Voir en action

Exemple d’utilisation

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.

Exemple d’entrée

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)

Exemple de résultat

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

Exemple original · ENVoir l’original · stanfordnlp/dspy Licence · MIT

Licence

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.

Exemple d’utilisation

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.

Exemple d’entrée

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)

Exemple de résultat

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

Exemple original · ENVoir l’original · stanfordnlp/dspy Licence · MIT

Licence

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.

FONCTIONS INDIQUÉES PAR LA SOURCE

Ce que propose DSPy

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

Texte original de la source · EN

Ce qui le distingue

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.

Texte original de la source · EN

À qui il convient

Developers evaluating language-model pipelines
Teams tuning repeatable AI tasks

Texte original de la source · EN

AVANT DE CHOISIR

Points à vérifier

Tester une vraie tâche et comparer le résultat à ses besoins.
Vérifier quelles données sont conservées, partagées ou utilisées pour l’entraînement.
Confirmer les tarifs, limites d’usage et conditions de résiliation sur le site officiel.

Dernière vérification

Lien officiel vérifié le 20 oct. 2026. Les fonctionnalités n’ont pas été testées indépendamment.