Could you please provide the steps for running the project?

[jibanli]

[Forcrush]

Sorry to reply so late, I didn't check this repo regularly. The main.py runs well in my console. I think maybe the version is the crux of ur problem. Python 3.6 or above should work.

[jibanli]

Hi. Could you give me any example for using the model already trained to predict intrusion? I want to input some data to predict whether it's anomaly or not.

[Forcrush]

Hi, I'm a little confucious about your demand. The model can be different on different dataset, how could I provide a model fitting for your data without knowing even the feature dimension, let alone the correlation. Did u really check the code?

…

------------------&nbsp;原始邮件&nbsp;------------------ 发件人:&nbsp;"jibanli"<[email protected]&gt;; 发送时间:&nbsp;2019年12月30日(星期一) 晚上8:48 收件人:&nbsp;"Forcrush/Efficient-GAN-based-method-for-cyber-intrusion"<Efficient-GAN-based-method-for-cyber-intrusion@noreply.github.com&gt;; 抄送:&nbsp;"鬱"<[email protected]&gt;;"Comment"<[email protected]&gt;; 主题:&nbsp;Re: [Forcrush/Efficient-GAN-based-method-for-cyber-intrusion] Could you please provide the steps for running the project? (#1) Hi. Could you give me any example for using the model already trained to predict intrusion? I want to input some data to predict whether it's anomaly or not. — You are receiving this because you commented. Reply to this email directly, view it on GitHub, or unsubscribe.

[jibanli]

Sorry I didn't express clearly and I just started learning tensorflow.

For the data 'kddcup.data_10_percent_corrected', I want to input data like below

0,icmp,ecr_i,SF,1032,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,511,511,0.00,0.00,0.00,0.00,1.00,0.00,0.00,255,255,1.00,0.00,1.00,0.00,0.00,0.00,0.00,0.00,smurf.(this is now in the kddcup.data_10_percent_corrected)

and use the model(# save training logs, weights, biases, etc.
def create_logdir(method, weight, rd):
return "bigan/train_logs/kdd/{}/{}/{}".format(weight, method, rd)
to predict whether its normal or smurf.

[jibanli]

[jibanli]

说中文？？？

[jibanli]

像这样的 https://stackoverflow.com/questions/58833810/loading-the-cnn-model-and-predict-the-csv-file
想用GAN网络做预测。。。

[Forcrush]

呃 预测样本标签的步骤已经写在代码里面了呀， 如果你是基于repo里面已经有的数据集 那么直接看代码就好了呀 比如kdd 样本如何拿来预测的步骤是在run_kdd.py这个文件里面。至于你所给的数据维度这些 你可以参考代码里面的dummy encoding方法或者你自己也可以用其他方法处理离散维度 （但是要注意输入输出维度一致 可能需要你自己在 kdd_utilities.py里面改）至于你要的已经训练好的模型 repo里面没有 因为模型文件比较大所以当时没有上传而且也没必要 因为如果你环境正确 可以直接run main.py训练 （我已经测试过 代码的运行应该没有任何问题 至于你要处理自己的其他数据集或者修改特征的扩维方式 你恐怕得自己修改相应的地方）

…

------------------&nbsp;原始邮件&nbsp;------------------ 发件人:&nbsp;"jibanli"<[email protected]&gt;; 发送时间:&nbsp;2019年12月31日(星期二) 上午9:12 收件人:&nbsp;"Forcrush/Efficient-GAN-based-method-for-cyber-intrusion"<Efficient-GAN-based-method-for-cyber-intrusion@noreply.github.com&gt;; 抄送:&nbsp;"鬱"<[email protected]&gt;;"Comment"<[email protected]&gt;; 主题:&nbsp;Re: [Forcrush/Efficient-GAN-based-method-for-cyber-intrusion] Could you please provide the steps for running the project? (#1) 像这样的 https://stackoverflow.com/questions/58833810/loading-the-cnn-model-and-predict-the-csv-file 想用GAN网络做预测。。。 — You are receiving this because you commented. Reply to this email directly, view it on GitHub, or unsubscribe.

[jibanli]

谢谢好多字的回复。

[Forcrush]

应该的 如果对paper里面提到的模型本身或者其他地方有不太清楚的地方或者发现错误 乐意一起讨论 :)

…

------------------&nbsp;原始邮件&nbsp;------------------ 发件人:&nbsp;"jibanli"<[email protected]&gt;; 发送时间:&nbsp;2019年12月31日(星期二) 中午1:55 收件人:&nbsp;"Forcrush/Efficient-GAN-based-method-for-cyber-intrusion"<Efficient-GAN-based-method-for-cyber-intrusion@noreply.github.com&gt;; 抄送:&nbsp;"鬱"<[email protected]&gt;;"Comment"<[email protected]&gt;; 主题:&nbsp;Re: [Forcrush/Efficient-GAN-based-method-for-cyber-intrusion] Could you please provide the steps for running the project? (#1) 谢谢好多字的回复。 — You are receiving this because you commented. Reply to this email directly, view it on GitHub, or unsubscribe.

[jibanli]

头像哪里搞的呀？好看

[Forcrush]

hhh 带土和琳呀&nbsp;

…

------------------&nbsp;原始邮件&nbsp;------------------ 发件人:&nbsp;"jibanli"<[email protected]&gt;; 发送时间:&nbsp;2019年12月31日(星期二) 下午5:18 收件人:&nbsp;"Forcrush/Efficient-GAN-based-method-for-cyber-intrusion"<Efficient-GAN-based-method-for-cyber-intrusion@noreply.github.com&gt;; 抄送:&nbsp;"鬱"<[email protected]&gt;;"Comment"<[email protected]&gt;; 主题:&nbsp;Re: [Forcrush/Efficient-GAN-based-method-for-cyber-intrusion] Could you please provide the steps for running the project? (#1) 头像哪里搞的呀？好看 — You are receiving this because you commented. Reply to this email directly, view it on GitHub, or unsubscribe.

[jibanli]

没看火影。。。

我找到这个代码
https://github.com/SnowMasaya/Efficient_GAN-Based_Anomaly_Detection_Keras

你这个也可以用keras写吧？
keras模型预测简单