Skip to content

benchmark: report classification results over five seeds #56

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Merged
Dimundel merged 1 commit into from
Sep 3, 2026
Merged

benchmark: report classification results over five seeds #56

Changes from all commits
Commits
Show all changes
1 commit
Select commit
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
25 changes: 13 additions & 12 deletions benchmarks/README.md
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -73,7 +73,7 @@ Approximate total runtime on an **NVIDIA RTX A4000**:

| Benchmark | Total runtime |
|----------------|--------------:|
| Classification | 4h 50min |
| Classification | 5h 20min |
| Regression | 1h 15min |

## Data
Expand All @@ -99,21 +99,22 @@ Datasets are downloaded automatically upon running the scripts. No manual prepar
Evaluates all classification / uncertainty estimation methods on:

- **In-distribution:** CIFAR-10
- **Distribution shift:** Gaussian-noised CIFAR-10
- **Distribution shift:** CIFAR-10 with Gaussian noise, σ = 0.1 in `[0, 1]` pixel space
- **Out-of-distribution:** SVHN
- **Protocol:** five seeds, reported as mean ± std

Implemented methods:

| Method | ID Acc | ECE ↓ | OOD AUROC ↑ |
| ------ | -----: | ----: | ----------: |
| Single Net | 92.72% | 0.0569 | 0.500 |
| Deep Ensemble | 93.80% | 0.0267 | 0.898 |
| MC Dropout | 92.90% | 0.0525 | 0.899 |
| Laplace (K-FAC) | 92.78% | 0.0550 | 0.848 |
| VI (ELBO) | 92.91% | **0.0197** | 0.749 |
| NES-BS | 93.73% | 0.0239 | **0.907** |
| NES-RS | **93.94%** | 0.0237 | **0.907** |
| NES-RE | 93.82% | 0.0243 | 0.891 |
| Method | ID Acc ↑ | ECE ↓ | Shift Acc ↑ | OOD AUROC ↑ |
| ------ | -----------: | -----------: | -----------: | -----------: |
| Single Net | 92.74 ± 0.15% | 0.0566 ± 0.0014 | 29.25 ± 2.54% | 0.5000 ± 0.0000 |
| Deep Ensemble | 93.68 ± 0.04% | 0.0256 ± 0.0005 | 27.43 ± 1.00% | 0.9015 ± 0.0074 |
| MC Dropout | 92.65 ± 0.14% | 0.0552 ± 0.0011 | 27.28 ± 1.04% | 0.8734 ± 0.0148 |
| Laplace (K-FAC) | 92.74 ± 0.15% | 0.0565 ± 0.0014 | **29.30 ± 2.15%** | 0.8634 ± 0.0237 |
| VI (ELBO) | 92.87 ± 0.08% | **0.0139 ± 0.0015** | 26.85 ± 1.60% | 0.7533 ± 0.0267 |
| NES-BS | **93.77 ± 0.17%** | 0.0249 ± 0.0015 | 27.55 ± 0.97% | 0.8926 ± 0.0078 |
| NES-RS | 93.77 ± 0.09% | 0.0254 ± 0.0011 | 29.20 ± 0.28% | 0.8879 ± 0.0088 |
| NES-RE | 93.76 ± 0.06% | 0.0253 ± 0.0007 | 29.21 ± 2.87% | **0.9062 ± 0.0082** |

## Regression benchmark

Expand Down
Loading