facebook / image-segmentation updated 3 years ago

mask2former-swin-large-ade-semantic

Mask2Former model trained on ADE20k semantic segmentation (large-sized version, Swin backbone). It was introduced in the paper Masked-attention Mask Transformer for Universal Image Segmentation and first released in this repository.

Params
220 M
Context
Downloads 30d
300 K
Likes
25
Licence: other Not gated SAFETENSORS View on Hugging Face ↗

Download history

daily snapshots · 55 days
▲ 92 K in the last 30 days (23.6%)
675 K284 K
Jul 28Aug 15Sep 2Sep 20

Can you run it?

Estimated VRAM at 8K context unless noted. Pick your hardware to see the verdict per quantization.

FileQuantSizeEst. VRAMVerdict on RTX 4090 · 24 GB
model.safetensors f32 0.9 GB 1.5 GB ✅ Runs comfortably
Estimate: file size × 1.1 + KV cache at 8K + 0.5 GB overhead. Not a benchmark — how we calculate this.

Specifications

Architecture
Mask2FormerForUniversalSegmentation
Parameters
220 M
Tensor type
F32
Layers / heads
10 / 8
Licence
other
First seen on the Hub
2023-01-05
Training datasets
coco
Added to our catalog
2026-07-28
Compare with any image-segmentation model