The following measurements use one 448x448 image at the model's compiled input resolution and a 20-token text question, so no image resizing is required.
Measured end-to-end on Modalix using five fresh requests after warm-up. Values are arithmetic means. TTFT includes image decoding and preprocessing, vision encoding and projection, language-model prefill, and the first generated token. Generation rate is measured after the first token.
Image input
Text tokens
Mean TTFT (seconds)
Mean generation rate (tokens/second)
448x448
20
0.22
43.92
Text-only generation
The following measurements exercise the language-model component without an image. Vision encoding and projection latency are not included, and input tokens refer only to text tokens.
Measured with MoLE using batch size 1, five samples per input length, and up to 128 generated tokens. Values are arithmetic means. TTFT includes language-model prefill and the first generated token; generation rate is measured after the first token.
Input tokens
Mean TTFT (seconds)
Mean generation rate (tokens/second)
128
0.05
44.71
256
0.11
43.80
512
0.21
42.12
1024
0.45
39.84
2048
1.01
30.57
3072
1.94
27.94
4096
3.12
25.40
5120
4.79
21.70
6144
6.80
18.93
7168
9.55
16.24
Prerequisites
To run this model, you need:
SiMa.ai Modalix Device
SiMa.ai CLI: Installed on your Modalix device.
SiMa.ai Neat Runtime: Install or update the Neat Library on Modalix. The LLiMa runtime is installed as part of the Neat runtime.
Hugging Face CLI: Optional, for downloading the model on a host before copying it to Modalix.
Installation & Deployment
Follow these steps to deploy the model to your Modalix device.
1. Install or Update Neat Runtime
Note: This is a one-time setup. If the Neat Library is already installed on your Modalix device, you can skip this step and continue with model download.
The llima CLI is available on Modalix after the Neat runtime is installed. It manages precompiled GenAI models under /media/nvme/llima/models by default. Set LLIMA_MODELS_PATH to use a different model directory.
2. Download the Model
Download the compiled model assets from this repository directly to your device.
bash
1# Download the model to a local directory2llima pull Qwen3-VL-2B-Instruct-GPTQ-a16w4
Alternatively, you can download the compiled model to a Host and copy it to the Modalix device:
Replace <modalix-ip> with the IP address of your Modalix device.
Expected Directory Structure:
text
1/media/nvme/llima/
2└── models/
3 └── Qwen3-VL-2B-Instruct-GPTQ-a16w4/ # The compiled model
Usage
Validate with LLiMa CLI
Run the model directly on Modalix:
llima run Qwen3-VL-2B-Instruct-GPTQ-a16w4
For all runtime options, run:
llima run -h
GenAI Demo Application
The GenAI demo application is separate from LLiMa installation. Use the GenAI Multimodal Assistant page to install and run the demo app. Once installed, the demo app can use precompiled models such as this one.
API Usage
To serve this model with OpenAI- or Ollama-compatible APIs and send requests to it, use the GenAI server workflow in Serve GenAI Models.
For direct VLM calls without setting up a server, see Run a VLM.
Limitations
Quantization: This model is quantized (A16W4) for optimal performance on embedded devices. While this maintains high accuracy, minor deviations from the full-precision model may occur.
Fixed Resolution: This version has been specifically optimized and fixed to 448x448 resolution at compile time to achieve maximum throughput and efficiency on the SiMa.ai MLA.
Troubleshooting
sima-cli not found: Ensure that sima-cli is installed on your Modalix device.
llima not found: Install or update the Neat Library. See Getting Started.
Model can't be run: Verify the model directory is exactly inside /media/nvme/llima/models/ and not nested (e.g., /media/nvme/llima/models/Qwen3-VL-2B-Instruct-GPTQ-a16w4/Qwen3-VL-2B-Instruct-GPTQ-a16w4).
Permission Denied: Ensure you have read/write permissions for the /media/nvme directory.
