repeat problem
#17
by
gulixin0922
- opened
I used model：command-r-plus with langchain. It has repeat problem.
chat_llm:
type: 'ChatCohere'
model: 'command-r-plus'
cohere_api_key: ''
max_tokens: 1000
temperature: 0.01
Try to pass a parameter called repetition_penalty with a value of 1.1.
Without this, the output begins repeating.