fitzwang

https://www.linkedin.com/in/fitzwang [email protected]

Tencent Shenzhen, Guangdong-province, China Engineer on big data and machine learning

Results 13 comments of


                                            fitzwang

99.9% percent of sample predicted label are -1

can you post you training hyper-parameter? the reason could be that the regularization parameters is so large that the model is under trained I gauss that the true positive/negative rate...

serving deepfm performance issue

I have several questions 1. how many dimension your models have? 2. how do you think 4ms latency? is it too long? we know why the CPU usage rate is...

serving one model multiple versions error

I got it. Since the "SharedConf" is single, that cause problem. We will remove some limitation to enable multiple model versions of the Angel.

serving deepfm error under high performance (fc_layer matrix null)

What do you mean by "sometimes"?

serving deepfm error under high performance (fc_layer matrix null)

can you post you graph json file?

Code work by ZhangYinHan

这是来自QQ邮箱的假期自动回复邮件。您好，我最近正在休假中，无法亲自回复您的邮件。我将在假期结束后，尽快给您回复。

multi-machine, multi-gpu sok core dump

这是来自QQ邮箱的假期自动回复邮件。您好，我最近正在休假中，无法亲自回复您的邮件。我将在假期结束后，尽快给您回复。

spark-on-angel-graph 编译报错

这是来自QQ邮箱的假期自动回复邮件。您好，我最近正在休假中，无法亲自回复您的邮件。我将在假期结束后，尽快给您回复。

编译问题，Maven build error

这是来自QQ邮箱的假期自动回复邮件。您好，我最近正在休假中，无法亲自回复您的邮件。我将在假期结束后，尽快给您回复。

使用doker的方式安装失败

这是来自QQ邮箱的假期自动回复邮件。您好，我最近正在休假中，无法亲自回复您的邮件。我将在假期结束后，尽快给您回复。

1
2
›