C-DiffSET · SAR2EO
C-DiffSET (TCSVT'26) retrained by the GeoSET authors on the SAR2EO split. These are the weights behind the C-DiffSET row for SAR2EO in Table 3 of the GeoSET paper (arXiv:2609.37496).
- Method: Do, Lee, Lee and Kim, "C-DiffSET: Leveraging latent diffusion for SAR-to-EO image translation with confidence-guided reliable object generation", IEEE Transactions on Circuits and Systems for Video Technology, 2026.
- Implementation: KAIST-VICLab/C-DiffSET at commit
6fc380215ae97c9d134e8f1434061b652339b45c, on Stable Diffusion 2.1-base (Manojb/stable-diffusion-2-1-base, revision0094d483a120f3f33dafbd187ea4aa60d10de75c). - Training and inference code:
baselines/c-diffsetin the GeoSET repository.
Files
| File | Bytes | SHA-256 | Role |
|---|---|---|---|
diffusion_pytorch_model.safetensors |
3,463,784,116 | 72148c1b9fe2bd354d6521d1317a01bbd931d73685355b653dd070df55f813c5 |
UNet weights (diffusers UNet2DConditionModel, fp32); input = SAR latent and noisy EO latent (8 channels), output = noise (channels 0-3) and a spatial variance whose inverse is the confidence (channel 4) |
config.json |
911 | a387cad525c0765440e96bad8b2ece4c9a63712c4df04ed95a95b57ddbd9927c |
UNet configuration |
Results (GeoSET paper, Table 3)
SAR2EO, first 4,000 pairs of the 21,260-pair test split, 256 × 256.
| FID↓ | DISTS↓ | KID↓ | DINO↑ | LPIPS↓ | SSIM↑ | PSNR↑ |
|---|---|---|---|---|---|---|
| 64.0 | 0.252 | 0.0257 | 0.527 | 0.462 | 0.512 | 17.09 |
Inference
git clone https://github.com/KAIST-VICLab/GeoSET && cd GeoSET
export BASELINES_ROOT=/path/to/upstream_code DATA_ROOT=/path/to/data WORK_DIR=/path/to/work CKPT_ROOT=/path/to/checkpoints
bash baselines/c-diffset/fetch.sh
python baselines/data/ldm_prepare.py sar2eo
hf download JeonghyeokDo/GeoSET "baselines/sar2eo/c-diffset/diffusion_pytorch_model.safetensors" "baselines/sar2eo/c-diffset/config.json" --local-dir hf
mkdir -p $CKPT_ROOT/c-diffset/sar2eo && cp hf/baselines/sar2eo/c-diffset/* $CKPT_ROOT/c-diffset/sar2eo/
bash baselines/c-diffset/test.sh sar2eo # writes $WORK_DIR/results/c-diffset/sar2eo/<stem>.png
python scripts/evaluate.py --dataset sar2eo --pred-dir $WORK_DIR/results/c-diffset/sar2eo
Licence
These weights are a derivative of Stable Diffusion 2.1-base and are released under the CreativeML Open RAIL++-M licence (text); the use-based restrictions of its Attachment A apply to any use and redistribution. The C-DiffSET code is MIT (Copyright (c) 2026 KAIST VICLab).