Skip to content

Is the Ray framework essential for Cell type identification on a single node with 2 GPUs? #14

@ddzhang3

Description

@ddzhang3

Thanks for your work on decomposition of spatial transcriptomic data. However, I am now stucked in the ray.init(), which always throws out of errors:

2024-04-15 08:07:58,016 ERROR services.py:1363 -- Failed to start the dashboard: Failed to start the dashboard
Failed to read dashboard log: [Errno 2] No such file or directory: '/tmp/ray/session_2024-04-15_08-06-57_711318_1989227/logs/dashboard.log'
2024-04-15 08:07:58,017 ERROR services.py:1364 -- Failed to start the dashboard
Failed to read dashboard log: [Errno 2] No such file or directory: '/tmp/ray/session_2024-04-15_08-06-57_711318_1989227/logs/dashboard.log'
Traceback (most recent call last):
File "/home/users/ddzhang/miniconda3/envs/SpatialScope/lib/python3.9/site-packages/ray/_private/services.py", line 1336, in start_dashboard
with open(dashboard_log, "rb") as f:
FileNotFoundError: [Errno 2] No such file or directory: '/tmp/ray/session_2024-04-15_08-06-57_711318_1989227/logs/dashboard.log'
During handling of the above exception, another exception occurred:
Traceback (most recent call last):
File "/home/users/ddzhang/miniconda3/envs/SpatialScope/lib/python3.9/site-packages/ray/_private/services.py", line 1347, in start_dashboard
raise Exception(err_msg + f"\nFailed to read dashboard log: {e}")
Exception: Failed to start the dashboard
Failed to read dashboard log: [Errno 2] No such file or directory: '/tmp/ray/session_2024-04-15_08-06-57_711318_1989227/logs/dashboard.log'

I tried several versions:
ray 2.0.0 with grpcio 1.43.0 and ray 2.0.0 with grpcio 1.46.3, and changed the temp_dir in the ray.init() of src/Cell_Type_Identification.py with no success. I also tried ray.init() with local_mode=True, which gave the same error.

What I wander is that, for my cluster's config with 1 head node and 1 compute node with 2 GPUs, is it possible to get rid of the ray framework and just run on a local mode?
I read a part of the source code, and it seems that I could modify the code to not use the Ray framework. What I'm not sure is that would the remove of ray influence the results of data analysis?

Metadata

Metadata

Assignees

No one assigned

    Labels

    No labels
    No labels

    Type

    No type

    Projects

    No projects

    Milestone

    No milestone

    Relationships

    None yet

    Development

    No branches or pull requests

    Issue actions