good first issue
Repository metrics
- Stars
- (21,634 stars)
- PR merge metrics
- (No merged PRs in 30d)
Description
Spent 2 hours training a model for it to fail the save with error prefixed with OSError: [Errno 22] Invalid argument :(
For now I'm just adding fail-safe but I am sure it can be written cleaner
custom_save_path = save_model_path
while True:
try:
torch.save(transformer.state_dict(), custom_save_path)
break
except:
print("Please enter new save path:")
custom_save_path = input()
print("\nDone, trained model saved at", custom_save_path)