Skip to content
Other

Pruned SegFormer-B0

a compact model for road surface defect detection

1 min readupdated

This research model performs eight-class road surface defect segmentation using aligned grayscale images and 3D depth data. Built as a lightweight version of an extended multimodal base model under FHWA EAR research, it uses pruning (a technique for reducing neural network size) to cut the parameter count by more than half and significantly boost speed. It is a solid fit for engineers, researchers, and developers building road infrastructure monitoring systems.

Key features

  • Reduces parameter count by 52.85% compared to the base model
  • Reaches processing speeds up to 14.16 FPS in FP32 on an RTX 4070
  • Performs eight-class segmentation of road defects and background
  • Includes a complete file bundle with a reconstruction recipe and checksums

What you can do

  • Analyze road conditions from imagery and 3D depth data
  • Integrate the provided load_pruned loader to deploy the lightweight architecture

Example

Download pruned.pth, extract the inference-code.zip archive, and run the loader script to reproduce the pruning recipe and start inference.

Read moreDetails
Parameters: 108,915,070
Validation mIoU (VAL522): 77.26%
Fixed test acceptance mIoU (TEST523): 73.47%
Throughput: 14.16 FPS (RTX 4070, FP32, batch size 1)
Discuss Telegram 0

Will it run?

Runs in BF16

Fits without quantisation.

  • BF16weights 0.4 GB · ready files0.9 GB
  • FP8 / INT8recommendedweights 0.2 GB0.7 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 · 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.

    # JunlongLiu/fhwa-ear-pruned-segformer-b0
    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 JunlongLiu/fhwa-ear-pruned-segformer-b0 --local-dir ~/models/fhwa-ear-pruned-segformer-b0
    # 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 fhwa-ear-pruned-segformer-b0 (quants, finetunes, LoRAs) will show up under the bell.

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

About the model

JunlongLiu/fhwa-ear-pruned-segformer-b0
Author
JunlongLiu
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