages %}{{message['role'] | capitalize}}{% if message['content'][0]['type'] == 'image' %}{{':'}}{% else %}{{': '}}{% endif %}{% for line in message['content'] %}{% if line['type'] == 'text' %}{{line['text']}}{% elif line['type'] == 'image' %}{{ '<image>' }}{% endif %}{% endfor %}<end_of_utterance>
                                         ^
{% endfor %}{% if add_generation_prompt %}{{ 'Assistant:' }}{% endif %}
 at row 1, column 13:
<|im_start|>{% for message in messages %}{{message['role'] | capitalize}}{% if message['content'][0]['type'] == 'image' %}{{':'}}{% else %}{{': '}}{% endif %}{% for line in message['content'] %}{% if line['type'] == 'text' %}{{line['text']}}{% elif line['type'] == 'image' %}{{ '<image>' }}{% endif %}{% endfor %}<end_of_utterance>
            ^
{% endfor %}{% if add_generation_prompt %}{{ 'Assistant:' }}{% endif %}
 at row 1, column 1:
<|im_start|>{% for message in messages %}{{message['role'] | capitalize}}{% if message['content'][0]['type'] == 'image' %}{{':'}}{% else %}{{': '}}{% endif %}{% for line in message['content'] %}{% if line['type'] == 'text' %}{{line['text']}}{% elif line['type'] == 'image' %}{{ '<image>' }}{% endif %}{% endfor %}<end_of_utterance>
^
{% endfor %}{% if add_generation_prompt %}{{ 'Assistant:' }}{% endif %}

srv    load_model: load_model: The chat template that comes with this model is not yet supported, falling back to chatml. This may cause the model to output suboptimal responses

[CISC]

You can bypass this with --no-jinja for now, but I've made a PR that fixes it.

[cchadowitz]

Thank you!