Instructions to use NCSOFT/VARCO-VISION-14B with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Libraries

How to use NCSOFT/VARCO-VISION-14B with Transformers:

# Use a pipeline as a high-level helper
from transformers import pipeline

pipe = pipeline("image-text-to-text", model="NCSOFT/VARCO-VISION-14B")
messages = [
    {
        "role": "user",
        "content": [
            {"type": "image", "url": "https://huggingface.co/datasets/huggingface/documentation-images/resolve/main/p-blog/candy.JPG"},
            {"type": "text", "text": "What animal is on the candy?"}
        ]
    },
]
pipe(text=messages)

# Load model directly
from transformers import AutoTokenizer, AutoModelForCausalLM

tokenizer = AutoTokenizer.from_pretrained("NCSOFT/VARCO-VISION-14B")
model = AutoModelForCausalLM.from_pretrained("NCSOFT/VARCO-VISION-14B")
messages = [
    {
        "role": "user",
        "content": [
            {"type": "image", "url": "https://huggingface.co/datasets/huggingface/documentation-images/resolve/main/p-blog/candy.JPG"},
            {"type": "text", "text": "What animal is on the candy?"}
        ]
    },
]
inputs = tokenizer.apply_chat_template(
	messages,
	add_generation_prompt=True,
	tokenize=True,
	return_dict=True,
	return_tensors="pt",
).to(model.device)

outputs = model.generate(**inputs, max_new_tokens=40)
print(tokenizer.decode(outputs[0][inputs["input_ids"].shape[-1]:]))

Notebooks
Google Colab
Kaggle
Local Apps Settings

vLLM

How to use NCSOFT/VARCO-VISION-14B with vLLM:

Install from pip and serve model

# Install vLLM from pip:
pip install vllm
# Start the vLLM server:
vllm serve "NCSOFT/VARCO-VISION-14B"
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:8000/v1/chat/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "NCSOFT/VARCO-VISION-14B",
		"messages": [
			{
				"role": "user",
				"content": [
					{
						"type": "text",
						"text": "Describe this image in one sentence."
					},
					{
						"type": "image_url",
						"image_url": {
							"url": "https://cdn.britannica.com/61/93061-050-99147DCE/Statue-of-Liberty-Island-New-York-Bay.jpg"
						}
					}
				]
			}
		]
	}'

Use Docker

docker model run hf.co/NCSOFT/VARCO-VISION-14B

SGLang

How to use NCSOFT/VARCO-VISION-14B with SGLang:

Install from pip and serve model

# Install SGLang from pip:
pip install sglang
# Start the SGLang server:
python3 -m sglang.launch_server \
    --model-path "NCSOFT/VARCO-VISION-14B" \
    --host 0.0.0.0 \
    --port 30000
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:30000/v1/chat/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "NCSOFT/VARCO-VISION-14B",
		"messages": [
			{
				"role": "user",
				"content": [
					{
						"type": "text",
						"text": "Describe this image in one sentence."
					},
					{
						"type": "image_url",
						"image_url": {
							"url": "https://cdn.britannica.com/61/93061-050-99147DCE/Statue-of-Liberty-Island-New-York-Bay.jpg"
						}
					}
				]
			}
		]
	}'

Use Docker images

docker run --gpus all \
    --shm-size 32g \
    -p 30000:30000 \
    -v ~/.cache/huggingface:/root/.cache/huggingface \
    --env "HF_TOKEN=<secret>" \
    --ipc=host \
    lmsysorg/sglang:latest \
    python3 -m sglang.launch_server \
        --model-path "NCSOFT/VARCO-VISION-14B" \
        --host 0.0.0.0 \
        --port 30000
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:30000/v1/chat/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "NCSOFT/VARCO-VISION-14B",
		"messages": [
			{
				"role": "user",
				"content": [
					{
						"type": "text",
						"text": "Describe this image in one sentence."
					},
					{
						"type": "image_url",
						"image_url": {
							"url": "https://cdn.britannica.com/61/93061-050-99147DCE/Statue-of-Liberty-Island-New-York-Bay.jpg"
						}
					}
				]
			}
		]
	}'

Docker Model Runner
How to use NCSOFT/VARCO-VISION-14B with Docker Model Runner:
```
docker model run hf.co/NCSOFT/VARCO-VISION-14B
```

Update tokenizer_config.json

by kaki-paper - opened Dec 12, 2024

base: refs/heads/main

←

from: refs/pr/4

Discussion Files changed

-1

kaki-paper

Dec 12, 2024

fixes this error:

Traceback (most recent call last):
File "/data/workspace/howard/workspace/LLaVA-NeXT/qwen2_vl_test_inference.py", line 35, in
prompt = tokenizer.apply_chat_template(conversation, add_generation_prompt=True, tokenize=False)
File "/data/workspace/howard/workspace/.visionzip_test/lib/python3.10/site-packages/transformers/tokenization_utils_base.py", line 1774, in apply_chat_template
compiled_template = self._compile_jinja_template(chat_template)
File "/data/workspace/howard/workspace/.visionzip_test/lib/python3.10/site-packages/transformers/tokenization_utils_base.py", line 1835, in _compile_jinja_template
return jinja_env.from_string(chat_template)
File "/data/workspace/howard/workspace/.visionzip_test/lib/python3.10/site-packages/jinja2/environment.py", line 1108, in from_string
return cls.from_code(self, self.compile(source), gs, None)
File "/data/workspace/howard/workspace/.visionzip_test/lib/python3.10/site-packages/jinja2/environment.py", line 768, in compile
self.handle_exception(source=source_hint)
File "/data/workspace/howard/workspace/.visionzip_test/lib/python3.10/site-packages/jinja2/environment.py", line 939, in handle_exception
raise rewrite_traceback_stack(source=source)
File "", line 7, in template
jinja2.exceptions.TemplateSyntaxError: Encountered unknown tag 'generation'. Jinja was looking for the following tags: 'endfor' or 'else'. The innermost block that needs to be closed is 'for'.

fixed chat template came from NCSOFT/VARCO-VISION-14B-HF

Update tokenizer_config.json8a355f6c

kaki-paper

Dec 12, 2024

Never mind, it seems like it was different problem, not with the chat template. After fixed, it successfully prints the output. closing PR for now.

kaki-paper changed pull request status to closed Dec 12, 2024

colossatr0n

Jun 15, 2025

Do you remember what the fix for this issue was? Running into the same thing for a different model and was looking into updating the template, but it looks like you didn't have to do that to fix the issue.

lfrance

Sep 5, 2025

i found that there is bug in tokenizer_config.json : (replace error with ok)

Error chat_template

"chat_template": "{% if messages[0]['role'] != 'system' %}{{ '<|im_start|>system\nYou are a helpful assistant.<|im_end|>\n' }}{% else %}{{ '<|im_start|>system\n' + messages[0]['content'] + '<|im_end|>\n' }}{% endif %}{% for message in messages %}{{ '<|im_start|>' + message['role'] + '\n' }}{# Render all images first #}{% for content in message['content'] | selectattr('type', 'equalto', 'image') %}{{ '\n' }}{% endfor %}{# Render all text next #}{% if message['role'] != 'assistant' %}{% for content in message['content'] | selectattr('type', 'equalto', 'text') %}{{ content['text'] }}{% endfor %}{% else %}{% for content in message['content'] | selectattr('type', 'equalto', 'text') %}{% generation %}{{ content['text'] }}{% endgeneration %}{% endfor %}{% endif %}{{ '<|im_end|>\n' }}{% endfor %}{% if add_generation_prompt %}{{ '<|im_start|>assistant\n' }}{% endif %}",

Ok chat_template

"chat_template": "{% if messages[0]['role'] != 'system' %}{{ '<|im_start|>system\nYou are a helpful assistant.<|im_end|>\n' }}{% else %}{{ '<|im_start|>system\n' + messages[0]['content'] + '<|im_end|>\n' }}{% endif %}{% for message in messages %}{{ '<|im_start|>' + message['role'] + '\n' }}{# Render all images first #}{% for content in message['content'] | selectattr('type', 'equalto', 'image') %}{{ '\n' }}{% endfor %}{# Render all text next #}{% if message['role'] != 'assistant' %}{% for content in message['content'] | selectattr('type', 'equalto', 'text') %}{{ content['text'] }}{% endfor %}{% else %}{% for content in message['content'] | selectattr('type', 'equalto', 'text') %}{{ content['text'] }}{% endfor %}{% endif %}{{ '<|im_end|>\n' }}{% endfor %}{% if add_generation_prompt %}{{ '<|im_start|>assistant\n' }}{% endif %}",

Upload images, audio, and videos by dragging in the text input, pasting, or clicking here.

Tap or paste here to upload images

· Sign up or log in to comment