This is the Linux app named JEPA whose latest release can be downloaded as jepasourcecode.tar.gz. It can be run online in the free hosting provider OnWorks for workstations.
Download and run online this app named JEPA with OnWorks for free.
برای اجرای این برنامه این دستورالعمل ها را دنبال کنید:
- 1. این برنامه را در رایانه شخصی خود دانلود کنید.
- 2. در فایل منیجر ما https://www.onworks.net/myfiles.php?username=XXXXX نام کاربری مورد نظر خود را وارد کنید.
- 3. این برنامه را در چنین فایل منیجر آپلود کنید.
- 4. OnWorks Linux آنلاین یا شبیه ساز آنلاین ویندوز یا شبیه ساز آنلاین MACOS را از این وب سایت راه اندازی کنید.
- 5. از سیستم عامل لینوکس OnWorks که به تازگی راه اندازی کرده اید، به مدیر فایل ما https://www.onworks.net/myfiles.php?username=XXXXX با نام کاربری که می خواهید بروید.
- 6. اپلیکیشن را دانلود کرده، نصب و اجرا کنید.
عکس ها
Ad
جپا
شرح
JEPA (Joint-Embedding Predictive Architecture) captures the idea of predicting missing high-level representations rather than reconstructing pixels, aiming for robust, scalable self-supervised learning. A context encoder ingests visible regions and predicts target embeddings for masked regions produced by a separate target encoder, avoiding low-level reconstruction losses that can overfit to texture. This makes learning focus on semantics and structure, yielding features that transfer well with simple linear probes and minimal fine-tuning. The repository provides training recipes, data pipelines, and evaluation utilities for image JEPA variants and often includes ablations that illuminate which masking and architectural choices matter. Because the objective is non-autoregressive and operates in embedding space, JEPA tends to be compute-efficient and stable at scale. The approach has become a strong alternative to contrastive or pixel-reconstruction methods for representation learning.
امکانات
- Predictive learning in embedding space instead of pixel reconstruction
- Separate context and target encoders with masked-region objectives
- Strong linear-probe and low-shot transfer performance
- Stable, efficient training without heavy negative sampling
- Clear recipes and ablations for masking and architecture choices
- Modular code for extending to new modalities or datasets
زبان برنامه نویسی
پــایتــون
دسته بندی ها
This is an application that can also be fetched from https://sourceforge.net/projects/jepa.mirror/. It has been hosted in OnWorks in order to be run online in an easiest way from one of our free Operative Systems.