Han-Jia Ye, Zhenguo Li, Da-Wei Zhou, Guannan Lai
We lifted 13 functions out of this paper's own repositories and ran 12 of them in a sandbox. "Ran" means the function executed on a synthesized input and returned a value. It is not a reproduction of the paper's results.
| Repository | Role | Ran |
|---|---|---|
| AIGNLAI/GOLD | canonical | 12 of 13 |
| Function | Status | Where it lives |
|---|---|---|
| complete_data_dir_path | Ran | AIGNLAI/GOLD/classification/conf.py pointer only (licence: NONE) · get_code("53b1e10b7cc3d6de") |
| create_file_list | Ran | AIGNLAI/GOLD/classification/datasets/imagenet_subsets.py pointer only (licence: NONE) · get_code("0c4495cb2187be84") |
| ds_name2pytorch_ds_name | Ran | AIGNLAI/GOLD/classification/conf.py pointer only (licence: NONE) · get_code("22a042db3ce6f8bb") |
| forward_decorator | Ran | AIGNLAI/GOLD/classification/methods/base.py pointer only (licence: NONE) · get_code("ef670fd2920b7ce2") |
| get_num_classes | Ran | AIGNLAI/GOLD/classification/conf.py pointer only (licence: NONE) · get_code("4d83f34a9766b4a9") |
| kernel | Ran | AIGNLAI/GOLD/classification/methods/cmf.py pointer only (licence: NONE) · get_code("b086cfc2dfe368f4") |
| refine_predictions | Ran | AIGNLAI/GOLD/classification/methods/adacontrast.py pointer only (licence: NONE) · get_code("6c5a8d4a69a975c2") |
| setup_adacontrast_optimizer | Ran | AIGNLAI/GOLD/classification/methods/adacontrast.py pointer only (licence: NONE) · get_code("ba822e549de6a087") |
| soft_k_nearest_neighbors | Ran | AIGNLAI/GOLD/classification/methods/adacontrast.py pointer only (licence: NONE) · get_code("110b394c4d611e0b") |
| softmax_entropy_cifar | Ran | AIGNLAI/GOLD/classification/methods/cotta.py pointer only (licence: NONE) · get_code("991d0f3401682804") |
| softmax_entropy_imagenet | Ran | AIGNLAI/GOLD/classification/methods/cotta.py pointer only (licence: NONE) · get_code("6bdfed740b3da7ce") |
| update_model_probs | Ran | AIGNLAI/GOLD/classification/methods/cmf.py pointer only (licence: NONE) · get_code("073ab5af80ddbe36") |
| create_imagenet_subset | Not yet run | AIGNLAI/GOLD/classification/datasets/imagenet_subsets.py pointer only (licence: NONE) · get_code("b723ea0fcdc2c948") |
Some links come from the archived Papers with Code dataset (CC BY-SA 4.0): attribution and licence.
Continual Test-Time Adaptation (CTTA) aims to enable models to adapt online to unlabeled data streams under distribution shift without accessing source data. Existing CTTA methods face an efficiency-generalization trade-off: updating more parameters improves adaptation but severely reduces online inference efficiency. An ideal solution is to achieve comparable adaptation with minimal feature updates; we call this minimal subspace the golden subspace. We prove its existence in a single-step adaptation setting and show that it coincides with the row space of the pretrained classifier. To enable online maintenance of this subspace, we introduce the sample-wise Average Gradient Outer Product (AGOP) as an efficient proxy for estimating the classifier weights without retraining. Building on these insights, we propose Guided Online Low-rank Directional adaptation (GOLD), which uses a lightweight adapter to project features onto the golden subspace and learns a compact scaling vector while the subspace is dynamically updated via AGOP. Extensive experiments on classification and segmentation benchmarks, including autonomousdriving scenarios, demonstrate that GOLD attains superior efficiency, stability, and overall performance. Our code is available at https://github.com/AIGNLAI/GOLD.
The same record, over MCP at https://syntology.ai/mcp:
get_harvested_code_for_paper("2603.21928")
get_code_for_paper("2603.21928")
have("2603.21928")
Connect an agent — have() is free.