ai-agents-llm-integration-stop-ollama-chatollama-with-llama3-model-streams-endles-e727c278

THEN Add an explicit stop condition by passing the stop parameter with the Llama3 end-of-turn token: <|eot_id|>. For example: ChatOllama(model='llama3:70b', stop=['<|eot_id|>']). This ensures the generation terminates properly.

from langchain_community.chat_models.ollama import ChatOllama
llm = ChatOllama(model='llama3:70b', stop=['<|eot_id|>'])
for chunk in llm.stream('Write a poem about fish'):
    print(chunk.content)