Sponsored by
Vmake Labs Video Enhancer
-Vmake Labs Video Enhancer uses AI to improve video quality, restore details, reduce blur, and upscale videos to 8K.
SmolVLM2-500M is a tiny vision-language model from Hugging Face: a
SigLIP
vision encoder +
pixel-shuffle connector feeding a
SmolLM2 (Llama-architecture) 360M
decoder. At just
361 MB
it
is one of the smallest on-device VLMs — give it an image and a question, get a grounded answer, fully
offline.
File
SmolVLM2-500M.litertlm
(~361 MB)
Vision
SigLIP encoder (512×512, 1024 patches, no CLS) + pixel-shuffle ×4 + Linear connector,
int8
→
64 image tokens
resized to 512×512 ((x−0.5)/0.5 normalization baked into the vision encoder)
Base model
HuggingFaceTB/SmolVLM2-500M-Video-Instruct
Quality
Single-image VQA produces coherent, image-grounded answers (CPU-verified; the SigLIP vision tower
converts bit-faithfully, float CPU-parity corr ≈ 1.0). It is a
very small (500M) model
— keep a
sensible
max_tokens
and use sampling (e.g. top-p); at pure greedy it can be repetitive/verbose.
⚠️ Best for single-image VQA — one image per conversation
Ask about
one image per chat
(start a new conversation for a different image). Single-image VQA is
the primary use case. (On the GPU backend, a second image in the same conversation may degrade — a
GPU-delegate trait shared across
fast_vlm
models; CPU handles multi-image.)
Run on iPhone / macOS
Use the LiteRT-LM Swift runtime (
swift-litert-lm
/
the
LiteRTDemo
sample). Load
SmolVLM2-500M.litertlm
with the
vision tower enabled
(modalities
Modality.textImage
/
[.vision]
— vision-only bundle, no audio tower), attach a photo,
ask a question.
Run on Android — Google AI Edge Gallery
Run this model
with image input
in the official
Google AI Edge Gallery
app — no custom app needed
(the bundle carries the tokenizer, chat template, and image preprocessing config):
Push the bundle onto the phone (or download it there directly from this repo):
adb push SmolVLM2-500M.litertlm /sdcard/Download/
Open the Gallery app, tap the
+
icon (bottom-right) and pick
SmolVLM2-500M.litertlm
in the file picker.
In the
Import Model
dialog,
check "Support image"
(required for image input), set a sensible
max tokens
, pick
GPU
(fast) or
CPU
, then tap
Import
.
Open the
Ask Image
task, choose the imported model, attach a photo, and ask.
Tip:
ask about
one image per conversation
. It's a tiny 500M model — keep max-tokens modest so it doesn't ramble.
The vision encoder uses the static
arange(1024)
position-embedding path (the model's dynamic
bucketize position logic is bypassed — numerically identical for a full 512×512 frame) and bakes the
(x−0.5)/0.5 normalization + NCHW transpose into the graph.
Single-image, no high-res splitting → a fixed 64 soft tokens; SmolLM2 (Llama) decoder exported with
externalized (tied) embedder.
SmolVLM2-500M huggingface.co is an AI model on huggingface.co that provides SmolVLM2-500M's model effect (), which can be used instantly with this litert-community SmolVLM2-500M model. huggingface.co supports a free trial of the SmolVLM2-500M model, and also provides paid use of the SmolVLM2-500M. Support call SmolVLM2-500M model through api, including Node.js, Python, http.
SmolVLM2-500M huggingface.co is an online trial and call api platform, which integrates SmolVLM2-500M's modeling effects, including api services, and provides a free online trial of SmolVLM2-500M, you can try SmolVLM2-500M online for free by clicking the link below.
litert-community SmolVLM2-500M online free url in huggingface.co:
SmolVLM2-500M is an open source model from GitHub that offers a free installation service, and any user can find SmolVLM2-500M on GitHub to install. At the same time, huggingface.co provides the effect of SmolVLM2-500M install, users can directly use SmolVLM2-500M installed effect in huggingface.co for debugging and trial. It also supports api for free installation.