-
Notifications
You must be signed in to change notification settings - Fork 45
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
example run issue. #20
Comments
Dear Friend, Which of the examples are you trying to run? I suspect that you might be running out of memory. Can you try reducing the size of the model to see if that helps? |
Friend and Brother. !! Memory already available is 16 GB. 24 cores machine. at the moment 12 GB free memory is still availeble. https://pypi.org/project/deepwave/ The starting example given here I am trying to get familiar with this library and modeling. If you can guide how much minimum memory should I have. or if I am making some nasty mistake? Greetings. |
I have just tried executing the example on a CPU and it ran successfully: This CPU only had 12GB of RAM, so 16GB should certainly be sufficient for this small example. Do the steps look the same as the ones that you followed? |
If it is still not working, one possible source of problems is the large number of cores you have. I have not tested it on a machine with that many cores. The CPU version of the code uses OpenMP to parallelize over cores within a node. To check if this is causing a problem, you could try setting the environmental variable OMP_NUM_THREADS to 1. |
Dear Brother, I found your tool interested for seismic inversion learning. Let me see .. !! I will update you shortly. Still I am working on it. |
Did you manage to get it running? |
HI. Greetings. |
Hi, Thank you for your continued efforts to use the code, and I am very sorry that you are experiencing difficulty. I have not managed to reproduce the kernel dying issue on any of the systems that I have tested, so unfortunately I do not know what to suggest. Please let me know if you discover any new information that might enable me to help you. |
HI. Thank you so much .. no issue. no worries. I will fix it soon and will share my experience. I would like to ask that are there any plans for the extension in 3D? Greetings. Khawar |
Hi Khawar, Thank you for offering to share your experiences. I am sure they will be valuable for others. The code should already run in 3D, but unfortunately it will only work for small models as it will probably run out of memory for larger models. I am very busy with other projects at the moment, so cannot dedicate the time to improve this. If you or someone else would like to write the code to improve it, I may be able to help. |
Dear Great Friend. That is super perfect. I will have a look how to cope with memory issue and will share my experience with 3D and will make some effort to use it for the larger 3D models. Just one question. This BIN format for the loading is special format or its from netcdf? Greetings. |
The model file that I use in the example is a plain 32-bit floating point binary file. The code expects a PyTorch Tensor as input - in the example I simply load the binary file using NumPy and then convert it to a Tensor before passing it to Deepwave. |
so simply the standard segy has been converted to binary? by using numpy or segyio can be used for deep wave or any guide? |
Yes, as long as the input to Deepwave is a PyTorch Tensor, it does not matter where that Tensor came from. |
Dear Friends. When I run ' pip install deepwave' , I observe this errors : c:\users\omen\appdata\local\temp\pip-install-h9i1pg1n\deepwave\deepwave\scalar\scalar.h(56): error C2143: syntax error: missing ',' before 'const'
Command ""c:\program files\python37\python.exe" -u -c "import setuptools, tokenize;file='C:\Users\OMEN\AppData\Local\Temp\pip-install-h9i1pg1n\deepwave\setup.py';f=getattr(tokenize, 'open', open)(file);code=f.read().replace('\r\n', '\n');f.close();exec(compile(code, file, 'exec'))" install --record C:\Users\OMEN\AppData\Local\Temp\pip-record-ibq1rxt_\install-record.txt --single-version-externally-managed --compile" failed with error code 1 in C:\Users\OMEN\AppData\Local\Temp\pip-install-h9i1pg1n\deepwave\ PLease help me that install deepwave Greetings |
It unfortunately appears that the compiler you are using (which seems to be
Visual Studio) does not support the "restrict" keyword. I apologise for
using a non-standard keyword. As a workaround, you might try cloning the
repository, removing the "restrict" keyword, and then installing from the
local copy.
For example, I think this should work (but is untested):
1. git clone https://github.com/ar4/deepwave.git
2. edit files in deepwave/deepwave/scalar/ to remove the "restrict" keyword
3. In the deepwave directory (the first one, where there should be a file
called setup.py), run "pip install -e ." (the final "." is part of the
command)
I hope that this solves the problem.
… |
Dear Friends, I have modified the code to change -Alan |
Dear Friends.
I have just executed the example and the kernel is dying at this step.
receiver_amplitudes = prop(source_amplitudes, x_s, x_r, dt)
I am using python 3.6.8 and pytorch 1.0.0
can anyone guide what the issue can be? I am not using any CUDA. Only CPU based cluster.
Greetings.
The text was updated successfully, but these errors were encountered: