running example of logistic error about tf-encrypted

[vincehong]

Which script are you running and what is the error message ?

[vincehong]

hi! the code I have ran is the vertical of logistic of example in Gemini-lab the instruction is as follows: https://alibaba-gemini-lab.github.io/docs/blog/tfe/ I don't know why the program cannot run as expected,but there is no error information just stopped at this moment:

Your picture shows that server1 and server2 are halting and waiting for server0, this is normal. When you ran server0, the result will be printed on server0.

On very rare cases (I encountered), server0 will halt too, this may be because of bad network condition. Try choose other IP/ports, or use all the three workers on 127.0.0.1 to test if everything works.

[vincehong]

the config.json I have wrote is: { "server0": "machine1:43198", "server1": "machine2:27919", "server2": "machine3:27921" } the port I have chosen is by inputting netstat -a and find three random port is listening

Make sure you replace the "machinex" with the real IP.

[Barry-ljf]

hi!
the code I have ran is the vertical of logistic of example in Gemini-lab
the instruction is as follows:
https://alibaba-gemini-lab.github.io/docs/blog/tfe/
I don't know why the program cannot run as expected,but there is no error information just stopped at this moment:

[Barry-ljf]

the config.json I have wrote is:
{
"server0": "machine1:43198",
"server1": "machine2:27919",
"server2": "machine3:27921"
}
the port I have chosen is by inputting netstat -a and find three random port is listening

[Barry-ljf]

by the way, my directory set is as follows:
├── commom.py
├── CONFIG.json
├── machine1
│   ├── aliceTrainFile.csv
│   ├── common.py
│   ├── config.json
│   ├── pycache
│   │   └── common.cpython-36.pyc
│   └── training_alice.py
├── machine2
│   ├── bobLabel.csv
│   ├── config.json
│   └── training_bob.py
├── machine3
│   ├── config.json
│   └── training_server.py
├── training_alice.py
├── training_bob.py
└── training_server.py
and the tensorflow I used is 1.15.2

[Barry-ljf]

Finally I made it!
Thanks！！
when I use tfe their own code , just one command windows can realize training jointly , so I want to learn how to realize a code in this way(like three directories). hh~
and very thanks for helping!

...
Batch 349
[[-0.02972108]
[ 0.0405426 ]
[-0.00335315]
...

[Barry-ljf]

By the way, Cheng ge, I recently read faster secure adaptive Gradient Descent by your team. is the code of this essay open source?

[vincehong]

By the way, Cheng ge, I recently read faster secure adaptive Gradient Descent by your team. is the code of this essay open source?

Not yet, but MP-SPDZ proposed an improved version based on ours, and is open-sourced. See https://arxiv.org/abs/2107.00501?context=cs

[Barry-ljf]

thanks a lot ,I will go for it !