jwohlwend
Stuck with 600aa protein
12th Gen Intel(R) Core(TM) i9-12900K NVIDIA GeForce RTX 4090 Ubuntu 22.04 64bit
Hello!
-
I have to run Boltz iteratively for different sequences (all size 600aa), but it gets stuck frequently with the following message, I'd be happy to know if its a memory issue or something else:
PENDING: 0%| Sleeping for 6s. Reason: PENDING | 0/300 [elapsed: 15:12:34 remaining: ?] PENDING: 0%| Sleeping for 6s. Reason: PENDING | 0/300 [elapsed: 15:12:41 remaining: ?] PENDING: 0%| Sleeping for 7s. Reason: PENDING | 0/300 [elapsed: 15:12:47 remaining: ?] PENDING: 0%| Sleeping for 6s. Reason: PENDING | 0/300 [elapsed: 15:12:54 remaining: ?] PENDING: 0%| Sleeping for 9s. Reason: PENDING | 0/300 [elapsed: 15:13:01 remaining: ?] PENDING: 0%| Sleeping for 8s. Reason: PENDING | 0/300 [elapsed: 15:13:10 remaining: ?] PENDING: 0%| Sleeping for 9s. Reason: PENDING | 0/300 [elapsed: 15:13:18 remaining: ?] PENDING: 0%| Sleeping for 7s. Reason: PENDING | 0/300 [elapsed: 15:13:28 remaining: ?] PENDING: 0%| Sleeping for 6s. Reason: PENDING | 0/300 [elapsed: 15:13:35 remaining: ?] -
While saving model cache, I use the same directory for all the iteration predictions, is this approach correct?
Thank you, Hassan
Hello, I also encountered such an error during use. Have you solved it? Thank you
I'm having the same issue on my end.
It happens when I run multiple de novo sequences using the same cache directory. Restarting from the next sequence sometimes helps, but the error reappears randomly.
Environment: Ubuntu 22.04
CPU:
Hope this helps. Thanks!
Has this issue been resolved / solved? I am having a similar issue.
From what I understand, the issue comes from multiple inference requests using the same model cache file path? I have not tried this, but does anyone have empirical evidence that this works?
