Training Initiation Problem (CONNECTION TIMED OUT)

[Dipankar1997161]

@biansy000 Hello,

I setup the training for my custom dataset with all the necessary parameters similar to the H36m parsed annotations provided.

After starting the training, I got the following error

[W socket.cpp:663] [c10d] The client socket has failed to connect to [::ffff:192.168.1.219]:23456 (errno: 110 - Connection timed out).
[W socket.cpp:663] [c10d] The client socket has failed to connect to [::ffff:192.168.1.219]:23456 (errno: 110 - Connection timed out).
[W socket.cpp:663] [c10d] The client socket has failed to connect to 192.168.1.219:23456 (errno: 110 - Connection timed out).

and finally stooped with

RuntimeError: The client socket has failed to connect to any network address of (192.168.1.219, 23456). The client socket has failed to connect to 192.168.1.219:23456 (errno: 110 - Connection timed out).

I used the following command to run the training

python scripts/train_smpl.py --cfg configs/NTOP.yaml --seed 123123 --nThreads 16 --launcher pytorch --rank 0 --exp-id 1   

so this should use the default tcp socket as given in the opt.py

can you tell me what needs to be done here..??

[Dipankar1997161]

solved