Veja em ação

Exemplo de uso

DSPy: see a signature produce the author’s haiku

Build the first haiku program from the author’s complete LM setup and signature example. The printed poem is the tutorial’s illustrative output; its exact generation model is unspecified. The copied recipe preserves the setup model string and API-key placeholder.

Entrada de exemplo

import dspy

# Pass the key explicitly...
lm = dspy.LM("openai/gpt-5-nano", api_key="YOUR_OPENAI_API_KEY")

dspy.configure(lm=lm)

haiku_signature = "subject -> haiku"
haiku_generator = dspy.Predict(haiku_signature)
result = haiku_generator(subject="computer science")
print(result.haiku)

Resultado de exemplo

Silent code unfolds
Logic threads through hidden paths
Bugs bloom, then resolve

Exemplo original · ENVer original · stanfordnlp/dspy Licença · MIT

Licença

Complete original license and any source copyright retained. Source-authored examples are separate from site-reproduced results; no original scripts, installations, model/API requests, account creation, memory writes, graph queries or tests were executed. Complete pinned source blocks are assembled in documented order, with their separate file bindings retained in the review record. The SDK setup and printed illustrative poem come from separate source sections; no exact generation-model binding is claimed.

MIT License

Copyright (c) 2023 Stanford Future Data Systems

Permission is hereby granted, free of charge, to any person obtaining a copy
of this software and associated documentation files (the "Software"), to deal
in the Software without restriction, including without limitation the rights
to use, copy, modify, merge, publish, distribute, sublicense, and/or sell
copies of the Software, and to permit persons to whom the Software is
furnished to do so, subject to the following conditions:

The above copyright notice and this permission notice shall be included in all
copies or substantial portions of the Software.

THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE
SOFTWARE.

Exemplo de uso

DSPy: specify setting, mood and season

The full source signature adds setting, mood and a Literal season before calling the haiku program for Bodega Bay. The author’s printed poem stays alongside the recipe. Setup and sample output come from separate documented sections, so the poem’s exact model is unspecified.

Entrada de exemplo

import dspy

# Pass the key explicitly...
lm = dspy.LM("openai/gpt-5-nano", api_key="YOUR_OPENAI_API_KEY")

dspy.configure(lm=lm)

from typing import Literal

Season = Literal[
    "spring", "summer", "autumn", "winter",
]

class HaikuBot(dspy.Signature):
    """
    Write a classical haiku given the provided inputs.
    """
    location: str = dspy.InputField()
    mood: str = dspy.InputField()
    season: Season = dspy.InputField()
    haiku: str = dspy.OutputField()

haiku_bot = dspy.Predict(HaikuBot)
result = haiku_bot(location="Bodega Bay", mood="mysterious", season="autumn")
print(result.haiku)

Resultado de exemplo

Fog drifts over waves,  
Crimson leaves swirl by the pier—  
Night whispers secrets.

Exemplo original · ENVer original · stanfordnlp/dspy Licença · MIT

Licença

MIT License

Copyright (c) 2023 Stanford Future Data Systems

Permission is hereby granted, free of charge, to any person obtaining a copy
of this software and associated documentation files (the "Software"), to deal
in the Software without restriction, including without limitation the rights
to use, copy, modify, merge, publish, distribute, sublicense, and/or sell
copies of the Software, and to permit persons to whom the Software is
furnished to do so, subject to the following conditions:

The above copyright notice and this permission notice shall be included in all
copies or substantial portions of the Software.

THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE
SOFTWARE.

RECURSOS LISTADOS NA FONTE

O que o DSPy oferece

01Define model tasks with typed signatures
02Compose reusable modules in Python
03Optimize prompts using examples and evaluation metrics

Texto original da fonte · EN

O que o diferencia

DSPy treats model calls as parts of a program that can be evaluated and optimized. Useful results depend on representative examples and the metric supplied by the developer.

Texto original da fonte · EN

Para quem é indicado

Developers evaluating language-model pipelines
Teams tuning repeatable AI tasks

Texto original da fonte · EN

ANTES DE ESCOLHER

O que confirmar

Experimente uma tarefa real e compare o resultado com os seus requisitos.
Verifique que dados são guardados, partilhados ou usados para treino.
Confirme preços, limites de utilização e condições de cancelamento no site oficial.

Última verificação

Link oficial verificado em 20 de out. de 2026. Os recursos não foram testados de forma independente.