Request for Assistance in Running Training Code

[MzHameed]

Dear Mathias Gehrig,

I hope this message finds you well. I am truly fascinated by your work on object detection and have been eager to explore it further. Following your instructions, I set up an environment and created a Weights & Biases (WandB) account. However, I’m struggling to get the training code running and I’m unsure how to use the WandB account to run the training script.

I am using PyCharm as my development environment, where I have set the necessary training parameters. Unfortunately, I keep encountering errors related to Hydra configuration when attempting to execute the training code.

Here’s how I configured the parameters in PyCharm:
C:\Users\muhammadz\AppData\Local\anaconda3\envs\rvt123\python.exe C:\Users\muhammadz\PycharmProjects\pythonProject3\train.py
model=rnndet
dataset=gen1
dataset.path="C:/Users/muhammadz/PycharmProjects/pythonProject3/gen1"
wandb.project_name=RVT
wandb.group_name=gen1
+experiment/gen1="C:/Users/muhammadz/PycharmProjects/pythonProject3/config/experiment/base.yaml"
hardware.gpus=0
batch_size.train=8
batch_size.eval=8
hardware.num_workers.train=6
hardware.num_workers.eval=2

Despite this setup, I encounter the following error:
Using python-based detection evaluation
== Timing statistics ==
Traceback (most recent call last):
File "C:\Users\muhammadz\AppData\Local\anaconda3\envs\rvt123\lib\site-packages\hydra_internal\utils.py", line 220, in run_and_report
return func()
File "C:\Users\muhammadz\AppData\Local\anaconda3\envs\rvt123\lib\site-packages\hydra_internal\utils.py", line 356, in
lambda: Hydra.create_main_hydra2(
File "C:\Users\muhammadz\AppData\Local\anaconda3\envs\rvt123\lib\site-packages\hydra_internal\hydra.py", line 61, in create_main_hydra2
config_loader: ConfigLoader = ConfigLoaderImpl(
File "C:\Users\muhammadz\AppData\Local\anaconda3\envs\rvt123\lib\site-packages\hydra_internal\config_loader_impl.py", line 48, in init
self.repository = ConfigRepository(config_search_path=config_search_path)
File "C:\Users\muhammadz\AppData\Local\anaconda3\envs\rvt123\lib\site-packages\hydra_internal\config_repository.py", line 65, in init
self.initialize_sources(config_search_path)
File "C:\Users\muhammadz\AppData\Local\anaconda3\envs\rvt123\lib\site-packages\hydra_internal\config_repository.py", line 73, in initialize_sources
source_type = SourcesRegistry.instance().resolve(scheme)
File "C:\Users\muhammadz\AppData\Local\anaconda3\envs\rvt123\lib\site-packages\hydra_internal\sources_registry.py", line 29, in resolve
raise ValueError(
ValueError: No config source registered for schema C, supported types: [file, pkg, structured]

I would greatly appreciate any guidance or suggestions you could provide to help me resolve this error and successfully run the training code. Being able to reproduce the results from your work would be invaluable for my understanding of object detection and its applications.

Thank you so much for your time and the incredible work you’ve contributed to the field. I look forward to any advice you can offer.

[magehrig]

I can't help with Pycharm. Can you run the gen1 training command from the readme?