Fine-Tuning Foundation Models in Amazon Bedrock
Short answer
A Provide labeled data with the prompt field and the completion field.
Question
Which strategy will successfully fine-tune the model?
Key Takeaways
The main trap is confusing the fine-tuning data preparation step with the post-training infrastructure setup (Provisioned Throughput).
To fine-tune a foundation model in Amazon Bedrock, you must provide labeled training data consisting of prompts and completions. This adapts the model to your specific domain.
Common mistake: Choosing Option C (Purchase Provisioned Throughput) because users associate custom models with reserved capacity, but provisioned throughput is for inference, not training.
Expert Analysis
The question asks for the strategy to fine-tune a foundation model in Amazon Bedrock. Fine-tuning is the process of training a pre-trained model on specific, labeled data to adapt it to a particular task or domain.
- Option A describes providing labeled data with prompt and completion fields. This is the correct method for fine-tuning in Bedrock, where input-output pairs are used to teach the model specific behaviors.
- Option C (Provisioned Throughput) is a common distractor. While provisioned throughput is often required after fine-tuning to ensure capacity for inference, it is not the mechanism for the fine-tuning process itself. As noted in comments, "Fine-tuning and provisioned throughput are two separate processes... Provisioned throughput is required after fine-tuning." Thus, purchasing it does not perform the fine-tuning.
- Option B suggests an incorrect file format (.txt containing.csv), which is technically invalid for structured training datasets.
- Option D suggests using external journals/textbooks, which ignores the requirement to use the company's own data for domain-specific adaptation.
The key insight is distinguishing between the training data structure (Option A) and the inference capacity configuration (Option C).
Exam Strategy
When asked about 'fine-tuning' or 'training', always look for options describing data preparation (labeled inputs/outputs). Reserve options about 'throughput', 'capacity', or 'deployment' for questions about inference scaling or post-training steps.
Frequently Asked Questions
Is Provisioned Throughput needed for fine-tuning?
What format should fine-tuning data be in?
Community Votes
100% of anonymous learners picked answer A. Votes are pick records left by other test-takers — they are not the verified answer.
Discussion 9 contributions
Compiled from public community threads and reader submissions, lightly edited for clarity. Handles and vote tallies are shown as recorded, and comments reflect the views of individual contributors.
From public community threads 9
-
Dandelion2025 👍 8Fine-tuning and provisioned throughput are two separate processes in Amazon Bedrock: Fine-tuning is the process of training the model on specific labeled data to improve its accuracy for particular tasks. This can be done without purchasing provisioned throughput.Provisioned throughput is required after fine-tuning, specifically for using the custom model for inference. It's not needed for the fine-tuning process itself.To test and deploy your model, you need to purchase Provisioned Throughput.
-
ap6491 👍 8 Answered AFine-tuning a foundation model involves providing labeled training data where each example consists of a prompt (the input to the model) and a completion (the desired output). This structure helps the model learn specific patterns or behaviors tailored to the company’s data and use case. In Amazon Bedrock, fine-tuning relies on a structured dataset that aligns with the model's learning requirements to improve its accuracy for domain-specific tasks.
-
Jessiii 👍 4 Answered AProvide labeled data with the prompt field and the completion field: Fine-tuning a foundation model (FM) typically involves providing labeled data in the form of prompts (input) and completions (desired output). The model learns from this data to generate more accurate responses tailored to the company’s specific use case. This approach is aligned with how fine-tuning works, where the model adjusts its parameters based on the relationship between inputs (prompts) and outputs (completions).
-
aldricstormcloak 👍 3Why is it not C? Finetuning cannot be done without provisioned throughput mode active.
-
jove 👍 3 Answered ALabeled Data: Fine-tuning requires labeled data
Show 4 more comments
-
VIJAY K T 👍 1 Answered ATo fine-tune a foundation model in Amazon Bedrock, you provide training examples consisting of prompts and their desired completions (responses). The model learns from these labeled input-output pairs and adapts to the company's domain and data.
-
Aamee 👍 1Correct. For fine-tuning the customized model. Here are the steps: 1. Fine-tune/train the custom model. 2. Purchase provisioned throughput for capacity. 3. Deploy the custom model.
-
Udyan 👍 1Fine Tuning is Done with Labeled Data so, A
-
65703c1 👍 1 Answered AA is the correct answer.
Reader comments & corrections
No reader corrections yet — spotted an error or have a note? Share it below.
Related Analysis
Practice All AIF-C01 Questions
Access 100 questions with complete answers and detailed explanations.
View Full AIF-C01 Practice Test →