crild / facies_net

Software for detection of given facies

facies_net

Software for detection of given facies

Functions needed to run facies_net.py are contained in the folder facies_net_func, logs hold TensorBoard data, and F3 holds trained models.

The user needs the F3 dataset in the same folder as facies_net.py, as well as classification in .pts files.

The file will save training results in TensorBoard-format, simply go into the terminal and write: tensorboard --logdir=logs/"name-of-folder" and then open a new web browser and write "localhost:6006" in the adress bar.

e.g. to view the results from F3_train write: tensorboard --logdir=logs/F3 then open a new tab in chrome and input localhost:6006

Optimized for TensorFlow 1.5.0

facies_net's Issues

How to generate F3_entire.segy file

Hi,

I have downloaded the F3 offshore demo from this link: F3 demo .
This opens up peacefully in OpenDtect.
However, the code provided here expects the data to be converted into one large segy file F3_entire.segy.

When I try to export the data cube using Survey -> Export -> Seismics -> Seg-Y -> 3D. It outputs a .sgy file that's 1.2gb in size. Not sure if this is all the data.
When I modify the given code to read this newly generated segy file, it gives me the following error:
ValueError: Invalid dimensions, ilines (631) * xlines (951) * offsets (1) should match the number of traces (600515)

Guys, any hints on how to obtain the full .segy file?
Any references, @crild ?
Thanks for making all this publicly available.

License

Hej Charles,

Thanks for tidying up and clearing out your MalenoV development. Would you mind specifying a license on this work?

I suspect as a successor to MalenoV it might be LGPL again?

StopIteration: index 4 is out of bounds for axis 1 with size 2

Is there any reason why I got the error? I think the error is because of keras updates but Im not sure how to fix it..

Request for help

Respeted sir/madam,
Kindly help me how to implement this facies prediction for the segy data.

Regards,
Satish