pyoe.OEBench.ADBench package¶
Subpackages¶
Submodules¶
pyoe.OEBench.ADBench.data_generator module¶
pyoe.OEBench.ADBench.myutils module¶
- class pyoe.OEBench.ADBench.myutils.Utils¶
Bases:
object- cal_loss(y, y_pred, mode='devnet')¶
- data_description(X, y)¶
- get_device(gpu_specific=False)¶
- grad_norm(grad_tuple)¶
- metric(y_true, y_score, pos_label=1)¶
- plot_grad_flow(named_parameters)¶
- result_process(result_show, name, std=False)¶
- sampler(X_train, y_train, batch_size)¶
- sampler_2(X_train, y_train, step, batch_size=512)¶
- sampler_pairs(X_train_tensor, y_train, epoch, batch_num, batch_size, s_a_a, s_a_u, s_u_u)¶
X_train_tensor: the input X in the torch.tensor form y_train: label in the numpy.array form
batch_num: generate how many batches in one epoch batch_size: the batch size
- set_seed(seed)¶
- torch_cdf_loss(tensor_a, tensor_b, p=1)¶
- unique(a, b)¶