Skip to content

Commit

Permalink
💄 style: support thinking for all non DeepSeek official api R1 models (
Browse files Browse the repository at this point in the history
…lobehub#5654)

* wip  ollama thinking

* revert

* update locale

* improve style

* support ollama thinking

* fix edge case
  • Loading branch information
arvinxx authored Feb 1, 2025
1 parent e5cda3a commit 9b32137
Show file tree
Hide file tree
Showing 49 changed files with 262 additions and 64 deletions.
4 changes: 2 additions & 2 deletions locales/ar/components.json
Original file line number Diff line number Diff line change
Expand Up @@ -88,7 +88,7 @@
},
"Thinking": {
"thinking": "في حالة تفكير عميق...",
"thought": "تم التفكير بعمق",
"thoughtWithDuration": "تم التفكير بعمق (استغرق {{duration}} ثانية)"
"thought": "لقد فكرت بعمق (استغرق الأمر {{duration}} ثانية)",
"thoughtWithDuration": "لقد فكرت بعمق"
}
}
3 changes: 3 additions & 0 deletions locales/ar/providers.json
Original file line number Diff line number Diff line change
Expand Up @@ -23,6 +23,9 @@
"deepseek": {
"description": "DeepSeek هي شركة تركز على أبحاث وتطبيقات تقنيات الذكاء الاصطناعي، حيث يجمع نموذجها الأحدث DeepSeek-V2.5 بين قدرات الحوار العامة ومعالجة الشيفرات، وقد حقق تحسينات ملحوظة في محاذاة تفضيلات البشر، ومهام الكتابة، واتباع التعليمات."
},
"doubao": {
"description": "نموذج كبير تم تطويره داخليًا بواسطة بايت دانس. تم التحقق من صحته من خلال أكثر من 50 سيناريو عمل داخلي، مع استخدام يومي يتجاوز تريليون توكن، مما يتيح تقديم قدرات متعددة الأنماط، ويعمل على توفير تجربة عمل غنية للشركات من خلال نموذج عالي الجودة."
},
"fireworksai": {
"description": "Fireworks AI هي شركة رائدة في تقديم خدمات نماذج اللغة المتقدمة، تركز على استدعاء الوظائف والمعالجة متعددة الوسائط. نموذجها الأحدث Firefunction V2 مبني على Llama-3، مُحسّن لاستدعاء الوظائف، والحوار، واتباع التعليمات. يدعم نموذج اللغة البصرية FireLLaVA-13B إدخال الصور والنصوص المختلطة. تشمل النماذج البارزة الأخرى سلسلة Llama وسلسلة Mixtral، مما يوفر دعمًا فعالًا لاتباع التعليمات وتوليدها بلغات متعددة."
},
Expand Down
4 changes: 2 additions & 2 deletions locales/bg-BG/components.json
Original file line number Diff line number Diff line change
Expand Up @@ -88,7 +88,7 @@
},
"Thinking": {
"thinking": "В процес на дълбочинно размисъл...",
"thought": "Дълбочинно размислено",
"thoughtWithDuration": "Дълбочинно размислено (време: {{duration}} секунди)"
"thought": "Дълбоко размислих (отне ми {{duration}} секунди)",
"thoughtWithDuration": "Дълбоко размислих"
}
}
3 changes: 3 additions & 0 deletions locales/bg-BG/providers.json
Original file line number Diff line number Diff line change
Expand Up @@ -23,6 +23,9 @@
"deepseek": {
"description": "DeepSeek е компания, специализирана в изследвания и приложения на технологии за изкуствен интелект, чийто най-нов модел DeepSeek-V2.5 комбинира способности за общи диалози и обработка на код, постигайки значителни подобрения в съответствието с човешките предпочитания, писателските задачи и следването на инструкции."
},
"doubao": {
"description": "Модел, разработен от ByteDance. Проверен в над 50 бизнес сценария в рамките на компанията, с ежедневна употреба на триллиони токени, който продължава да се усъвършенства, предоставяйки множество модални възможности и създавайки богато бизнес изживяване с висококачествени модели."
},
"fireworksai": {
"description": "Fireworks AI е водещ доставчик на напреднали езикови модели, фокусирайки се върху извикване на функции и мултимодална обработка. Най-новият им модел Firefunction V2, базиран на Llama-3, е оптимизиран за извикване на функции, диалози и следване на инструкции. Визуалният езиков модел FireLLaVA-13B поддържа смесени входове от изображения и текст. Други забележителни модели включват серията Llama и серията Mixtral, предлагащи ефективна поддръжка за многоезично следване на инструкции и генериране."
},
Expand Down
4 changes: 2 additions & 2 deletions locales/de-DE/components.json
Original file line number Diff line number Diff line change
Expand Up @@ -88,7 +88,7 @@
},
"Thinking": {
"thinking": "Tiefes Nachdenken...",
"thought": "Tief nachgedacht",
"thoughtWithDuration": "Tief nachgedacht (Dauer: {{duration}} Sekunden)"
"thought": "Tiefgründig nachgedacht (Dauer: {{duration}} Sekunden)",
"thoughtWithDuration": "Tiefgründig nachgedacht"
}
}
3 changes: 3 additions & 0 deletions locales/de-DE/providers.json
Original file line number Diff line number Diff line change
Expand Up @@ -23,6 +23,9 @@
"deepseek": {
"description": "DeepSeek ist ein Unternehmen, das sich auf die Forschung und Anwendung von KI-Technologien spezialisiert hat. Ihr neuestes Modell, DeepSeek-V2.5, kombiniert allgemeine Dialog- und Codeverarbeitungsfähigkeiten und hat signifikante Fortschritte in den Bereichen menschliche Präferenzanpassung, Schreibaufgaben und Befehlsbefolgung erzielt."
},
"doubao": {
"description": "Ein von ByteDance entwickeltes großes Modell. Durch die praktische Validierung in über 50 internen Geschäftsszenarien und die kontinuierliche Verfeinerung mit täglich Billionen von Tokens bietet es vielfältige Modalitäten und schafft mit hochwertigen Modellergebnissen ein reichhaltiges Geschäftserlebnis für Unternehmen."
},
"fireworksai": {
"description": "Fireworks AI ist ein führender Anbieter von fortschrittlichen Sprachmodellen, der sich auf Funktionsaufrufe und multimodale Verarbeitung spezialisiert hat. Ihr neuestes Modell, Firefunction V2, basiert auf Llama-3 und ist für Funktionsaufrufe, Dialoge und Befehlsbefolgung optimiert. Das visuelle Sprachmodell FireLLaVA-13B unterstützt gemischte Eingaben von Bildern und Text. Weitere bemerkenswerte Modelle sind die Llama-Serie und die Mixtral-Serie, die effiziente mehrsprachige Befehlsbefolgung und Generierungsunterstützung bieten."
},
Expand Down
4 changes: 2 additions & 2 deletions locales/en-US/components.json
Original file line number Diff line number Diff line change
Expand Up @@ -88,7 +88,7 @@
},
"Thinking": {
"thinking": "Deep in thought...",
"thought": "Thought deeply",
"thoughtWithDuration": "Thought deeply (took {{duration}} seconds)"
"thought": "Deeply thought (took {{duration}} seconds)",
"thoughtWithDuration": "Deeply thought"
}
}
3 changes: 3 additions & 0 deletions locales/en-US/providers.json
Original file line number Diff line number Diff line change
Expand Up @@ -23,6 +23,9 @@
"deepseek": {
"description": "DeepSeek is a company focused on AI technology research and application, with its latest model DeepSeek-V2.5 integrating general dialogue and code processing capabilities, achieving significant improvements in human preference alignment, writing tasks, and instruction following."
},
"doubao": {
"description": "A self-developed large model launched by ByteDance. Verified through practical applications in over 50 internal business scenarios, it continuously refines its capabilities with a daily usage of trillions of tokens, providing various modal abilities to create a rich business experience for enterprises with high-quality model performance."
},
"fireworksai": {
"description": "Fireworks AI is a leading provider of advanced language model services, focusing on functional calling and multimodal processing. Its latest model, Firefunction V2, is based on Llama-3, optimized for function calling, conversation, and instruction following. The visual language model FireLLaVA-13B supports mixed input of images and text. Other notable models include the Llama series and Mixtral series, providing efficient multilingual instruction following and generation support."
},
Expand Down
4 changes: 2 additions & 2 deletions locales/es-ES/components.json
Original file line number Diff line number Diff line change
Expand Up @@ -88,7 +88,7 @@
},
"Thinking": {
"thinking": "Pensando profundamente...",
"thought": "He pensado profundamente",
"thoughtWithDuration": "He pensado profundamente (durante {{duration}} segundos)"
"thought": "He pensado profundamente (durante {{duration}} segundos)",
"thoughtWithDuration": "He pensado profundamente"
}
}
3 changes: 3 additions & 0 deletions locales/es-ES/providers.json
Original file line number Diff line number Diff line change
Expand Up @@ -23,6 +23,9 @@
"deepseek": {
"description": "DeepSeek es una empresa centrada en la investigación y aplicación de tecnologías de inteligencia artificial, cuyo modelo más reciente, DeepSeek-V2.5, combina capacidades de diálogo general y procesamiento de código, logrando mejoras significativas en alineación con preferencias humanas, tareas de escritura y seguimiento de instrucciones."
},
"doubao": {
"description": "Un modelo grande desarrollado internamente por ByteDance. Validado a través de más de 50 escenarios de negocio internos, con un uso diario de tokens en billones que se perfecciona continuamente, ofrece múltiples capacidades modales y crea experiencias comerciales ricas para las empresas con un rendimiento de modelo de alta calidad."
},
"fireworksai": {
"description": "Fireworks AI es un proveedor líder de servicios de modelos de lenguaje avanzados, enfocado en la llamada de funciones y el procesamiento multimodal. Su modelo más reciente, Firefunction V2, basado en Llama-3, está optimizado para llamadas de funciones, diálogos y seguimiento de instrucciones. El modelo de lenguaje visual FireLLaVA-13B admite entradas mixtas de imágenes y texto. Otros modelos notables incluyen la serie Llama y la serie Mixtral, que ofrecen un soporte eficiente para el seguimiento y generación de instrucciones multilingües."
},
Expand Down
4 changes: 2 additions & 2 deletions locales/fa-IR/components.json
Original file line number Diff line number Diff line change
Expand Up @@ -88,7 +88,7 @@
},
"Thinking": {
"thinking": "در حال تفکر عمیق...",
"thought": "تفکر عمیق انجام شد",
"thoughtWithDuration": "تفکر عمیق انجام شد (مدت زمان {{duration}} ثانیه)"
"thought": "به‌طور عمیق فکر شده است (مدت زمان {{duration}} ثانیه)",
"thoughtWithDuration": "به‌طور عمیق فکر شده است"
}
}
3 changes: 3 additions & 0 deletions locales/fa-IR/providers.json
Original file line number Diff line number Diff line change
Expand Up @@ -23,6 +23,9 @@
"deepseek": {
"description": "DeepSeek یک شرکت متمرکز بر تحقیق و کاربرد فناوری هوش مصنوعی است. مدل جدید آن، DeepSeek-V2.5، توانایی‌های مکالمه عمومی و پردازش کد را ترکیب کرده و در زمینه‌هایی مانند هم‌ترازی با ترجیحات انسانی، وظایف نوشتاری و پیروی از دستورات بهبود قابل توجهی داشته است."
},
"doubao": {
"description": "مدل بزرگ خودساخته شده توسط بایت‌دANCE. با تأیید در بیش از 50 سناریوی تجاری داخلی بایت‌دANCE، با استفاده روزانه از تریلیون‌ها توکن، به طور مداوم بهبود یافته و توانایی‌های چندگانه‌ای را ارائه می‌دهد تا تجربه‌های تجاری غنی را با کیفیت مدل بالا برای شرکت‌ها ایجاد کند."
},
"fireworksai": {
"description": "Fireworks AI یک ارائه‌دهنده پیشرو در خدمات مدل‌های زبان پیشرفته است که بر فراخوانی توابع و پردازش چندوجهی تمرکز دارد. جدیدترین مدل آن، Firefunction V2، بر اساس Llama-3 ساخته شده و برای فراخوانی توابع، مکالمه و پیروی از دستورات بهینه‌سازی شده است. مدل زبان تصویری FireLLaVA-13B از ورودی‌های ترکیبی تصویر و متن پشتیبانی می‌کند. سایر مدل‌های قابل توجه شامل سری Llama و سری Mixtral هستند که پشتیبانی کارآمدی از پیروی دستورات چندزبانه و تولید ارائه می‌دهند."
},
Expand Down
4 changes: 2 additions & 2 deletions locales/fr-FR/components.json
Original file line number Diff line number Diff line change
Expand Up @@ -88,7 +88,7 @@
},
"Thinking": {
"thinking": "En pleine réflexion...",
"thought": "Réflexion terminée",
"thoughtWithDuration": "Réflexion terminée (durée : {{duration}} secondes)"
"thought": "Pensée approfondie (durée : {{duration}} secondes)",
"thoughtWithDuration": "Pensée approfondie"
}
}
3 changes: 3 additions & 0 deletions locales/fr-FR/providers.json
Original file line number Diff line number Diff line change
Expand Up @@ -23,6 +23,9 @@
"deepseek": {
"description": "DeepSeek est une entreprise spécialisée dans la recherche et l'application des technologies d'intelligence artificielle, dont le dernier modèle, DeepSeek-V2.5, combine des capacités de dialogue général et de traitement de code, réalisant des améliorations significatives dans l'alignement des préférences humaines, les tâches d'écriture et le suivi des instructions."
},
"doubao": {
"description": "Un grand modèle développé en interne par ByteDance. Validé par la pratique dans plus de 50 scénarios d'affaires au sein de ByteDance, avec un volume d'utilisation quotidien de plusieurs trillions de tokens, il offre diverses capacités multimodales, créant ainsi une expérience commerciale riche grâce à des performances de modèle de haute qualité."
},
"fireworksai": {
"description": "Fireworks AI est un fournisseur de services de modèles linguistiques avancés, axé sur les appels de fonction et le traitement multimodal. Son dernier modèle, Firefunction V2, basé sur Llama-3, est optimisé pour les appels de fonction, les dialogues et le suivi des instructions. Le modèle de langage visuel FireLLaVA-13B prend en charge les entrées mixtes d'images et de texte. D'autres modèles notables incluent la série Llama et la série Mixtral, offrant un support efficace pour le suivi et la génération d'instructions multilingues."
},
Expand Down
4 changes: 2 additions & 2 deletions locales/it-IT/components.json
Original file line number Diff line number Diff line change
Expand Up @@ -88,7 +88,7 @@
},
"Thinking": {
"thinking": "Pensando profondamente...",
"thought": "Pensato profondamente",
"thoughtWithDuration": "Pensato profondamente (tempo impiegato {{duration}} secondi)"
"thought": "Ho riflettuto a lungo (tempo impiegato {{duration}} secondi)",
"thoughtWithDuration": "Ho riflettuto a lungo"
}
}
3 changes: 3 additions & 0 deletions locales/it-IT/providers.json
Original file line number Diff line number Diff line change
Expand Up @@ -23,6 +23,9 @@
"deepseek": {
"description": "DeepSeek è un'azienda focalizzata sulla ricerca e applicazione della tecnologia AI, il cui ultimo modello DeepSeek-V2.5 combina capacità di dialogo generico e elaborazione del codice, realizzando miglioramenti significativi nell'allineamento delle preferenze umane, nei compiti di scrittura e nel rispetto delle istruzioni."
},
"doubao": {
"description": "Il grande modello sviluppato internamente da ByteDance. Validato attraverso oltre 50 scenari aziendali interni, con un utilizzo quotidiano di trilioni di token che affinano continuamente il modello, offre diverse capacità multimodali, creando esperienze aziendali ricche con risultati di alta qualità."
},
"fireworksai": {
"description": "Fireworks AI è un fornitore leader di servizi di modelli linguistici avanzati, focalizzato su chiamate funzionali e elaborazione multimodale. Il suo ultimo modello Firefunction V2, basato su Llama-3, è ottimizzato per chiamate di funzione, dialogo e rispetto delle istruzioni. Il modello di linguaggio visivo FireLLaVA-13B supporta input misti di immagini e testo. Altri modelli notevoli includono la serie Llama e la serie Mixtral, offrendo supporto efficiente per il rispetto e la generazione di istruzioni multilingue."
},
Expand Down
4 changes: 2 additions & 2 deletions locales/ja-JP/components.json
Original file line number Diff line number Diff line change
Expand Up @@ -88,7 +88,7 @@
},
"Thinking": {
"thinking": "深く考えています...",
"thought": "深く考えました",
"thoughtWithDuration": "深く考えました(所要時間 {{duration}} 秒)"
"thought": "深く考えました(所要時間 {{duration}} 秒)",
"thoughtWithDuration": "深く考えました"
}
}
3 changes: 3 additions & 0 deletions locales/ja-JP/providers.json
Original file line number Diff line number Diff line change
Expand Up @@ -23,6 +23,9 @@
"deepseek": {
"description": "DeepSeekは、人工知能技術の研究と応用に特化した企業であり、最新のモデルDeepSeek-V2.5は、汎用対話とコード処理能力を融合させ、人間の好みの整合、ライティングタスク、指示の遵守などの面で顕著な向上を実現しています。"
},
"doubao": {
"description": "バイトダンスが開発した独自の大規模モデルです。バイトダンス内部の50以上のビジネスシーンでの実践を通じて検証され、毎日数兆トークンの大規模な使用量で磨かれ、多様なモーダル能力を提供し、高品質なモデル効果で企業に豊かなビジネス体験を提供します。"
},
"fireworksai": {
"description": "Fireworks AIは、先進的な言語モデルサービスのリーダーであり、機能呼び出しと多モーダル処理に特化しています。最新のモデルFirefunction V2はLlama-3に基づいており、関数呼び出し、対話、指示の遵守に最適化されています。視覚言語モデルFireLLaVA-13Bは、画像とテキストの混合入力をサポートしています。他の注目すべきモデルには、LlamaシリーズやMixtralシリーズがあり、高効率の多言語指示遵守と生成サポートを提供しています。"
},
Expand Down
4 changes: 2 additions & 2 deletions locales/ko-KR/components.json
Original file line number Diff line number Diff line change
Expand Up @@ -88,7 +88,7 @@
},
"Thinking": {
"thinking": "심층적으로 생각 중...",
"thought": "심층적으로 생각했습니다",
"thoughtWithDuration": "심층적으로 생각했습니다 (소요 시간: {{duration}} 초)"
"thought": "심층적으로 생각했습니다 (소요 시간 {{duration}} 초)",
"thoughtWithDuration": "심층적으로 생각했습니다"
}
}
3 changes: 3 additions & 0 deletions locales/ko-KR/providers.json
Original file line number Diff line number Diff line change
Expand Up @@ -23,6 +23,9 @@
"deepseek": {
"description": "DeepSeek는 인공지능 기술 연구 및 응용에 집중하는 회사로, 최신 모델인 DeepSeek-V2.5는 일반 대화 및 코드 처리 능력을 통합하고 인간의 선호 정렬, 작문 작업 및 지시 따르기 등에서 상당한 향상을 이루었습니다."
},
"doubao": {
"description": "바이트댄스가 개발한 자체 대형 모델입니다. 바이트댄스 내부의 50개 이상의 비즈니스 시나리오에서 검증되었으며, 매일 수조 개의 토큰 사용량을 지속적으로 다듬어 다양한 모드 기능을 제공하여 우수한 모델 효과로 기업에 풍부한 비즈니스 경험을 제공합니다."
},
"fireworksai": {
"description": "Fireworks AI는 기능 호출 및 다중 모드 처리를 전문으로 하는 선도적인 고급 언어 모델 서비스 제공업체입니다. 최신 모델인 Firefunction V2는 Llama-3를 기반으로 하며, 함수 호출, 대화 및 지시 따르기에 최적화되어 있습니다. 비주얼 언어 모델인 FireLLaVA-13B는 이미지와 텍스트 혼합 입력을 지원합니다. 기타 주목할 만한 모델로는 Llama 시리즈와 Mixtral 시리즈가 있으며, 효율적인 다국어 지시 따르기 및 생성 지원을 제공합니다."
},
Expand Down
4 changes: 2 additions & 2 deletions locales/nl-NL/components.json
Original file line number Diff line number Diff line change
Expand Up @@ -88,7 +88,7 @@
},
"Thinking": {
"thinking": "Diep in gedachten...",
"thought": "Diep nagedacht",
"thoughtWithDuration": "Diep nagedacht (geduurd {{duration}} seconden)"
"thought": "Diep nagedacht (tijd gebruikt {{duration}} seconden)",
"thoughtWithDuration": "Diep nagedacht"
}
}
Loading

0 comments on commit 9b32137

Please sign in to comment.