Skip to content
Other

European Dashcam Vehicle Classifier

vehicle classification from dashcam footage

2 min readupdated

A model built on the ConvNeXt-Tiny architecture (an image recognition AI model) and fine-tuned on footage from European dashcams. It identifies vehicle makes and models across 117 different classes from cropped images. The backbone was pretrained on a massive dataset of 460,000 car images. This tool is useful for researchers, developers, and hobbyists working on traffic video analytics.

Key features

  • 117 classes: covers popular vehicle makes and models with enough training examples.
  • High accuracy: achieves 91.6% top-1 accuracy on the test split and 92.2% on human-audited data.
  • Efficient backbone: powered by the lightweight, high-performing ConvNeXt-Tiny model at 224px resolution.
  • Solid training pipeline: uses weighted class sampling and label smoothing to handle class imbalance.

What you can do

  • Dashcam video analysis: automate vehicle make recognition in surveillance camera recordings.
  • Traffic research: integrate the model into non-commercial academic traffic monitoring projects.
  • Data filtering: annotate and sort large collections of road video footage.

Example

import torch
from huggingface_hub import hf_hub_download
from torchvision import models, transforms
from PIL import Image

ckpt = torch.load(hf_hub_download("rdo1/euro-dashcam-vehicle-classifier-117", "model.pt"),
                  map_location="cpu", weights_only=False)
classes = ckpt["classes"]
model = models.convnext_tiny(weights=None)
model.classifier[2] = torch.nn.Linear(model.classifier[2].in_features, len(classes))
model.load_state_dict(ckpt["state_dict"]); model.eval()

tf = transforms.Compose([
    transforms.Resize(224, interpolation=transforms.InterpolationMode.BICUBIC),
    transforms.CenterCrop(224), transforms.ToTensor(),
    transforms.Normalize([0.485,0.456,0.406], [0.229,0.224,0.225]),
])
img = Image.open("crop.jpg").convert(
    "RGB")
with torch.inference_mode():
    p = model(tf(img).unsqueeze(0)).softmax(1)[0]
top = p.topk(5)
print([(classes[i], round(float(p[i]), 3)) for i in top.indices])
Read moreDetails
Metrics: 0.916 top-1 and 0.978 top-5 on a fixed split; 0.922 top-1 on manually verified frames.
Training: 30 epochs on 6,039 frames with CLIP-based label filtering.
Limitations: strictly for non-commercial research use under the CC-BY-NC-4.0 license.
Inputs: single-car images measuring around 100–400 pixels on the shorter side.
Discuss Telegram 0

Will it run?

Runs in BF16

Fits without quantisation.

  • BF16weights 0.1 GB · ready files0.6 GB
  • FP8 / INT8recommendedweights 0.1 GB0.6 GB
  • 4-bitweights 0.0 GB0.5 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 · BF16 — 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.

    # rdo1/euro-dashcam-vehicle-classifier-117
    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 rdo1/euro-dashcam-vehicle-classifier-117 --local-dir ~/models/euro-dashcam-vehicle-classifier-117
    # The model is in ~/models. How to run it is in the model card (link above).
  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 euro-dashcam-vehicle-classifier-117 (quants, finetunes, LoRAs) will show up under the bell.

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

About the model

rdo1/euro-dashcam-vehicle-classifier-117
Author
rdo1
Category
Other
Released
Posted

On NeyroLearn

Hugging Face

00

Ask AI about this model

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

Comments