beathahahaha

I don't think the data profile is right. The number of negative items should be 100 instead of 99. (original article:"we followed the common strategy that randomly samples 100 items...

补充说明，main.py中的一个随机采样的问题：建议改写源代码如下 train_data = CriteoDataset('./data', train=True) split_num = len(train_data) * 0.8 index_list = list(range(len(train_data))) train_idx, valid_idx = index_list[:split_num], index_list[split_num:] tr_sampler = sampler.SubsetRandomSampler(train_idx) val_sampler = sampler.SubsetRandomSampler(valid_idx) loader_train = DataLoader(train_data, batch_size=100, sampler=tr_sampler) #val_data...