multi-round editing

What is multi-round editing?

Iterative process where a model performs successive modifications to an output based on ongoing user instructions

In other languages

한국어다회차 편집

지시 사항을 한 번에 처리하지 않고 여러 번의 상호작용을 통해 점진적으로 음성을 수정하는 작업 방식

Related Terms

• Continual supervised fine-tuningTraining process where a model is repeatedly updated on new labeled datasets over time
• Test-time trainingA technique that continues to adapt or fine-tune a model during inference so it can fit the current test data better.
• Self-conditioningTechnique where a model uses its earlier prediction as additional input for later refinement
• Continual pre-trainingMachine learning process of further training an existing model on new data to acquire additional knowledge while retaining previously learned information
• Soft Prompt TuningA method of adapting a model to specific tasks by learning small, adjustable vectors (prompts) without changing the core weights of the model.
• Instruction alignmentThe process of fine-tuning a model so that its responses accurately follow the specific constraints, tone, and formatting requested by a user.