Common Warning Messages¶
Quick reference for common warning messages.
| Warning | Meaning | Action |
|---|---|---|
UserWarning: The dataloader does not have many workers |
Slow data loading | Increase num_workers |
UserWarning: Trying to infer the batch_size |
Can't determine batch size | Explicitly set in datamodule |
UserWarning: The number of training batches is very small |
Epoch finishes quickly | Increase dataset size or reduce batch size |
FutureWarning: Passing (type, 1) for ndim |
Deprecated numpy usage | Update to latest version |
UserWarning: Mixed precision is not supported on CPU |
Using wrong accelerator | Switch to GPU or remove precision flag |
Understanding Warnings¶
UserWarning¶
User warnings are informational and usually don't break training: - Data loader warnings: See Slow Training - Batch size warnings: Usually safe to ignore if training works - Mixed precision warnings: See Device Errors
FutureWarning¶
Future warnings indicate deprecated features: - Update dependencies to latest versions - Check AutoTimm documentation for new APIs
DeprecationWarning¶
Deprecation warnings suggest updating code: - Review recent changes in package documentation - Update to new recommended patterns
Suppressing Warnings¶
import warnings
# Suppress specific warning
warnings.filterwarnings("ignore", message=".*dataloader.*")
# Suppress all warnings (not recommended)
warnings.filterwarnings("ignore")
Related Pages¶
- Error Reference - Common errors
- Data Loading - Data loader warnings
- Device Errors - Hardware warnings