Optimization gets stuck

[thejobtan]

Hello, I've been running it for the past few months without issue and currently its just stuck on this, please help thanks!!

0%| | 0/100 [00:00<?, ?it/s]
Total progress: 100%|███████████████████████████| 16/16 [02:28<00:00, 9.31s/it]
strategy: 2-0-2: 100%|█████████▉| 99.99999999999976/100 [02:28<00:00, 1.49s/it]
0%| | 0/100 [02:28<?, ?it/s]
0%| | 0/100 [02:28<?, ?it/s]
0%| | 0/100 [02:28<?, ?it/s]
0%| | 0/100 [02:29<?, ?it/s]

strategy: 2-2-2: 100%|█████████▉| 99.99999999999976/100 [03:06<00:00, 2.21it/s]
strategy: 2-1-2: 100%|█████████▉| 99.99999999999976/100 [02:49<00:00, 26.26it/s]
strategy: 1-2-2: 100%|█████████▉| 99.99999999999976/100 [02:56<00:00, 66.38it/s]
Total progress: 19it [03:07, 9.85s/it]

[thejobtan]

my latest run came with an error

Users/jk/opt/anaconda3/envs/airsenalenv/lib/python3.12/multiprocessing/popen_fork.py:66: RuntimeWarning: os.fork() was called. os.fork() is incompatible with multithreaded code, and JAX is multithreaded, so this will likely lead to a deadlock.%| | 0/100 [00:00<?, ?it/s]
self.pid = os.fork() | 0/100 [00:00<?, ?it/s]
0%| | 0/100 [00:00<?, ?it/s]
Total progress: 100%|███████████████████████████| 16/16 [02:08<00:00, 8.01s/it]
0%| | 0/100 [02:08<?, ?it/s]
strategy: 2-1-2: 100%|█████████▉| 99.99999999999976/100 [02:08<00:00, 1.28s/it]
0%| | 0/100 [02:08<?, ?it/s]
0%| | 0/100 [02:08<?, ?it/s]
0%| | 0/100 [02:08<?, ?it/s]
0%| | 0/100 [02:08<?, ?it/s]
0%| | 0/100 [02:08<?, ?it/s]
0%| | 0/100 [02:08<?, ?it/s]

strategy: 1-2-2: 100%|█████████▉| 99.99999999999976/100 [02:19<00:00, 1.02it/s]

strategy: 2-0-2: 100%|█████████▉| 99.99999999999976/100 [02:09<00:00, 1.23s/it]

strategy: 2-2-2: 100%|████████▉| 99.99999999999976/100 [03:03<00:00, 794.75it/s]
Total progress: 19it [03:04, 9.72s/it]

[jack89roberts]

Hi @thejobtan - sorry it's taken a while to get back to you. The jax warning is something we need to look at but if the optimisation has started successfully I don't think it will be the problem.

This looks like a problem related to the free transfer rules changing, which we fixed a couple of months ago. Could you check what version of AIrsenal you're running with python -c "import airsenal; print(airsenal.__version__)"? If it's less than 1.9.0 could you trying pulling the latest version of the repo, reinstalling AIrsenal, and see if you have the same problem?

[brimell]

Hi @thejobtan I also had this issue and after I ran python -c "import airsenal; print(airsenal.__version__)" I got 1.8.0 so repulled the repo, removed the conda env conda remove --name airsenalenv --all and then did the whole reinstall process with conda env create. And it works now for me 🎉🎉

[Emengkeng]

Hi @thejobtan I also had this issue and after I ran python -c "import airsenal; print(airsenal.__version__)" I got 1.8.0 so repulled the repo, removed the conda env conda remove --name airsenalenv --all and then did the whole reinstall process with conda env create. And it works now for me 🎉🎉

Was facing the same issue, this solved it.