ModelScope Cookbook
A hands-on, open-source guide to building practical AI applications with open-source models, covering model selection, inference, fine-tuning, evaluation, and retrieval-augmented and agent workflows.
ModelScope Cookbook is a structured, hands-on guide to using open-source AI models in real applications. It connects practical decisions, such as which model fits a task, what hardware it needs, how to adapt it to data, and how to judge the result, to runnable examples built with tools such as EvalScope, ms-swift, DiffSynth, and Ollama, alongside RAG and agent workflows.
The book spans eight parts and 35 chapters, most of which are available, organized into learning paths such as getting started with a first inference, adapting a model through fine-tuning and evaluation, building an application that combines retrieval and tools, and creating with generative AI. Example applications include enterprise knowledge question answering, speech assistants, customer-service quality analysis, fitness coaching, product-image generation, and production-line inspection.
It can be read online with full-text search and guided paths, or served locally with a built-in web server; no build step, API key, or GPU is required for reading, though individual exercises may need them.