Databricks Generative AI Engineer Associate Question 168
Select 2You are tasked with selecting a pre-trained model from a model hub to perform text summarization for a financial dataset. The dataset contains highly technical financial jargon and requires a model capable of producing concise, accurate summaries. Which model metadata or model card information would be most relevant for making your selection?
- A
The model's training dataset and domain specificity
- B
The model's number of parameters
- C
The evaluation metrics for summarization tasks listed in the model card
- D
The license type of the model
- E
The model's inference latency
Show answer and explanation
Correct answers: A, C
Explanation
Selecting a model for a specific task like financial text summarization requires a thorough review of the model card or metadata. Key factors include the training dataset and domain specificity, as these determine the model's ability to handle domain-specific jargon, and the evaluation metrics, which indicate how well the model performs on summarization tasks. Other factors like license type and inference latency, while important in certain contexts, are secondary considerations for this specific scenario.
- A. Correct.
Understanding the model's training dataset and domain specificity is crucial because a model trained on financial or technical texts is more likely to perform well on your financial dataset.
- B. Incorrect.
While the number of parameters can provide some insight into model capacity, it is not the most relevant factor for selecting a model for a domain-specific text summarization task.
- C. Correct.
Evaluation metrics specific to summarization tasks (e.g., ROUGE scores) help determine the model's performance for the intended task, making them highly relevant.
- D. Incorrect.
The license type may be important for usage rights but does not directly impact the model's suitability for text summarization with financial data.
- E. Incorrect.
Inference latency is important in production environments for speed but does not indicate the model's capability to handle the specific task of summarizing financial texts.