Note - This workshop was recently updated to use the GA instance of watsonx.ai as well as the new Foundation Models module within the Watson Machine Learning Python SDK. The previous version was based on the "workbench" BAM environment, which will soon be sunset.
Completing these technical hands-on labs will take roughly 12 hours plus another 2-3 hours to complete the Apply Lessons Learned labs. You should have completed the attendee pre-requisites prior to starting.
| Time | Topic |
|---|---|
| 20 mins | Technical POV on watsonx.ai Platform |
| 10 mins | IBM's Foundation Models |
| Time | Topic |
|---|---|
| 30 mins | Lab 0: Setup Your Laptop Environment |
| 10 mins | Intro to Prompt Engineering |
| 90 mins | Lab 1: Intro Prompt Engineering Using watsonx.ai Prompt Builder |
| 90 mins | Lab 2: Advanced Prompt Engineering Challenge |
| 60 mins | Lab 3: Langchain Prompt Templates |
| 60 mins | Lab 5: watsonx.ai and Langchain |
| 60 mins | Lab 6: Retrieval Augmented Generation (RAG) for Contextual Search |
| 30 mins | Lab 7: watsonx.ai Demo in Streamlit |
| Time | Topic |
|---|---|
| 5 mins | Intro to Watson Code Assistant |
| 45 mins | Lab 1: Watson Code Assistant and Ansible Lightspeed |
You've made it to the end. Almost. You will now apply the new skills that you've learned to the challenges below. Select one of the challenges below and be creative. These challenges are specifically open-ended. Combine your own unique skills to extend the use case solutions provided in these challenges.
| Time | Topic |
|---|---|
| 1 hour | Challenge 1: Text Classification of News Articles |
| 2 hours | Challenge 2: Multi-Turn Model Interactions |
| 2 hours | Challenge 3: Visualize Model Conversation With Streamlit |
| 2+ hours | Challenge 4: RAG Search of SEC 10K Filings With Streamlit |
Did you complete a Challenge and want to share your results. Send your Github repo link to apsteven@us.ibm.com and we'll add it to this list.
| Topic | Author |
|---|---|
| Super Fast RAG With File Upload | Tyler Benson |