“single positional indexer is out-of-bounds”

[fuentesdt]

Results seem to be there but getting this “single positional indexer is out-of-bounds” error?

https://github.mdanderson.org/dtfuentes/hccgwas/blob/main/tumortraining.csv

https://github.mdanderson.org/dtfuentes/hccgwas/blob/main/README#L15

$ git log commit 47739e6b595e60183d06edeaba3560c83b8243c5 (HEAD -> main, origin/main, origin/HEAD) Author: David Fuentes fuentesdt@gmail.com Date: Fri Oct 18 17:58:05 2024 -0500

ENH: pdb

commit 39fc20a71aa5ea511ae68da445de7d794467493e Merge: 3fdfcfa 172075f Author: David Fuentes fuentesdt@gmail.com Date: Wed Oct 16 17:07:49 2024 -0500

Merge pull request #5 from mist-medical/main

updates

commit 172075fcb63d727eefa37dbe5b401246e41cd2a7 Merge: f1ef31b da0f83e Author: Adrian Celaya aec8@rice.edu Date: Thu Oct 10 18:37:30 2024 -0500

Merge pull request #49 from aecelaya/main

Refactor evaluation to production-quality code. Fix bug in computing

average surface distance.

innovador$ pwd /rsrch3/ip/dtfuentes/github/hccgwas innovador$ head misttrainresults/results.csv id,TUMOR_dice,TUMOR_haus95 id0,0.0510545909108354,231.22499864850255 id1,0.372221174119369,99.24212815130478 id2,0.9028247507823778,115.55950847939775 id3,0.3408212879340357,164.72401160729422 id4,0.5599213003702536,25.632011235952593 id5,0.5025447338794611,122.75585525749882 id6,0.113971721374634,82.29823813423954 id7,0.08889950701065076,107.34057946554975 id8,0.3863996992791348,35.90264614203248 . . .

From: Twam,Awj K ATwam@mdanderson.org Sent: Monday, October 21, 2024 11:16 AM To: Fuentes,David Thomas Alfonso DTFuentes@mdanderson.org Subject: Re: crash

I never got that specific index error- but whenever it crashed out without completing evaluation, it was mainly due to GPU usage. I think the error may be something in the dataset you are using.

---

From: Fuentes,David Thomas Alfonso [DTFuentes@mdanderson.org](mailto:DTFuentes@mdanderson.org) Sent: Monday, October 21, 2024 11:06 AM To: Twam,Awj K [ATwam@mdanderson.org](mailto:ATwam@mdanderson.org) Subject: crash

Awj, didn’t you have this problem before? Finishes all training and testing then crashes ? How did you fix ?

$ /opt/apps/miniforge/mistOct2024/bin/mist_train --data /rsrch3/ip/dtfuentes/github/hccgwas/misttrain/dataset.json --numpy /rsrch3/ip/dtfuentes/github/hccgwas/misttrainnumpy --results /rsrch3/ip/dtfuentes/github/hccgwas/misttrainresults --gpus 0 --amp --pocket --epochs 200

. . .

Testing on fold 0 ━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━ 53/53 • 0:03:12 Testing on fold 1 ━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━ 52/52 • 0:02:50 Testing on fold 2 ━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━ 52/52 • 0:03:03 Testing on fold 3 ━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━ 52/52 • 0:02:48 Testing on fold 4 ━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━ 52/52 • 0:02:55 Evaluating predictions ╸━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━ 5/272 • 0:00:20 rank0: Traceback (most recent call last): rank0: File "/opt/apps/miniforge/mistOct2024/bin/mist_train", line 8, in

rank0: File "/rsrch1/ip/dtfuentes/github/MIST/mist/scripts/train_entrypoint.py", line 28, in train_entry

rank0: File "/rsrch1/ip/dtfuentes/github/MIST/mist/main.py", line 68, in main

rank0: File "/rsrch1/ip/dtfuentes/github/MIST/mist/evaluate_preds/evaluate.py", line 294, in evaluate rank0: path_to_truth = paths.loc[

rank0: File "/opt/apps/miniforge/mistOct2024/lib/python3.12/site-packages/pandas/core/indexing.py", line 1191, in getitem rank0: return self._getitem_axis(maybe_callable, axis=axis)

rank0: File "/opt/apps/miniforge/mistOct2024/lib/python3.12/site-packages/pandas/core/indexing.py", line 1752, in _getitem_axis rank0: self._validate_integer(key, axis) rank0: File "/opt/apps/miniforge/mistOct2024/lib/python3.12/site-packages/pandas/core/indexing.py", line 1685, in _validate_integer rank0: raise IndexError("single positional indexer is out-of-bounds") rank0: IndexError: single positional indexer is out-of-bounds

[aecelaya]

That process happens entirely on the CPU. I did push some bug fixes for something like this, but it hasn't made it into a version yet. What happens if you run evaluation on this same dataset, but when you install from the latest repo?

[aecelaya]

I'm running this now with the latest code I just pushed. I'll post a screen shot of the results.

[aecelaya]

Looks like the latest code works with all metrics:

I'll release a v0.0.4b0 tonight with these updates.

[aecelaya]

The latest version (0.1.0-beta) addresses this issue. Please let me know if it comes up again.