Instructions

[princyi]

Project Instructions
Step 1: Upload Project Starter Files
Start your SageMaker Notebook Instance by clicking on Open JupyterLab once your notebook instance is ready.

Upload the 2 Python notebook files (.ipynb) in the project starter files(opens in a new tab) folder to JupyterLab using the upload arrow in the top left menu.

Remember to stop your notebook instance when you stop or pause work on your project.
Make sure to save your individual notebook files periodically in JupyterLab using the File -> Save menu, otherwise your work will be lost between your classroom Cloud Gateway sessions.
If you don't see your notebook instance when logging back in, make sure your region is set to us-west-2

Step 2: Choose your Dataset
In the project notebook files, you'll have to make a choice of what type of domain expert you'd like your model to be. Your choices are:

Financial domain
IT domain
Healthcare/Medical domain
Fill out the domain you've selected in the Project Documentation Report

Step 3: Deploy and Evaluate the model
Complete and run the cells in the Model_Evaluation.ipynb file
Take a screenshot of the Model_Evaluation.ipynb file with the cell output as proof you completed this step of the project
Save and download your Model_Evaluation.ipynb with the cell output, you'll be submitting this file.
Double check you've ran the cells that delete the model deployment and endpoint. Verify your model and endpoint have been deleted. If you do not delete these, you will run out of budget and will be unable to complete the next part of the project.
Fill out the Project Documentation section about your evaluation of the model's text generation capabilities and knowledge.

Step 4: Fine-tune the Model
Complete and run the cells in the Model_FineTuning.ipynb file
Download your Model_FineTuning.ipynb with the cell output, you'll be submitting this file.
Take a screenshot of the Model_FineTuning.ipynb file with the cell output as proof you completed this step of the project
Fill out the Project Documentation Report section about fine-tuning the model.

Step 5: Deploy and Evaluate the Fine-tuned Model
Visit the AWS S3 bucket where your fine-tuned model weights are stored after training and take a screenshot for your submission.
Complete and run the cells in the Model_FineTuning.ipynb file about deploying and evaluating the fine-tuned model
Take a screenshot of the Model_FineTuning.ipynb file with the cell output as proof you completed this step of the project
Double check you've ran the cells that delete the model deployment and endpoint. Verify your model and endpoint have been deleted. If you do not delete these, you will run out of budget and will be unable to complete the project.
Fill out the Project Documentation Report section about your evaluation of the fine-tuned model's text generation capabilities and knowledge.

Step 6: Collect Project Documentation and Submit
Revisit the project assessment rubric(opens in a new tab) and verify you've completed all of the project requirements and have collected all of the necessary files, proof, and documentation you need to submit the project
Model_evaluation.ipynb with cell output
Model_FineTuning.ipynb will cell output
Screenshots of both notebooks with cell output
The completed Project Documentation Report