libcity.executor.geosan_executor

class libcity.executor.geosan_executor.GeoSANExecutor(config, model, data_feature)[source]

Bases: libcity.executor.abstract_executor.AbstractExecutor

evaluate(test_dataloader)[source]

use model to test data

Parameters

test_dataloader (torch.Dataloader) – Dataloader

load_model(cache_name)[source]

加载对应模型的 cache

Parameters

cache_name (str) – 保存的文件名

static reset_random_seed(seed)[source]

重置随机数种子

Parameters

seed (int) – 种子数

save_model(cache_name)[source]

将当前的模型保存到文件

Parameters

cache_name (str) – 保存的文件名

train(train_dataloader, eval_dataloader)[source]

use data to train model with config

Parameters

  • train_dataloader (torch.Dataloader) – Dataloader

  • eval_dataloader (torch.Dataloader) – None