Functions — github.com/KarhouTam/FL-bench

hub / github.com/KarhouTam/FL-bench / functions

Functions580 in github.com/KarhouTam/FL-bench

⨍ Functions 580◇ Types & classes 144

↓ 184 callersMethodto

(self, device: torch.device)

src/utils/models.py:24

↓ 65 callersMethodstep

(self, local_parameters: list[torch.Tensor])

src/client/pfedme.py:79

↓ 62 callersMethodtrain

(self, server_package: dict[str, Any])

src/client/fedmd.py:20

↓ 43 callersMethodlog

(self, *args, **kwargs)

src/utils/logger.py:34

↓ 27 callersMethodeval

(self)

data/utils/datasets.py:61

↓ 21 callersMethodtrain

(self)

src/server/fedap.py:61

↓ 18 callersMethodupdate

(self, m, tm, tv)

src/server/fedap.py:226

↓ 16 callersMethodget_last_features

(self, x: torch.Tensor, detach=True)

src/client/fedem.py:150

↓ 14 callersMethod__init__

( self, root, args=None, test_data_transform=None, test_target_tr

data/utils/datasets.py:206

↓ 12 callersMethod__init__

(self, version, dataset, pretrained)

src/utils/models.py:453

↓ 12 callersFunctionevaluate_model

For evaluating the `model` over `dataloader` and return metrics. Args: model (DecoupledModel): Target model. dataloader (DataLoad

src/utils/functional.py:88

↓ 11 callersMethodupdate

(self, other)

src/utils/metrics.py:23

↓ 11 callersFunctionvectorize

Vectorize(Flatten) and concatenate all tensors in `src`. Args: `src`: The source of tensors. `detach`: Set as `True` to return `t

src/utils/functional.py:63

↓ 10 callersMethodaggregate_client_updates

( self, client_packages: OrderedDict[int, dict[str, Any]] )

src/server/pefll.py:60

↓ 10 callersMethodinit_trainer

Initiate the FL-bench trainier that responsible to client training. Args: `extras`: Arguments of `self.client_cls.__init__()` tha

src/server/fedavg.py:270

↓ 10 callersMethodstep

( self, clients_model_params_diff: list[OrderedDict[str, torch.Tensor]], weights: torc

src/server/fedopt.py:85

↓ 7 callersMethodforward

(self, x: Tensor)

src/utils/models.py:465

↓ 7 callersMethodpackage

(self)

src/client/apfl.py:35

↓ 7 callersMethodtrain_with_eval

Wraps `fit()` with `evaluate()` and collect model evaluation results. A model evaluation results dict: { `before`: {.

src/client/fedavg.py:85

↓ 6 callersMethod_calculate

(self, metric, **kwargs)

src/utils/metrics.py:29

↓ 5 callersMethoddisplay_metrics

Display aggregated client and server evaluation metrics at each round. This method aggregates metrics from selected clients for both

src/server/fedavg.py:704

↓ 5 callersMethodtest_client_models

(self)

src/server/ccvr.py:25

↓ 4 callersMethodadd_noise

(self, data: torch.Tensor, mean, std)

src/server/fedfed.py:285

↓ 4 callersMethodevaluate

Evaluating client model. Args: model: Used model. Defaults to None, which will fallback to `self.model`. Returns:

src/client/fedavg.py:226

↓ 4 callersMethodgrad

(self, server_package: dict[str, Any])

src/client/fediir.py:52

↓ 4 callersMethodinit_model

Initialize the global model and client personal model parameters. model: The global model. If not provided, will use the default model sp

src/server/fedavg.py:205

↓ 4 callersMethodtest

Test client model. If `finetune_epoch > 0`, `finetune()` will be activated. Args: server_package: Parameter package.

src/client/fedavg.py:284

↓ 4 callersMethodtrain

(self, server_package: dict[str, Any])

src/client/fedavg.py:155

↓ 3 callersMethodcalculate_prototypes

(self, mean=False)

src/client/fedpac.py:62

↓ 3 callersMethodclear

(self)

src/client/knnper.py:112

↓ 3 callersMethodclose

(self)

src/utils/logger.py:50

↓ 3 callersMethodencode

(self, x)

src/server/fedfed.py:297

↓ 3 callersMethodfit

(self)

src/client/moon.py:24

↓ 3 callersMethodget_data_batch

(self)

src/client/perfedavg.py:85

↓ 3 callersFunctionget_optimal_cuda_device

Dynamically select CUDA device (has the most memory) for running FL experiment. Args: use_cuda (bool): `True` for using CUDA; `False`

src/utils/functional.py:35

↓ 3 callersFunctionplot_distribution

(client_num: int, label_counts: np.ndarray, save_path: str)

data/utils/process.py:541

↓ 3 callersMethodset_parameters

(self, package: dict[str, Any])

src/client/metafed.py:53

↓ 3 callersMethodsize

(self)

src/utils/metrics.py:75

↓ 3 callersMethodwarn

Wrapper of `warnings.warn` to print the warning message onto `stdout` and log file (if enabled).

src/utils/logger.py:39

↓ 2 callersMethod__init__

(self, embed_net, hyper_net, **commons)

src/client/pefll.py:16

↓ 2 callersMethod__init__

(self, args)

src/server/fedfed.py:176

↓ 2 callersMethod__init__

(self, server: FedGenServer)

src/server/fedgen.py:159

↓ 2 callersMethod__init__

(self, args: DictConfig)

src/server/floco.py:34

↓ 2 callersFunction_get_domainnet_args

()

src/utils/constants.py:120

↓ 2 callersFunction_get_synthetic_args

()

src/utils/constants.py:129

↓ 2 callersMethod_log

Log message using server logger if available.

src/utils/dp_manager.py:61

↓ 2 callersFunctioncalculate_classifier_weights

(num_clients: int, P: torch.Tensor, idx: int)

src/server/fedpac.py:147

↓ 2 callersMethodcheck_and_preprocess

(self, args: DictConfig)

src/client/fedem.py:146

↓ 2 callersMethodcompute_mle_statistics

Compute Maximum Likelihood Estimation (MLE) statistics. Args: dataloader (DataLoader): DataLoader for the dataset. fe

src/client/pfedfda.py:239

↓ 2 callersFunctioncreate_jsons_for

Used in split-by-user case.

data/leaf_utils/split_data.py:14

↓ 2 callersMethodcreate_privacy_stats_dict

Create standardized privacy statistics dictionary structure for clients.

src/utils/dp_manager.py:72

↓ 2 callersMethodexp_coef_scheduler

(self, init_coef)

src/server/fedgen.py:147

↓ 2 callersMethodfeaturize

(self, x, num_samples=1, return_dist=False)

src/server/fedsr.py:28

↓ 2 callersFunctionfix_random_seed

Fix the random seed of FL training. Args: seed: Any number you like as the random seed. use_cuda: Flag indicates if using cuda.

src/utils/functional.py:18

↓ 2 callersMethodget_client_lr_scheduler_cls

Get the client-side learning rate scheduler class. Return None if lr_scheduler.name is NOne or no lr_scheduler arguement is provided.

src/server/fedavg.py:425

↓ 2 callersMethodget_client_optimizer_cls

Get client-side model training optimizer. Returns: A partial initiated optimizer class that client only need to add `params` arg.

src/server/fedavg.py:405

↓ 2 callersMethodget_privacy_report

Generate comprehensive privacy report.

src/utils/dp_manager.py:172

↓ 2 callersFunctioniid_divide

Divide list l among g groups each group has either int(len(l)/g) or int(len(l)/g)+1 elements returns a list of groups.

data/leaf_utils/util.py:14

↓ 2 callersMethodpairwise_distance

(self, tensor, how)

src/server/fedgen.py:223

↓ 2 callersFunctionparse_args

Purge arguments from default args dict, config file and CLI and produce the final arguments. Args: config: DictConfig set from .yaml

src/utils/functional.py:124

↓ 2 callersMethodset_parameters

(self, package: dict[str, Any])

src/client/fedfed.py:22

↓ 2 callersMethodset_parameters

(self, package: dict[str, Any])

src/client/fedavg.py:124

↓ 2 callersMethodset_parameters

(self, package: dict[str, Any])

src/client/moon.py:16

↓ 2 callersFunctionsubsample

Subsample the embeddings to a specified number of samples.

data/utils/schemes/semantic.py:39

↓ 2 callersMethodtest_discriminator

(self)

src/server/adcol.py:135

↓ 2 callersFunctionto_numpy

(x)

src/utils/metrics.py:6

↓ 2 callersMethodtrain

(self, server_package: dict[str, Any])

src/client/metafed.py:100

↓ 2 callersMethodtrain

(self, server_package: dict[str, Any])

src/client/scaffold.py:24

↓ 2 callersMethodtrain

(self, server_package: dict[str, Any])

src/client/fedproto.py:34

↓ 2 callersMethodtrain

(self)

src/server/fedavg.py:459

↓ 2 callersFunctiontraining_loop

( model, dataset, dataloader, local_epoch, optimizer, criterion, lr_scheduler,

src/client/floco.py:65

↓ 1 callersMethod__init__

(self, **commons)

src/client/fedem.py:14

↓ 1 callersMethod__init__

(self, generic_model: DecoupledModel, eval_per)

src/client/fedrod.py:147

↓ 1 callersMethod__init__

(self, args, dimension)

src/client/knnper.py:94

↓ 1 callersMethod__init__

(self, **commons)

src/client/pfedme.py:12

↓ 1 callersMethod__init__

(self, **commons)

src/client/apfl.py:11

↓ 1 callersMethod__init__

(self, args: DictConfig)

src/server/pfedla.py:28

↓ 1 callersMethod__init__

(self, args: DictConfig)

src/server/pfedhn.py:34

↓ 1 callersMethod__init__

(self, args: DictConfig)

src/server/fedrod.py:30

↓ 1 callersMethod__init__

(self, numpyform)

src/server/fedap.py:216

↓ 1 callersMethod__init__

(self, base_model: DecoupledModel, dataset)

src/server/fedsr.py:17

↓ 1 callersMethod__init__

(self, args: DictConfig)

src/server/adcol.py:72

↓ 1 callersFunction_get_celebrities_by_image

(identities)

data/celeba/preprocess/metadata_to_json.py:38

↓ 1 callersFunction_initialize_weight

(init_weight: torch.Tensor, seed: int)

src/server/floco.py:188

↓ 1 callersFunction_merge_configs

(defaults: DictConfig, config: DictConfig)

src/utils/functional.py:141

↓ 1 callersFunction_project_client_onto_simplex

(kappas, z)

src/server/floco.py:155

↓ 1 callersFunction_regularize_pers_model

(model, reg_model_params, lamda)

src/client/floco.py:93

↓ 1 callersMethod_rescale_data

(self)

data/utils/datasets.py:43

↓ 1 callersMethod_reset_privacy_engine

Reset Privacy Engine state to ensure client isolation

src/client/dpfedavg.py:54

↓ 1 callersFunction_riesz_s_energy

(simplex_points)

src/server/floco.py:167

↓ 1 callersFunction_sample_L1_ball

(center, radius)

src/server/floco.py:182

↓ 1 callersMethod_setup_dp_training

Setup DP training with provided configuration.

src/client/dpfedavg.py:63

↓ 1 callersMethod_update_privacy_stats

Update privacy budget consumption statistics.

src/client/dpfedavg.py:141

↓ 1 callersMethod_warn

Log warning using server logger if available.

src/utils/dp_manager.py:66

↓ 1 callersMethodaggregate_client_updates

Aggregate clients model parameters and produce global model parameters. Args: client_packages: Dict of client parameter p

src/server/fedavg.py:660

↓ 1 callersMethodaggregate_client_updates

Aggregate clients model parameters and produce global model parameters using FedADMM aggregation. Args: client_packages:

src/server/fedadmm.py:68

↓ 1 callersMethodaggregate_clusterwise

(self)

src/server/cfl.py:94

↓ 1 callersMethodaggregate_head

(self)

src/client/fedah.py:97

↓ 1 callersMethodaggregate_prototypes

( self, client_prototypes_list: list[dict[int, torch.Tensor]] )

src/server/fedproto.py:38

↓ 1 callersMethodaggregate_prototypes

(self, client_packages: list[dict[str, Any]])

src/server/fedpac.py:37

next →1–100 of 580, ranked by callers