Shadow removal - weights argument not defined in tensorflow/keras model call
#54
Labels
0_Code
for Zenhub Workspace Filtering
bug
Something isn't working
end user issue - general bug
Issues identified by end users
Comments
CameronBodine
added
bug
CameronBodine
added a commit
that referenced
this issue
Jan 11, 2023
|
I'm not able to recreate on my end, but I think I fixed the issue. I updated |
|
That fixed changed the resulting error a little bit, but it still ends with the same problem with the string |
|
Can you confirm that you have a model folder in the ping mapper directory and that there are hdf and json files in there? |
|
I have folders bankpick, bedpick and shadow in the folder PINGMapper/models, inside PINGMapper/models/shadow I have 1 .json file and 1.h5 file |
|
With |
|
|
Alrighty, can you post the contents of |
|
For example, on windows: |
|
Here is something to try (since you figured out how to install the environment way faster): Remove current ping mapper env: Clean up: Update base conda environment: Then try reinstalling: |
|
|
I'll give those steps a shot. |
|
I ran those steps you suggested and unfortunately got exactly the same error ending in |
|
PS here's every exception that comes out: |
|
I was able to successfully recreate your error! TLDR:Running Don't do this part! Just for documentationI changed the PINGMapper.yml to be: I then install the new environment: This gives me the following package version, close (I think the exact same) to your env: Then, when I run the test script The FixTo fix this, I force reinstalled This gave me the following env (note the updated tensorflow-gpu): Everything ran as expected after re-running the test script!!! NoteYou may want to consider updating your python version as well (python=3.9) to try and get your environment as close to mine as possible. I'm still stumped why PING-Mapper didn't install to the same versions as mine. Perhaps you installed an older version of miniconda3? Dunno. But hopefully updating tensorflow will work for you and you can keep on mapping! Please let me know how it goes. |
|
Hurray thanks Cam! I'm going to try this now. That is weird, I got the latest miniconda3 version on the site, and it says my installed conda version is 22.11.1. I tried to run |
|
In my experience, uninstalling Anaconda/miniconda doesn't always remove everything. So, in your case, I bet there were old files that were causing you to revert to older package versions. When you told me about I think the Welcome to the world of open-source computing! You truly get what you pay for ;-) Let me know once you successfully run the test script, and we can close this issue as a success! 🤜🪵🤛 |
|
Thanks for all your help Cam, I just got loaded with a ton of multibeam work but I'll get back to you on this soon! |
CameronBodine
added a commit
that referenced
this issue
Jan 24, 2023
Use weights stored in object #54
|
Hi @bthetford, I was wondering if you had a chance to check the new changes? If they work for you, I would like to close this issue. Thanks. |
|
Hi Cameron, I did not get it working but haven't had any time to test it! I can try to look at it again maybe Friday. |
|
I got back to this yesterday and tried to force reinstall tensorflow GPU, and as you said my python version 3.10 is an issue. Going to use 3.9 and retry. |
|
Hi @bthetford, I have updated |
|
Thanks I'll try now! I haven't been able to get the environment to create lately, I just got that same tensorflow error and my env list remained empty. I'll let you know how the new YAML does. |
|
Hurray! It's working now. Tried both the test file and |
|
Good deal! I will go ahead and close this issue. If you have any other problems, feel free to open a new issue. |
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
When I try to process a recording (my own, or the example data using test_PINGMapper.py), using
remShadow = 1or 2, I get the following error:The error is identical whether I try to remove all or only bank shadows.
The text was updated successfully, but these errors were encountered: