zhougr1993 DeepInterestNetwork issues

zhougr1993 / DeepInterestNetwork

1.6k stars 556 forks source link

issues

Newest

Newest Most commented Recently updated Oldest Least commented Least recently updated

Can not reproduce the DIN result in README.md

#92 liyangliu opened 1 year ago
0
Attention 中为什么用 softmax 把权重归一？和原论文的表述不一致

#91 wenmin-wu opened 2 years ago
1
对代码与论文模型图中的几个疑问

#90 LXXiaogege opened 2 years ago
1
train速度太慢，是否有多卡运行版本

#89 lljjgg opened 2 years ago
0
wide_deep 中 wide 部分的实现

#88 mikudehuane opened 2 years ago
0
TypeError: iter() returned non-iterator of type 'DataInputTest'

#87 ucasiggcas opened 3 years ago
1
请问代码当中的self attention与self all attention 与shine是什么意思？可以给我看看相关文章吗？

#86 GuoshenLi opened 3 years ago
0
感觉input.py那里hist_i用0补全是有问题的

#85 miaotianmiaodi closed 3 years ago
0
关于可以运行的Electronics数据集

#84 BelindaYanyiwuhua opened 3 years ago
0
关于代码中的bn实现部分的疑问

#83 ghost opened 3 years ago
2
我实现的mini-batch aware regularization，求拍砖

#82 bound2020 opened 3 years ago
1
输入问题

#81 huozijie1990 opened 3 years ago
0
why use sigmoid in Attention

#80 huaweiboy opened 3 years ago
1
attention_multi_items 的问题？

#79 Peterisfar opened 3 years ago
0
DIN模型中self.j

#78 xiaopp123 opened 4 years ago
3
不定长的hist如果不进行padding是否更合适

#77 DSXiangLi closed 4 years ago
2
代码存在严重bug: label泄露

#76 starspringcloud closed 4 years ago
2
movielen数据集如何预处理？

#75 Qian-Hao opened 4 years ago
1
原论文在介绍attention时候，叙述说放弃了softmax的归一化，但在代码中依旧保留了softmax，请问这是怎么回事？

#74 lcwy220 closed 4 years ago
2
关于负采样的问题

#73 Yanglazi opened 4 years ago
3
Why Scale

#72 LittleMisss opened 4 years ago
0
Different AUC implementation with the paper

#71 liuhu-bigeye closed 4 years ago
3
fix bn not share bugs for din

#70 liuhu-bigeye closed 4 years ago
0
max_sl 变量（用户点击序列的长度）在每个 batch 里都是不一样的值，模型结构在每个 batch 里是不固定的？

#69 lcshr123 opened 4 years ago
3
有关movielens数据集的问题

#68 yuanzhiyu opened 4 years ago
1
代码写的太差了，p8大佬

#67 weiyu322 closed 4 years ago
1
模拟随机数据 auc 特别高0.877

#66 LHFCOD opened 4 years ago
3
Not reusing batch_norm between hist_i and hist_j may

#65 liuhu-bigeye closed 4 years ago
0
DIN model中为何没有拼接user vector

#64 zhiweishen opened 4 years ago
4
Low AUC after fixed BN bugs

#63 liuhu-bigeye closed 4 years ago
1
cat aa ab ac 与python运行脚本生成的dataset.pkl不相同？大小差了两倍左右？！

#62 Ming-Qin-tech opened 4 years ago
4
代码的命名格式太差

#61 wind-meta closed 4 years ago
2
Mini-batch Aware Regularization正则化在代码哪里体现了呀。各位大佬

#60 cuiweizuishuai closed 4 years ago
1
我save model后model特别大可能是什么问题

#59 cuiweizuishuai opened 4 years ago
0
该模型是针对一个广告做训练吗？Is the model training for an AD？

#58 manji460 opened 4 years ago
1
关于DIN复现的结果问题

#57 hit-joseph closed 5 years ago
2
DataInput类hist的疑问

#56 Slayerxxx closed 5 years ago
0
DIN使用的都是现有（假设100个）itemID训练，如果出现新的（第101个）itemID,这个模型是否就不能用了？需要重新训练？

#55 lijian10086 opened 5 years ago
2
在attention里面，vec1*vec2的背后逻辑是什么？

#54 lijian10086 opened 5 years ago
12
DeepFM代码中的一个问题

#53 zhangbei121 opened 5 years ago
0
DIN的训练结果

#52 Rosehan opened 5 years ago
2
为什么GAUC的值一直为0？请问self.mf_auc = tf.reduce_mean(tf.to_float(x > 0))这一句代码的含义是？

#51 eckolemon opened 5 years ago
0
关于attention的实现

#50 ZhengYukai666 closed 4 years ago
1
关于实际问题中的序列构造问题

#49 caojiajia opened 5 years ago
0
DIN的输入只包括商品、类目以及历史行为商品吗？

#48 waterbeach opened 5 years ago
1
DIN模型处理dense feature的相关问题

#47 eckolemon opened 5 years ago
1
请问必须显存>=10G才能跑吗~

#46 Slayerxxx opened 5 years ago
2
Questions about Wide and deep model

#45 czisok closed 4 years ago
1
din模型实际业务应用

#44 andrew-zzz opened 5 years ago
4
使用其他模型（如LR）作为对比试验时，是否也需要输入历史ID信息

#43 tinkle1129 opened 5 years ago
2