Fine-Tune Skepticism

Most problems attributed to fine-tuning may be solved simply by formatting and pointing a powerful pre-trained model at the right data, rather than costly retraining.