Spaces:

usm3d

/

S23DR

like 15

Running on CPU Upgrade

Why is the space paused after my submission?

@Siromanec we are updating all previous leaderboard now with a fixed a re-evaluated metrics, you just got unlucky.
The space is expected to go back online in ~1 hour.

Hello, could you please mail me the logs on runs 04d5854d-434a-4cc7-9871-b0971d6fc514, 1ffda7b8-858d-4418-8208-d1733a416fe8, 65758f07-8643-46ce-9a0d-8da12934b4ab? My email: [email protected]

@Siromanec of course, not. The logs reveal too much information and could potentially leak the info about the test set.

However, I can post here some error traces, which hopefully help with debugging.

It seems that there is something wrong with packages on your 04d5854d submission

2024-05-26 20:00:54.203 | INFO     | __main__:generate_submission_file:41 - Generating submission file
Traceback (most recent call last):
  File "/tmp/model/script.py", line 34, in <module>
    import hoho; hoho.setup() # YOU MUST CALL hoho.setup() BEFORE ANYTHING ELSE
ModuleNotFoundError: No module named 'hoho'

1ffda7 finished succesfully.

65758f07 - same as 04d5854d submission

hoho module now is not preinstalled? It is extremely odd that this error did not occur three times in a row. Especially when considering the fact that I did not tweak any of the imports, when comparing to the baseline (or between the three runs).
And now the run identical to 65758f07 is a success?

@Siromanec the hoho module is preinstalled. We do not have a control over the huggingface infrastructure, so some weird stuff is possible. It is also possible, that the 65758f07 has been submitted at the times of the Space restart or something like that. We apologize for possible inconveniences, but, as I said, we do not have control over them.

Any information on 396768ac-d2af-4e66-bf80-c2f4860f0953?
Thanks in advance

@colin1842

32607fee-ac2f-4d08-8e03-f9308f8d6353

test: 0%| | 0/1 [00:00<?, ?it/s]CUDA kernel failed : no kernel image is available for execution on the device

87a66b43-aadb-4baf-8ac4-4552a634a12c

  0%|          | 0/1 [00:00<?, ?it/s]
Processing entries: 1it [00:02,  2.71s/it]
2024-06-10 10:05:33.726 | INFO     | __main__:generate_submission_file:74 - Subprocess terminated successfully
2024-06-10 10:05:33.726 | INFO     | __main__:generate_submission_file:76 - contents of submission_dir
2024-06-10 10:05:33.727 | INFO     | __main__:generate_submission_file:77 - ['params.json', '.gitattributes', 'LICENSE', 'README.md', '_script.py', 'checkpoint_epoch_90.pth', 'dataset', 'format_dataset.py', 'handcrafted_solution.py', 'hoho_train_checkpoint_epoch_90.pth', 'model', 'model_cfg.yaml', 'output', 'pc_util.cpython-310-x86_64-linux-gnu.so', 'pc_util.py', 'pc_util', 'pointnet2.py', 'script.py', 'test.py', 'test_util.py', 'train.py', 'train_utils.py', 'utils', 'wandb', 'socket-kit.so', 'data', '__pycache__']

And then it didn't save to submission.parquet

38311405-94bb-44e2-bd59-5a0a6397aa02
could u please check this?

135f90c7-e620-4de3-b6d1-a38a8ca00983?

@colin1842

38311405-94bb-44e2-bd59-5a0a6397aa02

you are trying to download Downloading: "http://__________.cz/_____-/models/_______.ckpt" to /app/.cache/torch/hub/checkpoints/____.ckpt

@kc92

135f90c7-e620-4de3-b6d1-a38a8ca00983?

Traceback (most recent call last):
  File "/tmp/model/script.py", line 142, in <module>
    pred_vertices, pred_edges = my_submission(house_data, local = local)        
  File "/tmp/model/script.py", line 91, in my_submission
    house_data.triangulate_all_2d_corner_pairs_new()
  File "/tmp/model/solution.py", line 660, in triangulate_all_2d_corner_pairs_new
    triangulated_corners, vertex_types, image_vertex_inds = triangulate_from_viewpoints(Ks = self.Ks, 
  File "/tmp/model/utils.py", line 148, in triangulate_from_viewpoints
    assert len(Ks) == len(Rs) == len(ts) == len(vertices)
AssertionError