| --- |
| license: other |
| license_name: lfm-open-license-v1.0 |
| license_link: LICENSE |
| base_model: |
| - LiquidAI/LFM2-VL-450M |
| tags: |
| - torq |
| - synaptics |
| - sl2619 |
| - npu |
| - edge |
| - lfm2 |
| - lfm2-vl |
| pipeline_tag: image-text-to-text |
| --- |
| |
| # LFM2-VL-450M — Torq build (Synaptics SL2619 NPU) |
| <img src="synaptics-logo-full-color.png" alt="Synaptics" width="600"/></img> |
|
|
| This repository provides compiled model files for LiquidAI's LFM2-VL-450M vision-language model, ready |
| to run on the **Synaptics SL2610-series Torq NPU**. |
| Give it an image and a natural-language question, and it answers questions about that image. |
|
|
| **Quick start guide**: |
| - [Buy a Machina kit](https://www.synaptics.com/products/embedded-processors/sl2610-product-line#devKit): Get an SL2600 Machina kit delivered to you |
| - [Torq Examples](https://synaptics-torq.github.io/torq-compiler/v/latest/): Use Torq-examples LiquidAI/LiquidAI-LFM2-VL-450M scripts to download and deploy on your Machina kit |
|
|
| <img src="astra_sl2600_machina.png" alt="SL2600 Machina kit" width="400"/></img> |
|
|
| ## Model Overview |
|
|
| LFM2‑VL is designed to process text and images with variable resolutions. Built on the LFM2 backbone, it is optimized for low-latency and edge AI applications. |
|
|
| LFM2-VL utilizes hybrid conv/attention text decoders that execute on the NPU in bf16; the token |
| embeddings run on the host CPU. |
|
|
| Image + prompt → caption / visual question answering. The image is encoded once and |
| its KV cache is reused, so follow-up questions about the same image stay fast. |
|
|
| ## Model Features |
|
|
| ### Contents |
|
|
| | File | Size | Role | |
| |---|--:|---| |
| | `vision_encoder_256.vmfb` | 203 MB | SigLIP vision encoder, 256-res → 64 image tokens | |
| | `decoder_image_2part_A.vmfb` | 353 MB | one-shot image-prefill decoder, layers 0–7 | |
| | `decoder_image_2part_B.vmfb` | 311 MB | one-shot image-prefill decoder, layers 8–15 | |
| | `decoder_nolm.vmfb` | 577 MB | LFM2 single-token decode body (hidden-state output) | |
| | `lm_head.vmfb` | 134 MB | tied LM head (hidden → 65 536 logits) | |
| | `token_embeddings.npy` | 134 MB | CPU embedding LUT / tied-LM-head weights (bf16) | |
| | `config.json`, `tokenizer.json` | — | model config + tokenizer | |
| | `cats-and-dogs-256.jpg` | — | sample 256-res image for the demo | |
| | `onnx/` | ~2 GB | reference ONNX exports (vision encoder, merged decoder, embeddings) for non-Torq runtimes | |
|
|
| ## Model Details |
|
|
| - **Base model:** LiquidAI LFM2-VL-450M (SigLIP vision tower + LFM2 language model). |
| - **Text decoder:** LFM2 — hidden size 1024, 16 layers, 16 attention heads, vocabulary 65 536, hybrid short-convolution + grouped-query attention. |
| - **Image tokens:** 64 per image (256-resolution input). |
| - **Precision:** bf16 on the NPU. |
| - **Target:** Synaptics SL2619, compiled with the Torq compiler. |
| - **On-device performance (SL2619, indicative):** vision encode ~2.4 s, image→KV prefill ~3.7 s, decode ~3.6–4.2 tok/s. |
|
|
| ## Tested Platforms |
|
|
| - [Synaptics Astra™ Machina SL2619 2GB](https://www.synaptics.com/products/embedded-processors/astra-machina-foundation-series) |
| ## Metrics |
|
|
| | Platform | Model / Stage | Environment | NPU Clock | TTFT | Infer / s | |
| | --------- | --------- | --------- | --------- | --------- | --------- | |
| | SL2619 2GB | LFM2-VL-450M | Torq v2.0.0 | 1 GHz | 2844 ms | 3.4 | |
|
|
|
|
| ## Deployment |
|
|
| The models have been tested with the following environment. |
|
|
| - Torq Compiler: *v2.0.0* |
| - Torq Runtime: *v2.0.0* included in Astra SDK release *scarthgap_6.12_v2.4.0* |
|
|
|
|
| ### Usage Tutorials / Example Apps |
|
|
| A usage example is provided in the [Torq Examples / LiquidAI-LFM2-VL-450M](https://github.com/synaptics-torq/torq-examples/tree/main/LiquidAI/LiquidAI-LFM2-VL-450M). |
|
|
| Check out the [README](https://github.com/synaptics-torq/torq-examples/blob/main/LiquidAI/LiquidAI-LFM2-VL-450M/README.md) for instructions. |
|
|
| ## License & attribution |
|
|
| This repository is a redistribution of a model created by **Liquid AI, Inc.**, licensed under the **LFM Open License v1.0**. Copies of the license and the attribution notices are included alongside the model files: |
|
|
| - **[LICENSE](https://huggingface.co/Synaptics/LiquidAI-LFM2-VL-450M/blob/main/LICENSE)** — a verbatim copy of the LFM Open License v1.0. |
| - **[NOTICE](https://huggingface.co/Synaptics/LiquidAI-LFM2-VL-450M/blob/main/NOTICE)** — the copyright, patent, trademark, and attribution |
| notices retained from the original Work (per Section 4(c) of the license). |
|
|
| Original model: **[LFM2.5-230M](https://huggingface.co/LiquidAI/LFM2-VL-450M)** · Copyright © Liquid AI, Inc. |
|
|
| ## Learn More |
|
|
| - [Synaptics AI Developer Zone](https://developer.synaptics.com?utm_source=hf): Get started with documentation, tutorials and resources for your Edge AI journey. |
| - [Torq Compiler Documentation](https://synaptics-torq.github.io/torq-compiler/v/latest/): Learn more about the Torq compiler based on MLIR and IREE. |
| - [Synaptics Astra SDK](https://synaptics-astra.github.io/doc/v/latest/): Learn more about the Yocto Project-based Linux software available for Astra SL processors. |
| - [Astra Support Portal](https://synacsm.atlassian.net/servicedesk/customer/portal/543?utm_source=hf): Connect with our engineering team and community. |
|
|
