Objectives

The participant:

• Knows the BMW GenAI tools and knows how to use them effectively.
• Knows which language models can be used for which use cases and what costs are associated with them.

Content

This training covers the basics of prompt engineering for the BMW GENEXT and IMAGINE tools, including:

• GENEXT:
  • Which GenAI tools can we use at BMW?
  • What is a prompt and how does it provide context and instructions for large language models?
  • Strategies for formulating an effective prompt, such as reformulating issues, adding relevant context, specifying the audience, and output format.
  • Rules for good prompting.
  • Advanced techniques such as chain of thought prompting.
  • Customizing models (e.g. OpenAI GPT 4 Turbo, Anthropic Claude v2.1, etc.) and model parameters such as Temperature or Respond Length.
• IMAGINE:
  • How are images generated with IMAGINE and the open source model Stable Diffusion XL?
  • What do good prompts look like for IMAGINE?
  • Prompt fine-tuning and negative prompting.
  • Parameter engineering, such as Sampler, Seed, N_Samples, N_Iter.
  • Where and in what form can I use the generated content?

Target Group

• Anyone who wants to learn the basics of prompting with the BMW GenAI tools Genext and Imagine.