Hello, sorry for the bad english. I’m using for a personal project wandb to train Yolov5 in a kaggle environment. Im getting this type of error:
Traceback (most recent call last):
File “train.py”, line 643, in
main(opt)
File “train.py”, line 539, in main
train(opt.hyp, opt, device, callbacks)
File “train.py”, line 95, in train
loggers = Loggers(save_dir, weights, opt, hyp, LOGGER) # loggers instance
File “/content/yolov5/utils/loggers/ init .py”, line 73, in init
self.wandb = WandbLogger(self.opt, run_id)
File “/content/yolov5/utils/loggers/wandb/wandb_utils.py”, line 185, in init
allow_val_change=True)
File “/usr/local/lib/python3.7/dist-packages/wandb/sdk/wandb_config.py”, line 181, in update
sanitized = self._update(d, allow_val_change)
File “/usr/local/lib/python3.7/dist-packages/wandb/sdk/wandb_config.py”, line 175, in _update
parsed_dict, allow_val_change, ignore_keys=locked_keys
File “/usr/local/lib/python3.7/dist-packages/wandb/sdk/wandb_config.py”, line 227, in _sanitize_dict
self._raise_value_error_on_nested_artifact(config_dict)
File “/usr/local/lib/python3.7/dist-packages/wandb/sdk/wandb_config.py”, line 266, in _raise_value_error_on_nested_artifact
“Instances of wandb.Artifact and wandb.apis.public.Artifact”
ValueError: Instances of wandb.Artifact and wandb.apis.public.Artifact can only be top level keys in wandb.config
Here the debug log:
With a dowgrade to version 0.12.10 just work.