Skip to main content
有时，在不将任何结果上传到LangSmith的情况下在本地运行评估会很有帮助。例如，如果您正在快速迭代一个提示，并希望在一些示例上进行冒烟测试，或者如果您正在验证目标函数和评估器函数是否正确定义，您可能不希望记录这些评估。 您可以通过使用LangSmith Python SDK，并向evaluate() / aevaluate()传递upload_results=False来实现这一点。 这将像往常一样运行您的应用程序和评估器，并返回相同的输出，但不会向LangSmith记录任何内容。这不仅包括实验结果，还包括应用程序和评估器的追踪记录。
如果您想将结果上传到LangSmith，但还需要在脚本中处理它们（用于质量门控、自定义聚合等），请参阅在本地读取实验结果。

​
示例

让我们来看一个示例： 需要langsmith>=0.2.0。示例也使用了pandas。 
from langsmith import Client

# 1. 创建和/或选择您的数据集
ls_client = Client()
dataset = ls_client.clone_public_dataset(
    "https://smith.langchain.com/public/a63525f9-bdf2-4512-83e3-077dc9417f96/d"
)

# 2. 定义一个评估器
def is_concise(outputs: dict, reference_outputs: dict) -> bool:
    return len(outputs["answer"]) < (3 * len(reference_outputs["answer"]))

# 3. 定义应用程序的接口
def chatbot(inputs: dict) -> dict:
    return {"answer": inputs["question"] + " is a good question. I don't know the answer."}

# 4. 运行评估
experiment = ls_client.evaluate(
    chatbot,
    data=dataset,
    evaluators=[is_concise],
    experiment_prefix="my-first-experiment",
    # 'upload_results' 是相关参数。
    upload_results=False
)

# 5. 在本地分析结果
results = list(experiment)

# 检查 'is_concise' 是否返回 False。
failed = [r for r in results if not r["evaluation_results"]["results"][0].score]

# 查看失败的输入和输出。
for r in failed:
    print(r["example"].inputs)
    print(r["run"].outputs)

# 将结果作为 Pandas DataFrame 查看。
# 必须已安装 'pandas'。
df = experiment.to_pandas()
df[["inputs.question", "outputs.answer", "reference.answer", "feedback.is_concise"]]

{'question': 'What is the largest mammal?'}
{'answer': "What is the largest mammal? is a good question. I don't know the answer."}
{'question': 'What do mammals and birds have in common?'}
{'answer': "What do mammals and birds have in common? is a good question. I don't know the answer."}
inputs.questionoutputs.answerreference.answerfeedback.is_concise
0What is the largest mammal?What is the largest mammal? is a good question. I don’t know the answer.The blue whaleFalse
1What do mammals and birds have in common?What do mammals and birds have in common? is a good question. I don’t know the answer.They are both warm-bloodedFalse
Edit this page on GitHub or file an issue.
Connect these docs to Claude, VSCode, and more via MCP for real-time answers.