Draft of image patch extraction pipeline - for feedback #167
Conversation
of the other notebooks.
and training data
up some old files. Added start of switch CNN training.
…ta. Moved some functions to scripts for neater code. Confirmed function with CNN training test.
… base image feature extraction everytime.
tim-rehnstrom
changed the title
| @@ -1,2 +1,3 @@ | |||
| FROM ghcr.io/nasa/isaac_analyst_notebook | |||
| ENV DEBIAN_FRONTEND=dialog | |||
| RUN pip install torch==2.2.1 torchvision==0.17.1 torchaudio==2.2.1 --index-url https://download.pytorch.org/whl/cu121 | |||
There was a problem hiding this comment.
Do we need to commit to these exact versions? What is this url?
There was a problem hiding this comment.
the link is from the pytorch install instructions here: https://pytorch.org/
There was a problem hiding this comment.
the exact versions might be remaining from when we had to support multiple ubuntu versions
| "mounts": [ | ||
| "source=${localWorkspaceFolder},target=/src/isaac/src,type=bind,consistency=cached", | ||
| "source=${localWorkspaceFolder}/../../astrobee/src,target=/src/astrobee/src,type=bind,consistency=cached" | ||
| "source=${localWorkspaceFolder}/../../astrobee/src,target=/src/astrobee/src,type=bind,consistency=cached", | ||
| "source=${localWorkspaceFolder}/../../data/Sock_example,target=/src/data/vent,type=bind,consistency=cached" |
There was a problem hiding this comment.
How does this directory get here? Will this cause problems for people who don't have your exact setup or use the analyst notebook for other purposes?
There was a problem hiding this comment.
That is left from testing I had done in the past. I will remove it, so no issues occur for others using the notebook
There was a problem hiding this comment.
Yes it should not be included. It is an old file that slipped through. Files that are deleted within jupyterlab get put in the .trash-0 folder.
| "source": [ | ||
| "from scripts.load_bag_database import LoadBagDatabase\n", | ||
| "path=\"data/bags/\"\n", | ||
| "path=\"data/bags/2024-03-21_tim/bsharp/Fixed\"\n", |
There was a problem hiding this comment.
not sure you want to commit the changes to this file. They seem specific to you.
There was a problem hiding this comment.
They are specific to me. I will adjust it for the final version.
There was a problem hiding this comment.
How do other people get the data to run this?
There was a problem hiding this comment.
The bagfiles needs to be uploaded to the ArangoDB first before this can be used to extract the targets. The upload can be done with the 1_import_bagfiles.ipynb notebook.
There was a problem hiding this comment.
Make this a function query_image.py? Probably doesn't need a separate module?
There was a problem hiding this comment.
Probably better not to have _scripts in filename
There was a problem hiding this comment.
That makes sense, an oversight on my part. I will adjust the filenames and the notebooks
There was a problem hiding this comment.
What is this? Add some explanation
There was a problem hiding this comment.
This is one of the subfeatures. It allows the user to select a target in the image and get the coordinates. It is integrated into the gather_training_data notebook for the actual pipeline. I will add some more explanation and it can act as a tutorial as we talked about during Tuesday's meeting.
There was a problem hiding this comment.
Also a subfeature. Will add comment so it can act as a tutorial.
|
I recommend that you work from a branch with a relevant name to your feature and not develop. This makes it easier for us to checkout your branch and fix something that comes up too. To create a branch you can do: Also, keep your branch updated with our develop, since we push fixes and updates frequently. Right now, your tests in this PR are failing because you haven't merged in the latest opencv updates |
This is a draft containing the patch extraction pipeline. Please let me know if you find any errors or if you see something where I can optimize the code.