Shortcuts

TransE-FB15K-single-gpu || TransE-FB15K-single-gpu-wandb || TransE-FB15K-single-gpu-hpo || TransE-FB15K-multigpu || TransE-FB15K-multigpu-wandb || TransE-FB15K237-single-gpu-wandb || TransE-WN18RR-single-gpu-adv-wandb

TransE-FB15K-multigpu

这一部分介绍如何用多个 GPU 在 FB15k 知识图谱上训练 TransE [BUGD+13]

导入数据

pybind11-OpenKE 有两个工具用于导入数据: pybind11_ke.data.TrainDataLoaderpybind11_ke.data.TestDataLoader

from pybind11_ke.config import trainer_distributed_data_parallel
from pybind11_ke.module.model import TransE
from pybind11_ke.module.loss import MarginLoss
from pybind11_ke.module.strategy import NegativeSampling
from pybind11_ke.data import TrainDataLoader

pybind11-KE 提供了很多数据集,它们很多都是 KGE 原论文发表时附带的数据集。 pybind11_ke.data.TrainDataLoader 包含 in_path 用于传递数据集目录。

# dataloader for training
train_dataloader = TrainDataLoader(
    in_path = "../../benchmarks/FB15K/",
    nbatches = 200,
    threads = 8,
    sampling_mode = "normal",
    bern = True,
    neg_ent = 25,
    neg_rel = 0)

导入模型

pybind11-OpenKE 提供了很多 KGE 模型,它们都是目前最常用的基线模型。我们下面将要导入 pybind11_ke.module.model.TransE,它是最简单的平移模型。

# define the model
transe = TransE(
    ent_tol = train_dataloader.get_ent_tol(),
    rel_tol = train_dataloader.get_rel_tol(),
    dim = 50,
    p_norm = 1,
    norm_flag = True)

损失函数

我们这里使用了 TransE [BUGD+13] 原论文使用的损失函数:pybind11_ke.module.loss.MarginLosspybind11_ke.module.strategy.NegativeSamplingpybind11_ke.module.loss.MarginLoss 进行了封装,加入权重衰减等额外项。

# define the loss function
model = NegativeSampling(
    model = transe,
    loss = MarginLoss(margin = 1.0),
    batch_size = train_dataloader.get_batch_size()
)

训练模型

pybind11-OpenKE 将训练循环包装成了 pybind11_ke.config.trainer_distributed_data_parallel() 函数, 进行并行训练,该函数必须由 if __name__ == '__main__' 保护。

if __name__ == "__main__":

    print("Start parallel training...")

    trainer_distributed_data_parallel(model = model, data_loader = train_dataloader,
            epochs = 1000, lr = 0.01, opt_method = "adam",
            test = True, valid_interval = 100, log_interval = 100, save_interval = 100,
            save_path = "../../checkpoint/transe.pth", delta = 0.01, type_constrain = True)

Total running time of the script: ( 0 minutes 0.000 seconds)

Gallery generated by Sphinx-Gallery

Docs

Access comprehensive developer documentation for Pybind11-OpenKE

View Docs