Bài kiểm tra có liên kết hướng dẫn · Trung bình Cấp độ

Xây dựng ứng dụng LLM cục bộ với bài kiểm tra Ollama

Xây dựng bằng API cục bộ của Ollama đồng thời phân biệt suy luận cục bộ, định tuyến đám mây, khả năng tương thích, bảo mật và cấp phép mô hình.

Đường dẫn hướng dẫn liên quanXây dựng ứng dụng LLM cục bộ với Ollama
Câu hỏi 1 của 8

Where can a local Ollama API server run for development?

Xem trước mọi câu hỏi8 câu hỏi

Câu trả lời và giải thích xuất hiện khi bạn làm bài kiểm tra.

  1. Where can a local Ollama API server run for development?
    • Only on a hosted third-party endpoint
    • Inside the model weights
    • On the developer’s machine, with a client configured for its local API address
    • Only in a browser extension
  2. What does Ollama mean by OpenAI API compatibility?
    • It supports a subset of the OpenAI API, so feature compatibility should be checked
    • It implements every OpenAI endpoint and feature identically
    • It requires sending every prompt to OpenAI
    • It guarantees model outputs match
  3. When does Ollama’s local privacy statement apply?
    • To every app that calls Ollama, including cloud tools
    • Only when no model is installed
    • Only to account passwords
    • To prompts and responses processed locally; cloud model use has separate handling
  4. Does local execution automatically make an application secure?
    • Yes, all local apps are private by definition
    • No; network exposure, logs, backups, tools, and host controls still matter
    • Yes, because local APIs cannot be reached by other processes
    • No, because models cannot run locally
  5. What should a developer verify if offline use is required?
    • That the API uses JSON
    • That the model has a large context window
    • That model files are present and the app does not call cloud or external services
    • That the client uses an OpenAI SDK
  6. Why check the model’s license separately from Ollama?
    • All model licenses are identical to Ollama’s
    • The runtime and each model can have different license terms
    • The runtime automatically grants commercial rights
    • Licensing only applies to cloud models
  7. How should a local API be configured for safer use?
    • Expose it only to trusted interfaces or clients and apply appropriate host controls
    • Bind it publicly without authentication by default
    • Store API secrets in prompt text
    • Disable operating-system updates
  8. What should be measured before choosing a local model for a device?
    • Only model file size
    • Only whether the name contains “instruct”
    • The number of available API routes
    • Memory use, context needs, speed, quality, and supported features