Closed aThinkingNeal closed 7 months ago
from your screenshot, the dataset# is 0. This looks suspicious, can you confirm the data dir is given correctly?
Thanks for your help!
Usually what should the #dataset look like?
If data dir means the --train_data_path parameter:
I have run the command "ls" on the --train_data_path parameter path, the result looks fine to me. So I guess the data dir is given correctly?
My process of getting this data is as follows:
download the dataset of bleach_cleanser from this website: https://archive.cs.rutgers.edu/archive/a/2020/pracsys/Bowen/iros2020/YCB_traindata/
extract it under the YCBV_data folder
The following point may or may not be relevant to this issue:
in the repo there is the data_organized dataset, but I have no idea on how to use it.
https://archive.cs.rutgers.edu/archive/a/2020/pracsys/Bowen/iros2020/YCB_traindata/
this is the training data. If you want to run inference, you need to download the data_organized
which is an organized version of YCB-Video dataset. If you have configured those correctly, expect to see #dataset
>0
this is my current dataset configuration, I have downloaded the data_organized
and put them under the ycb_dir
folder, which in my case is /home/neal/debug/YCBV_data
.
(By the way, Could you share your dataset folder structure? It seems a bit unclear from the README session, I am afraid that others may face the same issue when setting up their dataset configuration.)
However, I am still getting the error: IndexError: list index out of range
I have tried to debug by adding a print statement in the predict.py:
it seems that the program didn't manage to go to the test_data_dir
step, otherwise the print should show up in the error message.
Below is my command to run the program, note that I have updated the ycb_dir
.
can you print args.ycb_dir
here?
https://github.com/wenbowen123/iros20-6d-pose-tracking/blob/18dc5bac30024e4d44465f3202ba6ffbb62e8e3b/predict.py#L315
If it's not updated like what you passed in the command line, maybe it's argparse's issue. Then try to directly edit it here
Thanks for the help, I have managed to solve the IndexError
problem.
Essentially it is the test_data_path
is not correctly configured.
Here is what I changed to make it work:
change the line of code from here
to
test_data_path = '{}/data_organized/%04d'.format(args.ycb_dir)%(seq_id)
organized my dataset configuration as:
/home/neal/debug/YCBV_data/data_organized/0051
In this way the test dataset is correctly passed and the problem is solved
I think the confusion comes from the README and codebase mismatch:
Here in the README session implies that should be a data_organized
folder
However in the codebase this folder path is not reflected.
Hope this helps!
Hi I am running the docker image but got this IndexError, could you take a look at it?
Thank in advance for your help!
the content of test.sh is: