An optimized prompt tuning strategy comparable to fine-tuning across model scales and tasks.
Primary LanguagePythonApache License 2.0Apache-2.0