search

jcorsetti / oryon

Official implementation of CVPR24 Highlight paper "Open-vocabulary object 6D pose estimation"
20 stars 2 forks source link

Missed utils folders #2

Closed taeyeopl closed 4 months ago

taeyeopl commented 5 months ago

Hi @jcorsetti, Thanks for sharing the codes. It looks like you have missed uploading the utils folders.

jcorsetti commented 5 months ago

Hi Taeyeop, indeed you are right, I just added them along with some missing scripts. I will keep the issue open for a few days in case other files are missing.

taeyeopl commented 5 months ago

Thanks for the quick reply with updating.

taeyeopl commented 5 months ago

Do you have any plans for releasing the checkpoints to reproduce the report results? (e.g. exp_data/baseline/ files)

jcorsetti commented 5 months ago

The best checkpoint has been released along with the main code, is the oryon_best.zip file in the latest release. It can be evaluated as detailed in the example in the README

taeyeopl commented 5 months ago

python run_train.py exp_name=baseline Run the following to obtain results with the basic four configurations. The last checkpoint is used by default. python run_test.py -cp exp_data/baseline/ dataset.test.name=nocs test.mask=predicted

As far as I understand, the README suggests using the last checkpoint after training. what does it mean release the best checkpoint? Does the best checkpoint denotes pretrained_models.zip folders?

taeyeopl commented 5 months ago

https://github.com/jcorsetti/oryon/releases/download/v1.0.0-data/oryon_best.zip I understood what you said. Thanks :)

jcorsetti commented 5 months ago

Sorry for the late reply. The pretrained_models folder is used to store the checkpoints of third-party models we need, such as PointDSC. The best checkpoint is the one released (oryon_best.zip), in order to evaluate it it should be unzipped in exp_data, and the run_test.py script should be run, using -cp exp_data/oryon_best as argument.

Let me know if something is unclear

ismael-esp commented 4 months ago

hi, @jcorsetti. first of all, thanks for sharing your code.

I encountered an error when trying to execute the run test Python file for either real275 or toyl. The error is the following.

FileNotFoundError: [Errno 2] No such file or directory: 'exp_data/oryon_final/results/toyl_cross_scene_test_all_19042024_2039_431.csv'

As highlighted before, I downloaded the oryon_best.zip file and extracted it into the exp_data folder. However, it appears that it doesn't include a results folder. Should I run another Python code to generate them, or am I doing something wrong? Thanks in advance for your help.

jcorsetti commented 4 months ago

Hi @ismael-esp, thanks for your interest. Yes, the code expects a "result" folder inside the oryon_final folder, as all the result files for the evaluation will be placed there. For now you can create the results folder manually in exp_data/oryon_final and it should work. I will update the code so that it automatically creates the folder when needed. Let me know if this solution does not work.