HomeModel Memory Fit › SmolVLM 500M Instruct
Model Memory Fit · VLM

SmolVLM 500M Instruct: memory fit on every Jetson module.

Method v1.0 · dataset 2026-09-07 · verified 2026-09-07 · last updated September 2026

0.51 B parameters (text 0.42 B + vision tower) · 32 layers · 8,192-token max context. Hugging Face: HuggingFaceTB/SmolVLM-500M-Instruct.

Model facts

FactValueNote
Parameters0.508 Btext 0.422 B + vision 0.085 B (estimated)
Layers32
Hidden size960
Attention heads15
KV heads5
Head dimension64
Max context8,192 tokens
dtypebfloat16
Vocabulary49,280

HuggingFaceTB · hf:HuggingFaceTB/SmolVLM-500M-Instruct · config.json · verified 2026-09-07

“"num_attention_heads": 15, "num_key_value_heads": 5, "head_dim": 64”

Parameters source

Vision tower

FactValue
Layers12
Hidden size768
Patch size16
Image size512
Model typeidefics3 (SigLIP-based)

Vision config source

  • HuggingFaceTB · hf:HuggingFaceTB/SmolVLM-500M-Instruct · vision_config · verified 2026-09-07 · config.json vision_config · class A
    “"hidden_size": 768, "num_hidden_layers": 12, "image_size": 512”

mmproj (vision weights): GGUF mmproj Q8_0 (vision), 0.109 GB.

mmproj source

  • ggml-org (llama.cpp maintainers) · hf:ggml-org/SmolVLM-500M-Instruct-GGUF · verified 2026-09-07 · SmolVLM-500M-Instruct-GGUF repo file listing · class A
    “mmproj-SmolVLM-500M-Instruct-Q8_0.gguf size 108783360 bytes”

Image tokens: 64 tokens per image (512² patch / 16 = 32² tokens, pixel-shuffle ×4 → 64 tokens per 512×512 tile; one tile assumed; class D).

Published quantised artefacts

A file the publisher or a community mirror actually ships, used as the class-A weight figure when the requested quant matches. Any quant without a row here falls back to parameters × bytes-per-parameter (class D for block quants).

QuantFormatFileSizeSource
INT8 / GGUF Q8_0GGUF Q8_0 (text)SmolVLM-500M-Instruct-Q8_0.gguf0.437 GBggml-org (llama.cpp maintainers) · hf:ggml-org/SmolVLM-500M-Instruct-GGUF · verified 2026-09-07

Jetson tokens/s measurements

Published or archived throughput numbers, not modelled. Class C (external measured benchmark).

No Jetson-measured tokens/s reported for this model yet.

Every module × every quantisation

Verdict and total memory at 4096-token context, 1 sequence, llama.cpp, headless. Q4 cells link to the static breakdown page; every other cell links to the live engine at that quantisation.

Notes

Model card states 'run inference on one image with 1.23GB of GPU RAM' (generic GPU, not Jetson-specific).

Method and data: Model Memory Fit methodology. Full registry: model-memory.json.

Change the context, concurrency or runtime.

The live engine covers any context length, sequence count, KV precision and runtime, with a permanent link.

OPEN MODEL MEMORY FIT →