Skip to content
Other

vehicle-reid-veri776

a model for matching vehicles across different cameras

2 min readupdated

This model tackles vehicle re-identification (vehicle Re-ID): it takes a photo of a car and converts it into a 2048-dimensional vector (embedding). Capture the same vehicle from different angles or on other cameras, and the resulting vectors will sit very close to each other. Built on a ResNet50 backbone with architectural enhancements (including BNNeck and center loss) and trained on the VeRi-776 dataset, it is well-suited for video analytics developers, computer vision researchers, and anyone working on duplicate detection or visual object matching.

Key features

  • High accuracy: reaches 78.3% mAP and 96.0% Rank-1 on the VeRi-776 test set.
  • Re-ranking post-processing: applying k-reciprocal encoding boosts mAP to 80.6%.
  • Solid architecture: pairs a strong ResNet50 baseline with triplet loss and center loss for better class separation.
  • Efficient training: trained on a single RTX 4050 GPU in just 8.8 hours using mixed precision.

What you can do

  • Search vehicles in a database: compare a target car's vector against an image gallery to automatically retrieve matching frames.
  • Integrate into a pipeline: use the ready-made PyTorch implementation to extract vehicle embeddings from surveillance footage.
  • Experiment with re-ranking: apply the k-reciprocal algorithm to refine result ordering in your search gallery.
  • Explore the code: inspect a clean PyTorch strong baseline implementation with detailed augmentation and loss configurations.

Example

import torch
from reid.models.baseline import Baseline
from reid.data.transforms import build_transforms
from PIL import Image

model = Baseline(num_classes=576, pretrained=False)
model.load_state_dict(torch.load("vehicle_reid_baseline.pt", map_location="cpu"))
model.eval()

transform = build_transforms(train=False)
image = transform(Image.open("carro.jpg").convert("RGB")).unsqueeze(0)
with torch.no_grad():
    embedding = torch.nn.functional.normalize(model(image), dim=1)
Read moreDetails
Benchmarks: 78.3% mAP (80.6% with re-ranking), 96.0% Rank-1 (96.5% with re-ranking).
Hardware requirements: trained on 1x RTX 4050 (6 GB VRAM).
Limitations: trained solely on daytime VeRi-776 data from 20 cameras; sensitive to severe lighting changes and does not recognize license plates.
Discuss Telegram 0

Will it run?

Runs in BF16

Fits without quantisation.

  • BF16weights 0.1 GB · ready files0.6 GB
  • FP8 / INT8recommendedweights 0.0 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.

    # sarahcalixto/vehicle-reid-veri776
    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 sarahcalixto/vehicle-reid-veri776 --local-dir ~/models/vehicle-reid-veri776
    # The model is in ~/models. How to run it is in the model card (link above).
    Add more LoRAs to vehicle-reid-veri776
  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 vehicle-reid-veri776 (quants, finetunes, LoRAs) will show up under the bell.

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

About the model

sarahcalixto/vehicle-reid-veri776
Author
sarahcalixto
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