CIFAR-10 Self-Supervised Results Set New SOTA
wandb · x · 2026-07-16
The post mentions that while experimenting with self-supervised learning on CIFAR-10, the author felt their previous judgment was correct—until a new competitor emerged.
Key details include:
- Model used: RN18 (ResNet-18)
- Linear classification top-1 accuracy of 94.4%
- This beats the known SOTA by about 1 percentage point
- Next up: evaluating on ImageNet-1K
More from Research
- OpenAI says long-horizon models need safety and alignment checks across full action sequences — rhiever · 2026-07-22
- A Reddit user proposes a consistency LoRA to keep anime and game scenes visually stable — ThirdWorldBoy21 · 2026-07-22
- Graph workload 854.graph500 enters SPEC CPU 2026 as a new CPU benchmark — Prof_DavidBader · 2026-07-22
- BlackboxNLP 2026 is recruiting extra reviewers after a high submission volume — hanjie_chen · 2026-07-22
- AWS shows self-distilled reasoning can preserve math and coding skills during SFT — AWS ML Blog · 2026-07-22
- UI2App shows screenshot fidelity still lags real interaction recovery — Grace Man Chen · 2026-07-22