Skip to main content

deepseek_r1_8b

What it is​

  • A module that defines a preconfigured ChatModel instance for DeepSeek R1 8B served via Ollama using langchain_ollama.ChatOllama.

Public API​

  • Module constant: model: ChatModel
    • A ready-to-use chat model wrapper with:
      • model_id: "deepseek-r1:8b"
      • name: "DeepSeek R1 8B"
      • description: reasoning-oriented model description
      • image: Ollama logo URL
      • provider: "ollama"
      • context_window: 32768
      • underlying LangChain model: ChatOllama(model="deepseek-r1:8b", temperature=0.1)

Configuration/Dependencies​

  • Dependencies
    • langchain_ollama.ChatOllama
    • naas_abi_core.models.Model.ChatModel
  • Runtime requirements
    • An Ollama runtime with the deepseek-r1:8b model available (as implied by provider="ollama" and model_id="deepseek-r1:8b").

Usage​

from naas_abi_marketplace.ai.deepseek.models.deepseek_r1_8b import model

# Access metadata
print(model.model_id)     # deepseek-r1:8b
print(model.provider)     # ollama

# Use the underlying LangChain chat model
llm = model.model
response = llm.invoke("Explain the Pythagorean theorem briefly.")
print(response)

Caveats​

  • The module only exposes a pre-instantiated ChatModel; it does not define additional helper functions.
  • Token prediction limits are not configured (the num_predict setting is present but commented out).