AI-102 Question 492
Select 3You are tasked with fine-tuning an Azure OpenAI model to improve its performance for a customer support chatbot that handles domain-specific queries. Which of the following steps are required to fine-tune the model successfully in Azure OpenAI?
- A
Prepare a dataset in JSONL format with training examples and validation examples.
- B
Configure a custom endpoint URL for your fine-tuned model during the fine-tuning process.
- C
Upload the training dataset to an Azure Blob Storage container accessible to the Azure OpenAI resource.
- D
Use the Azure OpenAI Studio or Azure CLI to initiate the fine-tuning process with the prepared dataset.
- E
Manually adjust the learning rate and epochs during the fine-tuning process.
Show answer and explanation
Correct answers: A, C, D
Explanation
Fine-tuning an Azure OpenAI model involves preparing a dataset in JSONL format, uploading the dataset to an accessible storage location, and initiating the fine-tuning process using tools like Azure OpenAI Studio or Azure CLI. Azure OpenAI handles the hyperparameters for fine-tuning, such as learning rate and epochs, automatically, and the endpoint URL for the fine-tuned model is generated by Azure.
- A. Correct.
Correct. The dataset must be in JSONL format, and it should include labeled training examples for fine-tuning the model. This is a required step.
- B. Incorrect.
Incorrect. The endpoint URL is automatically generated for the fine-tuned model. You do not configure it manually during the fine-tuning process.
- C. Correct.
Correct. The training dataset must be uploaded to Azure Blob Storage or another accessible location for the Azure OpenAI service to access it.
- D. Correct.
Correct. Fine-tuning in Azure OpenAI can be initiated using Azure OpenAI Studio, Azure CLI, or REST API with the prepared dataset.
- E. Incorrect.
Incorrect. Learning rate and epochs are pre-configured by Azure OpenAI during the fine-tuning process, and manual adjustments are not required.