Please enable JavaScript.

Coggle requires JavaScript to display documents.

Nova - Coggle Diagram

- - - - Task
        
        Role (System Role)
        
        Response Style
        
        Prefill (direct the model's actions, bypass preambles, and enforce specific output formats such as JSON
        
        Instructions
        
        Chain of thought
        
        Few-shot prompting
        
        Provide supporting text (grounding content)
    - - Provide contextual and specific information
        
        Specifying the desired format and style of the output (e.g. "only generating response in JSON")
        
        Structured prompt with clear sections (e.g. Task, Content, Tone& Style)
      - Draft prompt is necessary to initiate the iterative process of prompt engineering
  - - - Character Count (<1024)
        
        Do not use negation words like “no”, “not”, “without”, use instead the negativeText parameter
        
        Refine Results:
        
        Using a consistent seed value, make small changes to your prompt or negative prompt and re-run the prompt
        
        Once the prompt has been refined to your liking, generate more variations using the same prompt but a different seed value
        
        Phrase the prompt as an image caption rather than a command
      - Prompt formula:
        
        Subject
        
        Environment
        
        (optional) Position/Pose of the subject
        
        (optional) Lighting description
        
        (optional) Camera position/framing
        
        (optional) The visual style or medium (“photo,” “illustration,” “painting,” etc.)
    - - Characters count (<512)
        
        Camera movement (aerial shot, arc shot, clockwise rotation, dolly in)
      - Prompting for video generation models differs from prompting for LLMs
        
        Video generation models do not have the ability to reason or interpret explicit commands
        
        It’s best to phrase your prompt as if it were an image caption or summary of the video rather than a command
        
        Place camera movement descriptions at the beginning or end of the prompt
- - - - Micro
        
        Use Cases:
        
        latency-sensitive
        
        high-volume text,
        
        including: classification, sentiment analysis, and search
    - - Lite
        
        Use Cases:
        
        multimodal use cases with balance required of speed, cost
        
        intelligence- in-line productivity tools
        
        document understanding
      - Pro
        
        Use Cases:
        
        Enterprise application use cases built on RAG-agentic applications
        
        Complex multimodal understanding
  - - - Use Cases:
        
        Image generation, content creation
    - - Use Case:
        
        Video generation, content creation