shenweichen/

DeepCTR

Easy-to-use,Modular and Extendible package of deep-learning based CTR models.

Owners:

$0.00

USD raised

Recent activities

shenweichen added the repository

almost 6 years ago

Issues

Funded Unfunded Rewarded Closed

use din error？

Unfunded#410created bylonngxiang

$0.00

训练集不平衡的问题

Unfunded#381created bywhk6688

$0.00

How can I concat two embeddings in sparse feature with DeepFM

Unfunded#382created byllfzllfz

$0.00

Estimator 如何保存模型，没有找到对应的例子

Unfunded#386created bywhk6688

$0.00

run_multivalue_movielens_hash.py中对string类型的变长序列pad_sequences，padding后的结果无法理解

Unfunded#388created byrogerxsj

$0.00

run_multivalue_movielens_hash.py中对string类型的变长序列pad_sequences，padding后的结果无法理解

Unfunded#389created byrogerxsj

$0.00

RuntimeError: Unable to create link (name already exists

Unfunded#397created byeshaoliu

$0.00

请问DIN的离散型特征支持字符串类型吗

Unfunded#398created byKatouMegumiMoe

$0.00

new methods

Unfunded#73created byyufengwhy

$0.00

Multi-value Input 不支持全部都是 Nan, 即 all 0 的情况

Unfunded#139created bycivilman628

$0.00

如何保存原始数据 (raw data) 和用于模型输入数据 (encoding data) 的映射关系？

Unfunded#140created bycivilman628

$0.00

Eastimator 中增加 loggingTensorHook 以便在日志中输出 loss

Unfunded#401created byworkingloong

$0.00

ValueError: A `Concatenate` layer should be called on a list of at least 2 inputs

Unfunded#404created bylonngxiang

$0.00

FM的实现疑问

Unfunded#405created byxuanzebi

$0.00

run_classification_criteo.py error in tf1.2

Unfunded#380created byzanshuxun

$0.00

调用DeepCTR的DIN模型中的example出错

Unfunded#411created byechohandsome

$0.00

sparse_feature中embeddings_initializer维度如何设置？

Unfunded#412created byfxxxxxxz

$0.00

CustomMaskWarning

Unfunded#413created byhaywertz

$0.00

DIN构建模型时报错

Unfunded#414created byMotsen

$0.00

How does DeepFM handle values of a feature not seen in training set?

Unfunded#415created byjoelpv25

$0.00

Why both CrossNet and CrossMix use x_l = x_0 when making feature crossing?

Unfunded#416created byxuChenSJTU

$0.00

run_dien的数据集问题

Unfunded#417created bywu-ruijie

$0.00

How to share embeddings among features?

Unfunded#418created byjoelpv25

$0.00

MTL多目标损失函数权重设置

Unfunded#421created bywushupei1997

$0.00

请教下大佬几个deepctr din模型的问题

Unfunded#422created bynmslandwsnd

$0.00

请教大佬，怎么在建模时使用预训练的embedding呢？

Unfunded#423created byYukeyy

$0.00

请教下大佬几个deepctr din模型的问题

Unfunded#424created bynmslandwsnd

$0.00

Change this DeepCTR model to fit multi-classification task

Unfunded#425created bylearninlearn

$0.00

sparse_embedding_list is empty and gives an error A `Concatenate` layer should be called on a list of at least 2 inputs

Unfunded#118created byshriramhr

$0.00

Python 3.8 什么时候支持？

Unfunded#352created bytigerneil

$0.00

使用DeepFMEstimator，怎么打印训练过程每轮的loss和auc情况

Unfunded#355created by1980695671

$0.00

When I create multiple dnn hidden layers using a list, it will only produce 1 dnn layer

Unfunded#356created byelysego-kumu

$0.00

DeepFMEstimator 保存模型时是否要同时保存 linear_feature_columns 和 dnn_feature_columns？

Unfunded#357created byFrancisKong

$0.00

训练一段时间后，会被kill掉，模型是DCN模型

Unfunded#358created byWujiangXu

$0.00

Missing parameter 'training' during DICE activation function call

Unfunded#359created bywc591723150

$0.00

DIN model misses parameters 'training' when calling the DICE activation function DIN模型在调用DICE时缺少参数training

Unfunded#360created bywc591723150

$0.00

GPU Utilization

Unfunded#49created bydeepkshikha

$0.00

如何在此项目里使用经过one-hot编码的YOYI数据集

Unfunded#79created byqlz1995

$0.00

请问DeepFM能支持多维的dense feature输入吗

Unfunded#81created bysunqiang85

$0.00

请问如何使用梯度截断功能呢？

Unfunded#85created byzhuangjiayue

$0.00

How do you solve the cold-start problem in DSIN?

Unfunded#90created byvikasmalhotra08

$0.00

some questions about the dsin example

Unfunded#113created byyouchika

$0.00

使用din模型的时候，y_pred会出现nan，

Unfunded#426created byzhangxueren9

$0.00

请问支持超大维度的uid和itemid么？

Unfunded#124created bysneffort

$0.00

参考样例使用自定义的模型输入时，DIEN无法正常建模

Unfunded#130created bymoneyml

$0.00

question in DIEN example

Unfunded#131created bybetter629

$0.00

数据处理

Unfunded#132created byccfccl

$0.00

How to get the embedding vector of all FM models?

Unfunded#136created bypursueorigin

$0.00

关于deepfm的FM()输入的疑问

Unfunded#155created byDoubleYing

$0.00

多分类实现的困惑

Unfunded#166created bystupidHIGH

$0.00

为啥保存个模型要半小时，模型大小只有不到10m

Unfunded#177created byyyy1987

$0.00

Fail to save model with Transformer layers with use_positional_encoding=True.

Unfunded#365created byworkingloong

$0.00

请问下多谢，有个CTR特征的问题问下，就是如果挑3-5个最有效的特征，商家列表排序，选哪几个特征？商家类型、用户对商家类型的偏好，还有？

Unfunded#368created byguotong1988

$0.00

feature_column 支持embedding向量特征输入吗？

Unfunded#372created bylonngxiang

$0.00

读取大文件CSV的时候怎么办

Unfunded#374created bymaoyaozong

$0.00

sincos位置编码问题

Unfunded#379created bychenxli

$0.00

java调用模型的问题

Unfunded#162created bysunjiadong

$0.00

sequence.py transformer 关于多个 layer_norm 参数 name 重复的问题

Unfunded#496created byBlackcOVER

$0.00

tf2.4保存模型失败

Unfunded#498created byArcturusMensk

$0.00

什么是LinkageEmbedding

Unfunded#156created byliyinan1001

$0.00

bilstm

Unfunded#141created byshidaide2016

$0.00

GPU out of memory with reasonable data

Unfunded#29created byLiChen2012

$0.00

Training model in multi gpu environment

Unfunded#55created bykhalidhuseynov

$0.00

你好，请问支持多分类问题吗

Unfunded#71created byxingyuezhiji

$0.00

Varlen feature model save error

Unfunded#120created bykhlee88

$0.00

get_varlen_pooling_list without mask parameter when supports_masking=True

Unfunded#123created bydulm

$0.00

Multiclass classfication task

Unfunded#128created bylicesonw

$0.00

标准化的疑问

Unfunded#147created byDJy4ever

$0.00

din能否支持pandas的输入？

Unfunded#151created by2shou

$0.00

DIN模型，run_din遇到的问题

Unfunded#158created byDoubleYing

$0.00

sequence.py 文件第535行bug

Unfunded#494created byBlackcOVER

$0.00

How to build prediction system with DeepCTR?

Unfunded#197created bybatrlatom

$0.00

Why is the parameter for the embedding size not present in case of fibinet?

Unfunded#164created byrajeshshrestha

$0.00

自动驾驶更新笔记 Autopilot Updating Notes

Unfunded#504created bygotonote

$0.00

DIN等模型为什么对目标物品和历史物品的embedding分别建立矩阵？与原论文实现不符

Unfunded#507created byTangJiakai

$0.00

How to get expert utilization for multiple tasks of mmoe?

Unfunded#512created bylunachy

$0.00

关于导入deepctr包后程序执行完不退出的问题

Unfunded#513created byT0L0ve

$0.00

Linear logic in DCNMIX

Unfunded#523created byidoblankfiverr

$0.00

如何保存deepctr-torch训练好的deepfm模型

Unfunded#527created byRainFung

$0.00

多值特征代码有bug

Unfunded#531created byshelson-woo

$0.00

model.predict only support np.array ?

Unfunded#533created byRainFung

$0.00

feature interaction visualization

Unfunded#537created bybyunsunyoung

$0.00

StagingError: in user code: IndexError: pop from empty list

Unfunded#541created byfood-bin

$0.00

deepctr.models import error

Unfunded#546created bybrokendays611

$0.00

tensorflow.python.framework.errors_impl.InvalidArgumentError: Cannot assign a device for operation model/transformer/position_encoding/embedding_lookup

Unfunded#460created bybill2766

$0.00

unsupported operand type(s) for +: 'int' and 'str'

Unfunded#427created bykamskanagi

$0.00

run_estimator_tfrecord_classification.py AUC is 0.49

Unfunded#429created byuniversewill

$0.00

tf2.7版本报错

Unfunded#432created bymoseshu

$0.00

deepfm network structure difference

Unfunded#434created byPengtai9928

$0.00

小白没懂SparseFeat和VarLenSparseFeat的区别，求大神指点啊

Unfunded#440created byguotong1988

$0.00

小白弱问下linear_feature和dnn_feature的区别是？

Unfunded#441created byguotong1988

$0.00

训练阶段最大最小归一化后，能否将最大最小值保存，便于预测阶段使用

Unfunded#447created by930083287

$0.00

请问m1 mac怎么安装？python版本3.9，tensorflow版本2.5.0

Unfunded#449created bybill2766

$0.00

xDeepFM网络的CIN层数是否可调？

Unfunded#450created byuser-120

$0.00

w&d的模型是否总是过拟合

Unfunded#451created byuser-120

$0.00

单机单卡GPU环境利用率低

Unfunded#452created by930083287

$0.00

这是keras版本不对吗？

Unfunded#453created byDimitri666

$0.00

run_dien.py test reports "localhost/dnn/bias0/N10tensorflow3VarE does not exist"

Unfunded#458created bymingxiaoh

$0.00

WDL 模型进行TF serving报错

Unfunded#351created byHanlos

$0.00

run_classification_criteo遇到TypeError: Cannot convert a symbolic Keras input/output to a numpy array.问题

Unfunded#462created byVinnyHu

$0.00

sparsefeat densefeat 线上预测使用时需不需要处理这一步？

Unfunded#468created byzaibian

$0.00

It is available to design the loss function by myself? If so, how can i change the loss function?

Unfunded#44created byJimmy880

$0.00

load_model 出现 AttributeError: 'Concatenate' object has no attribute 'outbound_nodes'

Unfunded#127created bywziji

$0.00

Wrong practice when using SequencePoolingLayer in inputs.get_varlen_pooling_list

Unfunded#474created byivan-ge677

$0.00

增加多个行为序列出错

Unfunded#475created bylittleshutong

$0.00

相同环境运行xDeepFM报错，运行DeepFM是好的

Unfunded#481created byArtificialZeng

$0.00

用xDeepFM和DeepFM模型，结果几乎一样，甚至稍低

Unfunded#482created byArtificialZeng

$0.00

官网deepfm example运行报错AttributeError

Unfunded#484created bySYSU-Linxp

$0.00

h5py==3.0.0 causes model save issues in tf 2.1.0

Unfunded#487created byzanshuxun

$0.00

关于 DCN 模型实现中，LR部分的疑问

Unfunded#488created byLALBJ

$0.00

Embedding size in dien

Unfunded#165created byrajeshshrestha

$0.00

Is values calculate way right in layers.sequence.transformer?(Transformer里似乎V的计算有一点点问题吧)

Unfunded#490created byDarryYoung

$0.00

run_dssm_neg_sample.py

Unfunded#243created bynon-swimmer

$0.00

加载保存好的模型然后预测返回的结果全是 nan

Unfunded#224created bymiaochaowei

$0.00

加载模型后，预测结果都是0

Unfunded#226created byrickyhuw

$0.00

关于FLEN模型的问题

Unfunded#228created bydiwuxiaoxu

$0.00

训练WDL模型使用 t f.saved_model.save()保存时报错。

Unfunded#229created bysangyongjia

$0.00

稀疏编码one-hot保持一致性

Unfunded#231created byaiot-tech

$0.00

FGCNN

Unfunded#235created byyyyzhao

$0.00

ValueError: Failed to convert a NumPy array to a Tensor (Unsupported object type numpy.ndarray).

Unfunded#236created byNDS-98

$0.00

How i change the Initializers

Unfunded#61created byInkdyeHuang

$0.00

SparseFeat import failed

Unfunded#98created byjt120

$0.00

Use of linear_feature_columns and dnn_feature_columns

Unfunded#111created byulrych8

$0.00

怎么把 estimator 和tfrecord 和这个框子联动起来呢

Unfunded#88created byBigAN

$0.00

InvalidArgumentError错误求助

Unfunded#116created bydkicenan

$0.00

urllib3.connectionpool:DEBUG:https://pypi.org:443 "GET /pypi/deepctr/json HTTP/1.1" 200 7556

Unfunded#137created byzkailinzhang

$0.00

训练DeepFM时指定了loss，但是没有用。

Unfunded#223created byelzizyel

$0.00

feature_column import error problem

Unfunded#245created bykang1916

$0.00

DeepFM embedding concat 维度错误

Unfunded#248created bywujunzhuo

$0.00

Why are sparse features treated as 1-dim embeddings in WND?

Unfunded#249created byhibayesian

$0.00

How to get prediction result with identification information(e.g. ID)?

Unfunded#250created byhibayesian

$0.00

DeepFM运行时出错

Unfunded#251created byRobbieing

$0.00

feature_column和inputs有循环依赖

Unfunded#252created byzscwind

$0.00

deepfm.py ValueError: '_emb_feature' is not a valid scope name

Unfunded#256created bySue52nlp

$0.00

请问下DeepFM如何实现为每个用户推荐物品列表？

Unfunded#257created byRobbieing

$0.00

deepfm.py error

Unfunded#258created byfrajabi

$0.00

Error, just run "run_classification_criteo.py" in examples, tf version == 2.3.0

Unfunded#259created byxinyu009

$0.00

SparseFeat函数的embedding_dim设为'auto'时报错

Unfunded#260created bylinmengnan

$0.00

训练时正则loss变成nan了是怎么回事？

Unfunded#261created bychenjinting123

$0.00

如何正确运行example中的程序？

Unfunded#262created byrobinren03

$0.00

cannot import name 'get_fixlen_feature_names' from 'deepctr.inputs

Unfunded#202created byYutingSun163

$0.00

Linear logit in DeepFM - embedding dim

Unfunded#178created byornellamarciano

$0.00

How to edit learning rate?

Unfunded#168created byEagleYing

$0.00

"run_dien" leads to ValueError for 'seq_length'

Unfunded#186created byCurlykonda

$0.00

What's the feature "item_gender" in DIN / DIEN?

Unfunded#187created byCurlykonda

$0.00

在deepctr/models/din.py中，第86-88行，输入只有query,keys，为什么没有keys_length，或者keys_length怎么计算得到的

Unfunded#188created bywuyunxiangwyx

$0.00

DIN run_din.py执行失败

Unfunded#189created byjingbao12

$0.00

训练完DeepFM模型后，加载权重，输入参数不变的情况下每次predict结果不一致，经调试发现是由deepctr.layers.utils.Linear导致

Unfunded#190created bywuyunxiangwyx

$0.00

ZeroDivisionError - Fibinet

Unfunded#191created bypavansvn

$0.00

example中DIN、DIEN中的score代表什么

Unfunded#192created bywangwenyu01

$0.00

训练从第二个epoch开始后面的loss不变，是什么原因？

Unfunded#193created byHelloWorld19930113

$0.00

Unable to train on very simple criteo like data

Unfunded#196created bybatrlatom

$0.00

DIEN报错ValueError: The name "seq_length" is used in 2 times in the model. All layer names should be unique.

Unfunded#200created byMildAdam

$0.00

Where to download the criteo_sample.txt?

Unfunded#201created byvba34520

$0.00

A merge layer should be called on a list of inputs

Unfunded#263created by1073521013

$0.00

你这个模型在进行推理的时候性能很差，不知道是什么原因

Unfunded#203created byrickyhuw

$0.00

可以提供estimator接口么~

Unfunded#205created bywannibar

$0.00

保存模型的时候会失败，比如model.save("./data/")？

Unfunded#206created byhecongqing

$0.00

如何在现有tf-gpu版本基础上安装deepctr

Unfunded#207created byshm007g

$0.00

从v0.7.1升级到v0.7.4，model.fit_generator接口报错

Unfunded#208created bycddesire

$0.00

这个不就是tf.keras.layers.Dense吗，为啥要重新写一遍？

Unfunded#209created byJie-Yuan

$0.00

运行失败

Unfunded#213created byyyyzhao

$0.00

AutoInt模型疑问

Unfunded#214created byWegenPan

$0.00

about auto embedding size

Unfunded#215created byNemoSailingHeart

$0.00

which model is the basemodel (Embedding + MLP)

Unfunded#216created byjimmylao

$0.00

Maybe this is a mistake

Unfunded#219created byTanTingyi

$0.00

把DeepFM改成CCPM，报错DenseFeat is not supported in dnn_feature_columns

Unfunded#220created byjc-moon

$0.00

supporting pretrain embedding ?

Unfunded#221created bylsq357

$0.00

话说现在用TensorFlow分布式训练CTR模型怎么搞速度快啊？

Unfunded#336created byguotong1988

$0.00

请问deepfm训练的模型，在tfserving部署后，能不能提供grpc or https预测的demo？构造和训练数据一样的输入，传入tfserving一直报错

Unfunded#320created byRONG-SONG

$0.00

csv转tfrecords,送入模型失败，提示: Data type int64, but except type float

Unfunded#321created byNicolezjy

$0.00

这样操作是否是增量训练

Unfunded#322created bywhk6688

$0.00

SparseFeat(feat, vocabulary_size=train_df[feat].nunique(),embedding_dim=32) 用java实时预测有bug

Unfunded#323created bywhk6688

$0.00

数据输入不支持TFRecord，无法大规模应用，不知道作者有没有想法

Unfunded#324created byzhenghang

$0.00

tensorboard scalar只显示loss，没有显示其它的metrics

Unfunded#328created bypengyuan

$0.00

有没有人会这个入门问题？多谢多谢多谢

Unfunded#329created byguotong1988

$0.00

attention layer weights in DIN

Unfunded#330created byminarastgar

$0.00

训练过慢

Unfunded#331created bywhk6688

$0.00

关于模型保存加载部署问题

Unfunded#332created bylonngxiang

$0.00

Unfunded#333created byguotong1988

$0.00

我小白，请问下，特征筛选这事是不是看加了某特征离线auc涨了就行？如果不能上线实验

Unfunded#334created byguotong1988

$0.00

话说现在用TensorFlow的PS分布式架构用CPU训练CTR模型这条路速度行吗？

Unfunded#335created byguotong1988

$0.00

使用DeepFMEstimator，GPU物理机 , 速度不增反降（每轮40分钟），训练数据大概35G

Unfunded#314created bywhk6688

$0.00

请问一下如何做多模型集成

Unfunded#52created byzhouhang95

$0.00

Would you please explain VarLenFeat's parameters

Unfunded#72created byshowkeyjar

$0.00

请问关于preprocess_input_embedding函数的输出结果

Unfunded#83created byweiyang112

$0.00

使用deepctr 包 DCN 训练模型没有报错，但是 model = tf.keras.models.load_model('model_657.h5') 会报错， ValueError: Unknown layer: DNN

Unfunded#107created byhuminwudi

$0.00

用deepfm为什么每次跑出来的结果都不一样，有什么方法可以使他们一样？

Unfunded#101created byhuminwudi

$0.00

关于FiBiNet bilinear_type为Field-Interaction Type时输出向量尺寸不同的问题

Unfunded#103created byae2leowen

$0.00

Din模型调用报错

Unfunded#117created bywaterbeach

$0.00

SingleFeat and VarLenFeat removed?

Unfunded#105created byCarlosVaquero

$0.00

离散数据编码问题

Unfunded#339created bydiwuxiaoxu

$0.00

建议序列模型din、disn等可增加有数据集的案例

Unfunded#342created bylonngxiang

$0.00

keras 的batchnormal 参数训练过程中即使设置tranable=true ,但是预测时还是有问题的，你们的dice激活函数中使用了这个类，所以这个类是不是也有问题

Unfunded#343created byareafather

$0.00

Can not save the model with *tf* format because in the *DNN* layer modify the list

Unfunded#344created bydengc367

$0.00

请教一下为什么DCN最后有Linear的部分？原论文不是只有cross和deep的部分吗？

Unfunded#347created byBenLucas7

$0.00

FLEN - Linear Component (S) of FWBI missing in the model ?

Unfunded#289created bysaoanuj89

$0.00

PNN InnerProductLayer & OutterProductLayer中构建p、q逻辑错误

Unfunded#264created byYang-Lv

$0.00

lbe = HashEncoder()

Unfunded#266created bychen849157649

$0.00

AttentionSequencePoolingLayer类，call()中未传递mask参数。

Unfunded#267created bycodefish1990

$0.00

关于DeepFM中FM的实现问题

Unfunded#270created byGforky

$0.00

Error on loading saved model

Unfunded#275created byPavelevivan

$0.00

DeeoFM模型run_classification_criteo.py文件

Unfunded#56created bywangtao666666

$0.00

DIN模型样例代码run_din.py运行错误

Unfunded#112created bywaterbeach

$0.00

损失函数

Unfunded#277created bydiwuxiaoxu

$0.00

Converting sparse IndexedSlices to a dense Tensor of unknown shape. This may consume a large amount of memory

Unfunded#278created byMesilenceki

$0.00

Low precision of classifiers

Unfunded#282created bymjuarezm

$0.00

Can we standardize the DenseFeat when building the models.

Unfunded#284created byworkingloong

$0.00

Error - Failed to convert a NumPy array to a Tensor (Unsupported object type list).

Unfunded#286created bypythonmjs

$0.00

run_estimator_tfrecord_classification 不具有可复现性

Unfunded#287created byluckmoon

$0.00

Training the DeepFM fails if we choose different embedding_dim for each feature ...

Unfunded#184created bybachephysicdun

$0.00

单机多卡训练问题

Unfunded#290created byMandule

$0.00

单机多卡训练问题

Unfunded#291created byMandule

$0.00

没有sparse特征，只有dense特征时，DeepFMEstimator报错

Unfunded#295created byFanoid

$0.00

请问deepctr支持处理imbalanced数据集吗？怎么处理

Unfunded#296created byjqsl2012

$0.00

关于DeepFM中Sparse特征的Embedding过程。

Unfunded#298created bySTHSF

$0.00

Maybe exposing some model hyper-parameter options to the user?

Unfunded#306created byamznero

$0.00

can not use deepctr

Unfunded#307created byhallucination11

$0.00

[数据集下载] Criteo CTR prediction

Unfunded#308created byhuangbiubiu

$0.00

call() missing 'state' argument when saving DIEN model

Unfunded#310created byLitchilitchy

$0.00

deepctr0.8.2 DSIN怎样在淘宝ctr数据集上跑起来

Unfunded#311created byArclabs001

$0.00

SparseFeat 的用法

Unfunded#312created bywhk6688

$0.00

特征统一编码

Unfunded#313created bywhk6688

$0.00

Recent activities

shenweichen added the repository

almost 6 years ago

DeepCTR

$0.00

Issues

Company

Community

Support

Connect