-
This esmm don't have the share embedding layer?
-
你好最近我在学习多任务推荐,我看到esmm的方法是通过pcvr*pctr计算出pctcvr,并对pctr和pctcvr进行损失计算训练网络,但是在esmm的模型中我只看出来计算了pctr和pcvr,好像并没有计算pctcvr?
-
On required dropdowns the red asterisk is visible even when the dropdown is open.
-
-
感觉x-deeplearning 中auc 计算batch auc,把其中只包含negative的batch 算作invalid 抛弃,这个做法跟tensorflow里面做法不一样。会比较大的影响auc的计算,因为一个batch中没有positive 也会影响全局的FP,auc应该算全局的。
具体请看: https://github.com/alibaba/x-deeplearning/issu…
-
train
![image](https://github.com/PaddlePaddle/PaddleRec/assets/28783826/9dcb8e7e-f49f-41cf-add3-fb2e038b99c7)
infer
![image](https://github.com/PaddlePaddle/PaddleRec/assets/28783826/389e154b-43…
-
我用gen_ctr.py产生了100K用户数据,运行ESSM任务。
python main.py --task_name=mtl --seed=100 --model_name=esmm --dataset_path="ctr_data_100000.csv" --train_batch_size=4096 --val_batch_size=4096 --test_batch_size=4096…
-
你好,请问原始训练集经过处理后(reader.py)样本格式是 index,click_label,conversion_label,field_index:feature_index... 此处并没有用到特征具体的value值。可参照PaddleRec/models/multitask/esmm/data/train/small.txt目录下的测试样本。为什么没有特征的val,此处是把所有的…
-
不太清楚这个 reader.py是怎么实现的
PaddleRec/models/multitask/esmm/data/run.sh
-
我看您使用的是csv文件输入,现在我也是这样做的,只不过csv文件中有多值特征(例如,以'|'分隔),请问有什么好的处理方法吗?尝试了tf.string_split,但提示'SparseTensor is not supported'