SAVRN
Search Contact SAVRN

Open-weight model

skin-lesion-segmentation-unet

by Ashraf Farhan2000/skin-lesion-segmentation-unet

Trained on ISIC 2017 Task 1 lesion-boundary masks. Educational project — not diagnostic-grade, not a medical device.

Parameters
Context
Weights97.9 MB
Licensemit
AccessOpen weights
Monthly Downloads

Model Card

By Ashraf, published under mit, revision fabc3db482aa.

Trained on ISIC 2017 Task 1 lesion-boundary masks. Educational project — not diagnostic-grade, not a medical device.

Read Ashraf's full model card

Skin Lesion Segmentation (U-Net, ResNet34 encoder)

Trained on ISIC 2017 Task 1 lesion-boundary masks. Educational project — not diagnostic-grade, not a medical device.

Load

import segmentation_models_pytorch as smp, torch
from huggingface_hub import hf_hub_download

model = smp.Unet(encoder_name="resnet34", encoder_weights=None, in_channels=3, classes=1)
weights_path = hf_hub_download(repo_id="Farhan2000/skin-lesion-segmentation-unet", filename="model.pth")
model.load_state_dict(torch.load(weights_path, map_location="cpu"))
model.eval()

Mean validation Dice: 0.8520

Identity and Version

Repository
Farhan2000/skin-lesion-segmentation-unet
Publisher
Ashraf
Task
Not stated by the source
Modality
Other
Library
Not stated by the source
Parameters
Not stated by the source
Languages
Not stated by the source
Revision
fabc3db482aa45ad47232f70a0ce85c6180f7647
First published
2026-09-18
Last updated
2026-09-18

Files and Weights

3 files, 97.9 MB in total. The weights are 1 file totalling 97.9 MB in pth.

Weights1 file · 97.9 MB
Documentation1 file · 735 B
Repository1 file · 1.5 KB
Every file
FileTypeSizeSHA-256
model.pthWeights97.9 MB ccb8d5e95999
README.mdDocumentation735 B
.gitattributesRepository1.5 KB

License and Download

License
mit
Access
Open weights, no gate
Download size
97.9 MB
Download from Ashraf

Released by Ashraf through its official repository on Hugging Face. Read the license.

Memory Requirements

PrecisionWeights in memory
As published97.9 MB

Weights only, from the published parameter count; the key-value cache and runtime add to this.

Questions About skin-lesion-segmentation-unet

Can I use skin-lesion-segmentation-unet commercially?

Yes. skin-lesion-segmentation-unet is released under MIT License. The MIT License is a short permissive license. It permits commercial use, modification and redistribution, provided the copyright notice and permission notice are included.