LongmaoTeamTf / deep_recommenders

Deep Recommenders
Apache License 2.0
325 stars 108 forks source link

google_tt模型中batch_softmax疑问 #4

Closed Andy1314Chen closed 3 years ago

Andy1314Chen commented 3 years ago
def corrected_batch_softmax(x, y, sampling_p=None):
    """logQ correction softmax"""
    correct_inner_product = log_q(x, y, sampling_p=sampling_p)
    return tf.math.exp(correct_inner_product) / tf.math.reduce_sum(tf.math.exp(correct_inner_product))

这里softmax计算的是不是有些问题呢? 原论文: image 中分母是x_i与batch内不同y_j点乘的和, 而代码中分母是batch内x_i与y_i点乘的和

GrogusBall commented 3 years ago

相关实现确实有问题,多谢支持~ v0.1.0版本的google_tt已经不在维护,请关注v1.1.x版本~ 双塔实验,请参考Demo实验: https://github.com/LongmaoTeamTf/deep_recommenders/blob/master/experiments/deep_retrieval.ipynb

Andy1314Chen commented 3 years ago

相关实现确实有问题,多谢支持~ v0.1.0版本的google_tt已经不在维护,请关注v1.1.x版本~ 双塔实验,请参考Demo实验: https://github.com/LongmaoTeamTf/deep_recommenders/blob/master/experiments/deep_retrieval.ipynb

感谢回复~