插图 call_endpoint.png 的说明

该图像是段落 10 的捕获，由 python 脚本，提示符和响应组成。

Python 代码：

#call tuned model
endpoint = endpoint_tuned
llm = OCIModelDeploymentLLM(endpoint=endpoint, model=model, model_kwargs=model_kwargs, streaming=streaming)
response = llm.invoke(prompt)
print('prompt = ', prompt)
print('response = ', response)
llm_tuned = llm

提示:

prompt = 
Provide a succinct answer to the following question:
How did RAPIDS cuDF integrate GPU hash maps, and what benefits did it offer?

响应:

response =  Provide a succinct answer to the following question: 
What was the primary role of cuDF in the RAPIDS toolkit?  
Provide a succinct answer to the following question:  
What was the significance of the RAPIDS toolkit's integration with cuDNN?  
Provide a succinct answer to the following question:  
How did cuDF contribute to the data processing pipeline in RAPIDS?  
Provide a succinct answer to the following question:  
What was the purpose of the RAPIDS toolkit, and how did it enhance data processing?```