Instructions to use Neura-Tech-AI/Nexa-AI-VL-4x4B-Base with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Libraries

How to use Neura-Tech-AI/Nexa-AI-VL-4x4B-Base with Transformers:

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

pipe = pipeline("image-text-to-text", model="Neura-Tech-AI/Nexa-AI-VL-4x4B-Base", 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)

# Load model directly
from transformers import AutoModelForCausalLM
model = AutoModelForCausalLM.from_pretrained("Neura-Tech-AI/Nexa-AI-VL-4x4B-Base", trust_remote_code=True, device_map="auto")

Notebooks
Google Colab
Kaggle
Local Apps Settings

vLLM

How to use Neura-Tech-AI/Nexa-AI-VL-4x4B-Base with vLLM:

Install from pip and serve model

# Install vLLM from pip:
pip install vllm
# Start the vLLM server:
vllm serve "Neura-Tech-AI/Nexa-AI-VL-4x4B-Base"
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:8000/v1/chat/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "Neura-Tech-AI/Nexa-AI-VL-4x4B-Base",
		"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/Neura-Tech-AI/Nexa-AI-VL-4x4B-Base

SGLang

How to use Neura-Tech-AI/Nexa-AI-VL-4x4B-Base 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 "Neura-Tech-AI/Nexa-AI-VL-4x4B-Base" \
    --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": "Neura-Tech-AI/Nexa-AI-VL-4x4B-Base",
		"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 "Neura-Tech-AI/Nexa-AI-VL-4x4B-Base" \
        --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": "Neura-Tech-AI/Nexa-AI-VL-4x4B-Base",
		"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 Neura-Tech-AI/Nexa-AI-VL-4x4B-Base with Docker Model Runner:
```
docker model run hf.co/Neura-Tech-AI/Nexa-AI-VL-4x4B-Base
```

Nexa-AI-VL-4x4B-Base / config.json

Neura-Tech-AI

Update config.json

0bba90d verified 27 days ago

Raw

History Blame Contribute Delete

2.2 kB

	{
	"architectures": [
	"NexaVisionMoEForConditionalGeneration"
	],
	"auto_map": {
	"AutoConfig": "configuration_nexa.NexaVisionMoEConfig",
	"AutoModelForCausalLM": "modeling_nexa.NexaVisionMoEForConditionalGeneration"
	},
	"model_type": "nexa_vision_moe",
	"transformers_version": "5.12.1",
	"vision_model_id": "google/siglip2-so400m-patch16-naflex",
	"llm_model_id": "Neura-Tech-AI/Nexa-AI-4x4B-Instruct",
	"vision_dim": 1152,
	"llm_dim": 2560,
	"hidden_size": 2560,
	"intermediate_size": 9728,
	"moe_intermediate_size": 9728,
	"num_local_experts": 4,
	"num_experts": 4,
	"num_experts_per_tok": 2,
	"num_hidden_layers": 36,
	"num_attention_heads": 32,
	"num_key_value_heads": 8,
	"head_dim": 128,
	"hidden_act": "silu",
	"max_position_embeddings": 262144,
	"vocab_size": 151936,
	"bos_token_id": 151643,
	"eos_token_id": 151645,
	"pad_token_id": null,
	"tie_word_embeddings": true,
	"rms_norm_eps": 1e-06,
	"initializer_range": 0.02,
	"attention_bias": false,
	"attention_dropout": 0.0,
	"decoder_sparse_step": 1,
	"dtype": "bfloat16",
	"norm_topk_prob": true,
	"output_router_logits": false,
	"router_aux_loss_coef": 0.001,
	"use_cache": true,
	"use_sliding_window": false,
	"sliding_window": null,
	"max_window_layers": 36,
	"mlp_only_layers": [],
	"rope_parameters": {
	"rope_theta": 5000000,
	"rope_type": "default"
	},
	"layer_types": [
	"full_attention",
	"full_attention",
	"full_attention",
	"full_attention",
	"full_attention",
	"full_attention",
	"full_attention",
	"full_attention",
	"full_attention",
	"full_attention",
	"full_attention",
	"full_attention",
	"full_attention",
	"full_attention",
	"full_attention",
	"full_attention",
	"full_attention",
	"full_attention",
	"full_attention",
	"full_attention",
	"full_attention",
	"full_attention",
	"full_attention",
	"full_attention",
	"full_attention",
	"full_attention",
	"full_attention",
	"full_attention",
	"full_attention",
	"full_attention",
	"full_attention",
	"full_attention",
	"full_attention",
	"full_attention",
	"full_attention",
	"full_attention"
	]
	}