Instructions to use NyxKrage/TrinityVLM-Nano with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Libraries

How to use NyxKrage/TrinityVLM-Nano with Transformers:

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

pipe = pipeline("image-text-to-text", model="NyxKrage/TrinityVLM-Nano", trust_remote_code=True)
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)

# pip install -U transformers accelerate
# Load model directly
from transformers import AutoModelForCausalLM
model = AutoModelForCausalLM.from_pretrained("NyxKrage/TrinityVLM-Nano", trust_remote_code=True, device_map="auto")

Notebooks
Google Colab
Kaggle
Local Apps Settings

vLLM

How to use NyxKrage/TrinityVLM-Nano with vLLM:

Install from pip and serve model

# Install vLLM from pip:
pip install vllm
# Start the vLLM server:
vllm serve "NyxKrage/TrinityVLM-Nano"
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:8000/v1/chat/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "NyxKrage/TrinityVLM-Nano",
		"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/NyxKrage/TrinityVLM-Nano

SGLang

How to use NyxKrage/TrinityVLM-Nano 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 "NyxKrage/TrinityVLM-Nano" \
    --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": "NyxKrage/TrinityVLM-Nano",
		"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 "NyxKrage/TrinityVLM-Nano" \
        --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": "NyxKrage/TrinityVLM-Nano",
		"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 NyxKrage/TrinityVLM-Nano with Docker Model Runner:
```
docker model run hf.co/NyxKrage/TrinityVLM-Nano
```

TrinityVLM

Trinity VLM is a vision model built on top of arcee-ai/Trinity-Nano-Preview using the vision encoder extracted from moondream/moondream3-preview

This is not inteded to be a good model, but is only an experiment in adding vision capabilites to a text-only model from scratch.

The model is trained using the following datamix:

20% anthracite-org/pixmo-cap-images
30% anthracite-org/pixmo-cap-qa-images
25% anthracite-org/pixmo-point-explanations-images
25% nvidia/Llama-Nemotron-Post-Training-Dataset chat examples with irrelevant PixMo images attached to avoid overfitting on image explaination when the prompt do not require image context.

The model is licensed under the BSL 1.1 terms of Moondream 3.

Downloads last month: 14

Safetensors

Model size

7B params

Tensor type

F32

BF16

Model tree for NyxKrage/TrinityVLM-Nano

Base model

arcee-ai/Trinity-Nano-Base-Pre-Anneal

Finetuned

arcee-ai/Trinity-Nano-Base

Finetuned

arcee-ai/Trinity-Nano-Preview

Finetuned

(6)

this model