Skip to content
Other

LiteCalib ViT-B

Fast single-image camera calibration

2 min readupdated

The LiteCalib model, built on the DINOv2 ViT-B/14 backbone (a vision transformer with 106M parameters), tackles single-view camera calibration. It estimates the vertical field of view (vFoV) and optical distortion parameters from a single static image using symmetric camera models. This development will be useful for computer vision researchers, 3D reconstruction creators, and augmented reality app developers who need to recover scene geometry without special calibration patterns.

Key features

  • DINOv2 backbone: Uses a powerful visual foundation for feature extraction.
  • High speed: Frame processing takes just 6.3 ms on an NVIDIA A100 (fp16).
  • Accuracy: The median angular error is 1.50° according to the paper benchmarks.
  • TensorRT support: Optimized for running on edge devices like the Jetson Orin Nano.

What you can do

  • Geometry analysis: Determine the field of view and optical distortions for any single snapshot.
  • Integration into 3D pipelines: Use calibration as a preprocessing step for scene reconstruction or photogrammetry.
  • Hardware deployment: Deploy the model on robotic or embedded modules via TensorRT for real-time operation.

Example

import litecalib

model = litecalib.load("litecalib_vitb")
calib = model.calibrate("photo.jpg", "radial:1")
print(calib.K, calib.distortion)
Read moreDetails
Median angular error: 1.50°
Median vFoV error: 3.20°
Performance: 6.6 ms total on NVIDIA A100 (PyTorch fp16) and 27.1 ms on Jetson Orin Nano 8 GB (TensorRT fp16)
Training: 24 epochs on OpenPano, CommonsPano, and MegaScenes data
License: Apache License 2.0
Discuss Telegram 0 29

Will it run?

Runs in F32

Fits without quantisation.

  • F32weights 0.4 GB · ready files0.9 GB
  • BF16weights 0.2 GB0.7 GB
  • FP8 / INT8recommendedweights 0.1 GB0.6 GB
  • 4-bitweights 0.1 GB0.6 GB

Tip: go 8-bit.FP8 / INT8 is within about a percent of full precision in quality, takes half the memory and runs faster on RTX 40/50 and H100. Full precision only pays off for fine-tuning.

What to get for this model

Massed ComputeOpen Massed Compute

Your own GPU: FP8 / INT8 — from RTX 2060 · F32 — from RTX 2060

Massed Compute hourly prices, November 2026. The discounted price applies with the coupon at checkout.

Estimate: weights plus working memory. Long videos, high resolutions and long contexts need more.

Run it on Massed Compute

The script downloads the model to the rented machine. How to run it is in the model card.

  1. 1. Rent a machine with a fitting GPU

    Open Massed Compute
  2. 2. Paste the script into the machine's terminal

    Connect over SSH (the Massed Compute panel shows the command) or open a terminal on the machine's desktop.

    HF_TOKEN / CIVITAI_TOKENThe script starts with token lines: paste yours between the quotes. The Hugging Face token is for gated models, the Civitai one for Civitai models.

    # LiteCalib/litecalib-vitb
    mkdir -p ~/models
    curl -LsSf https://astral.sh/uv/install.sh | sh
    source "$HOME/.local/bin/env"
    uv tool install "huggingface_hub[cli]"
    hf download LiteCalib/litecalib-vitb --local-dir ~/models/litecalib-vitb
    # The model is in ~/models. How to run it is in the model card (link above).
    Add more LoRAs to litecalib-vitb
  3. 3. Use it

    The model is in ~/models. How to run it is in the model card (link above).

Stop the machine in the panel when you are done: it bills by the minute while it runs.

How to speed it up

Picked for this model type. Choose your hardware above to see what fits you exactly.

  • FlashAttention 44.0.0b40from November 18, 2026new release

    fastest exact attention on B200

    Kernels for Blackwell (beta). Exact attention without quantisation — fine for everything, LLMs included.

    How to enable: pip install flash-attn-4; vLLM and SGLang pick it up themselves.

    Earlier generations (2)
    • FlashAttention 3 — The Hopper (H100/H200) version with FP8.
    • FlashAttention 2 — The standard for RTX 30/40/50 and A100: exact attention without the memory cost of long contexts.
  • torch.compile2.15.0from October 28, 2026new release

    10–30% for free

    Compiles the model for your GPU. Slower first run, faster afterwards.

    How to enable: model = torch.compile(model); in ComfyUI the TorchCompileModel node.

Model family

New posts about versions of litecalib-vitb (quants, finetunes, LoRAs) will show up under the bell.

No known relatives of this model yet. All derivatives on Hugging Face

About the model

LiteCalib/litecalib-vitb
Author
LiteCalib
Category
Other
Released
Posted

On NeyroLearn

Hugging Face

290

Ask AI about this model

Answered by Gemini 4 · the chat is kept only in this tab

Comments