Upload README.md with huggingface_hub

README.md

@@ -38,18 +38,20 @@ Five ground-roll intensity levels produce paired noisy / noise-label records:
 
 | Level | SNR (dB) | PSNR (dB) | SSIM | MAE | MSE | RMSE |
 |-------|----------|-----------|------|-----|-----|------|
-| 1.0   | 2.7129   | 21.8322   | 0.9527 | 0.015312 | 0.006558 | 0.080982 |
-| 3.0   | -6.8295  | 18.3104   | 0.9477 | 0.022968 | 0.014756 | 0.121473 |
-| 5.0   | -11.2665 | 17.3952   | 0.9466 | 0.025520 | 0.018217 | 0.134970 |
-| 7.0   | -14.1891 | 16.9715   | 0.9461 | 0.026796 | 0.020084 | 0.141719 |
-| 9.0   | -16.3720 | 16.7268   | 0.9458 | 0.027561 | 0.021248 | 0.145768 |
+| 1.0   | -2.7129  | 19.1193   | 0.6766 | 0.049940 | 0.012248 | 0.110671 |
+| 3.0   | 6.8295   | 25.1399   | 0.8047 | 0.024970 | 0.003062 | 0.055335 |
+| 5.0   | 11.2665  | 28.6618   | 0.8604 | 0.016647 | 0.001361 | 0.036890 |
+| 7.0   | 14.1891  | 31.1605   | 0.8929 | 0.012485 | 0.000766 | 0.027668 |
+| 9.0   | 16.3720  | 33.0987   | 0.9149 | 0.009988 | 0.000490 | 0.022134 |
 
 *Metrics computed on the test set (2D flattened shot gathers) in the normalized domain before denoising.*
 
 ## Model Architectures
 
 - **DDPM cDDPM** (`ddpm`) — Conditional DDPM (DDPM-2c) jointly modeling signal and ground-roll distributions. Modified U-Net with time embedding, self-attention bottleneck, 5 ResNet levels.
-- **physics** (`physics`) — 
+- **Enhanced Atten-UNet** (`enhanced_atten_unet`) — U-Net with residual blocks + attention-gated skip connections, trained with hybrid MSE + AFM (adaptive frequency modulation) loss. Base channels: 32, depth: 4.
+- **Pix2Pix cGAN** (`pix2pix`) — Pix2Pix cGAN with 7-level U-Net generator (Conv4×4 stride 2, LeakyReLU) and 4-level PatchGAN discriminator. Trained with adversarial + L1 (λ=100) loss.
+- **SANet** (`sanet`) — Soft Attention Network: multi-branch parallel convs (3×3, 5×5, 7×7) + spatial soft attention + residual blocks. Base channels: 32, 8 blocks.
 
 ### Preprocessing
 
@@ -113,7 +115,62 @@ See the companion benchmark documentation for detailed experimental setup and fu
 
 ## Results
 
-*Results pending — run batch_evaluate.py to populate.*
+Mean ± std over 3 seeds on the held-out test shot (FFID=9), evaluated on 2D-flattened data in the normalized domain. Raw (noisy) is the input before denoising.
+
+### Noise Level 1.0
+
+| Method | Params (M) | SNR (dB) | PSNR (dB) | SSIM | MAE | MSE | RMSE |
+| --- | :---: | :---: | :---: | :---: | :---: | :---: | :---: |
+| Raw (noisy) | — | 2.7129±0.0000 | 21.8322±0.0000 | 0.9527±0.0000 | 0.015312±0.000000 | 0.006558±0.000000 | 0.080982±0.000000 |
+| Enhanced Atten-UNet | 8.2 | 11.9244±0.4043 | 33.7566±0.4043 | 0.9288±0.0024 | 0.008774±0.000263 | 0.000422±0.000039 | 0.020535±0.000949 |
+| SANet | 0.23 | 21.4884±0.3003 | 40.6078±0.3002 | 0.9886±0.0009 | 0.003088±0.000175 | 0.000087±0.000006 | 0.009328±0.000324 |
+| Physics CNN | 7.27 | 22.3319±0.5078 | 36.2584±0.5077 | 0.9728±0.0028 | 0.006262±0.000573 | 0.000238±0.000027 | 0.015402±0.000889 |
+| Pix2Pix cGAN | 48.84 | 15.9398±0.4189 | 35.0591±0.4189 | 0.9755±0.0041 | 0.004174±0.001219 | 0.000313±0.000031 | 0.017676±0.000864 |
+| DDPM cDDPM | 50.61 | 15.4040±1.4496 | 29.3304±1.4496 | 0.9066±0.0416 | 0.012975±0.002109 | 0.001208±0.000366 | 0.034466±0.005503 |
+
+### Noise Level 3.0
+
+| Method | Params (M) | SNR (dB) | PSNR (dB) | SSIM | MAE | MSE | RMSE |
+| --- | :---: | :---: | :---: | :---: | :---: | :---: | :---: |
+| Raw (noisy) | — | -6.8295±0.0000 | 18.3104±0.0000 | 0.9477±0.0000 | 0.022968±0.000000 | 0.014756±0.000000 | 0.121473±0.000000 |
+| Enhanced Atten-UNet | 8.2 | 18.5837±0.9329 | 36.8940±0.9330 | 0.9580±0.0027 | 0.006279±0.000549 | 0.000208±0.000045 | 0.014354±0.001559 |
+| SANet | 0.23 | 17.1129±0.2490 | 42.2528±0.2490 | 0.9876±0.0006 | 0.002294±0.000126 | 0.000060±0.000003 | 0.007717±0.000222 |
+| Physics CNN | 7.27 | 21.2933±0.2677 | 36.0215±0.2677 | 0.9613±0.0015 | 0.007664±0.000260 | 0.000250±0.000016 | 0.015815±0.000491 |
+| Pix2Pix cGAN | 48.84 | 8.9322±0.3070 | 34.0721±0.3071 | 0.9725±0.0015 | 0.003538±0.000607 | 0.000392±0.000028 | 0.019796±0.000702 |
+| DDPM cDDPM | 50.61 | 14.6481±1.7167 | 29.3762±1.7167 | 0.9154±0.0353 | 0.012519±0.001434 | 0.001211±0.000418 | 0.034405±0.006392 |
+
+### Noise Level 5.0
+
+| Method | Params (M) | SNR (dB) | PSNR (dB) | SSIM | MAE | MSE | RMSE |
+| --- | :---: | :---: | :---: | :---: | :---: | :---: | :---: |
+| Raw (noisy) | — | -11.2665±0.0000 | 17.3952±0.0000 | 0.9466±0.0000 | 0.025520±0.000000 | 0.018217±0.000000 | 0.134970±0.000000 |
+| Enhanced Atten-UNet | 8.2 | 21.4265±0.8323 | 38.8217±0.8323 | 0.9708±0.0019 | 0.004890±0.000354 | 0.000133±0.000024 | 0.011488±0.001072 |
+| SANet | 0.23 | 14.2619±0.1390 | 42.9236±0.1390 | 0.9875±0.0004 | 0.001949±0.000145 | 0.000051±0.000002 | 0.007142±0.000115 |
+| Physics CNN | 7.27 | 21.8948±0.1265 | 36.8837±0.1264 | 0.9690±0.0001 | 0.006944±0.000087 | 0.000205±0.000006 | 0.014317±0.000209 |
+| Pix2Pix cGAN | 48.84 | 5.6983±0.6128 | 34.3600±0.6128 | 0.9688±0.0031 | 0.004315±0.001284 | 0.000369±0.000053 | 0.019175±0.001360 |
+| DDPM cDDPM | 50.61 | 14.4170±1.4478 | 29.4060±1.4478 | 0.9140±0.0382 | 0.012570±0.001544 | 0.001187±0.000353 | 0.034165±0.005404 |
+
+### Noise Level 7.0
+
+| Method | Params (M) | SNR (dB) | PSNR (dB) | SSIM | MAE | MSE | RMSE |
+| --- | :---: | :---: | :---: | :---: | :---: | :---: | :---: |
+| Raw (noisy) | — | -14.1891±0.0000 | 16.9715±0.0000 | 0.9461±0.0000 | 0.026796±0.000000 | 0.020084±0.000000 | 0.141719±0.000000 |
+| Enhanced Atten-UNet | 8.2 | 23.4293±1.1926 | 40.4008±1.1926 | 0.9785±0.0012 | 0.003988±0.000376 | 0.000093±0.000023 | 0.009608±0.001265 |
+| SANet | 0.23 | 12.0953±0.1912 | 43.2559±0.1912 | 0.9877±0.0008 | 0.001822±0.000184 | 0.000047±0.000002 | 0.006875±0.000151 |
+| Physics CNN | 7.27 | 22.2912±0.4163 | 37.4043±0.4163 | 0.9701±0.0034 | 0.006623±0.000452 | 0.000182±0.000018 | 0.013494±0.000653 |
+| Pix2Pix cGAN | 48.84 | 3.4396±0.3167 | 34.6001±0.3168 | 0.9692±0.0026 | 0.003104±0.000211 | 0.000347±0.000025 | 0.018629±0.000672 |
+| DDPM cDDPM | 50.61 | 14.3006±1.6218 | 29.4137±1.6218 | 0.9035±0.0324 | 0.012902±0.001538 | 0.001195±0.000397 | 0.034213±0.006054 |
+
+### Noise Level 9.0
+
+| Method | Params (M) | SNR (dB) | PSNR (dB) | SSIM | MAE | MSE | RMSE |
+| --- | :---: | :---: | :---: | :---: | :---: | :---: | :---: |
+| Raw (noisy) | — | -16.3720±0.0000 | 16.7268±0.0000 | 0.9458±0.0000 | 0.027561±0.000000 | 0.021248±0.000000 | 0.145768±0.000000 |
+| Enhanced Atten-UNet | 8.2 | 24.8580±1.5442 | 41.5847±1.5442 | 0.9820±0.0028 | 0.003487±0.000463 | 0.000072±0.000023 | 0.008417±0.001415 |
+| SANet | 0.23 | 10.5793±0.1366 | 43.6780±0.1367 | 0.9884±0.0006 | 0.001715±0.000084 | 0.000043±0.000001 | 0.006548±0.000104 |
+| Physics CNN | 7.27 | 23.0216±0.3917 | 38.2117±0.3917 | 0.9726±0.0024 | 0.006038±0.000219 | 0.000151±0.000013 | 0.012295±0.000550 |
+| Pix2Pix cGAN | 48.84 | 0.0373±2.6620 | 33.1360±2.6620 | 0.9660±0.0036 | 0.004698±0.000950 | 0.000555±0.000367 | 0.022771±0.007399 |
+| DDPM cDDPM | 50.61 | 14.2884±1.1511 | 29.4785±1.1511 | 0.9251±0.0252 | 0.012315±0.001360 | 0.001153±0.000280 | 0.033771±0.004296 |
 
 
 ## References