Perplexity - GlobalFeed

Perplexity

Perplexity开源基于Qwen的决策模型pplx-decider

Perplexity 在 Hugging Face 上以 Apache 2.0 许可证发布了基于 Qwen3.8-27B 微调的决策模型 pplx-decider-v1-27b。该模型无需生成长篇回答，直接为预设选项给出概率，同时也是 Perplexity 新推出的 Decisions API 背后的模型。

Perplexity

Perplexity releases pplx-decider, an open decision model built on Qwen

Perplexity has published pplx-decider-v1-27b, a decision model fine-tuned from Qwen3.8-27B, on Hugging Face under Apache 2.0. It returns probabilities for predefined options without writing long replies, and it also powers Perplexity's new Decisions API.

Perplexity

Perplexity, karar modeli pplx-decider'ı açık model olarak yayımladı

Perplexity, Qwen3.8-27B üzerine geliştirdiği karar modeli pplx-decider-v1-27b'yi Hugging Face'te Apache 2.0 lisansıyla yayımladı. Model uzun yanıt yazmadan seçeneklere olasılık veriyor ve Perplexity'nin yeni Decisions API hizmetinin de arkasında çalışıyor.

Perplexity

Perplexity'den metin ve görselde arama yapan açık model ailesi: pplx-embed-v2-late

Perplexity, metin ve görsellerde arama için geliştirdiği pplx-embed-v2-late gömme modeli ailesini tanıttı. 0.6B ve 9B olmak üzere iki boyu olan modeller aynı vektör uzayını paylaşıyor ve PDF sayfalarını metne çevirmeden arayabiliyor. Model dosyaları Hugging Face'te MIT lisansıyla indirilebiliyor.

Perplexity

Perplexity تطرح نموذج القرار pplx-decider مفتوحا

نشرت Perplexity على Hugging Face نموذج القرار pplx-decider-v1-27b المبني على Qwen3.8-27B بترخيص Apache 2.0. ويعطي النموذج احتمالا لكل خيار محدد مسبقا دون كتابة رد طويل، كما يشغّل خدمة Decisions API الجديدة من Perplexity.

Perplexity

Perplexity تطرح نماذج التضمين pplx-embed-v2-late مفتوحة للبحث في النصوص والصور

قدمت Perplexity عائلة نماذج التضمين pplx-embed-v2-late للبحث في النصوص والصور. ويتشارك النموذجان 0.6B و9B فضاء متجهات واحدا، ويستطيعان البحث في صفحات PDF دون تحويلها إلى نص أولا. وملفات النماذج متاحة على Hugging Face بترخيص MIT.

Perplexity

Perplexity开源可检索图文的多向量嵌入模型pplx-embed-v2-late

Perplexity 发布了用于文本与图像检索的嵌入模型系列 pplx-embed-v2-late。0.6B 和 9B 两种规格共享同一向量空间，无需先把 PDF 页面转换成文本即可直接检索。模型文件以 MIT 许可证在 Hugging Face 上开放下载。

Perplexity

Perplexity releases pplx-embed-v2-late, open embedding models for text and image search

Perplexity introduced pplx-embed-v2-late, a family of embedding models for searching text and images. The 0.6B and 9B models share one embedding space and can search PDF pages without converting them to text first. The weights are on Hugging Face under the MIT license.