Extensions

Use large language models with your favourite framework, like LangChain / LLamaIndex or Transformers (HuggingFace). If you want to use your model but still use the SDK, try the LocalServer extension provided by the SDK.

Before you start

To use one of the following extensions, first install it by running pip install 'ibm-generative-ai[extension_name]' (the extension name is always lower-cased).

• Transformers (HuggingFace)
  • Run Transformers Agents
• lm_eval
  • lm-evaluation-harness CLI usage
  • lm-evaluation-harness advanced usage
• LangChain
  • Streaming response from LangChain
  • Chat with a model using LangChain
  • Use LangChain generation with a custom template.
  • LangChain SQL Agent
  • Text generation using LangChain
  • LangChain agent
  • LangChain Embeddings
  • Serialize LangChain model to a file
  • QA using native LangChain features
• LLamaIndex
  • Use a model through LLamaIndex
  • LlamaIndex Embeddings
• LocalServer
  • Customize behavior of local client
  • Use a local server with a custom model