This is the Linux app named Retro whose latest release can be downloaded as v0.8.0sourcecode.tar.gz. It can be run online in the free hosting provider OnWorks for workstations.
Download and run online this app named Retro with OnWorks for free.
Follow these instructions in order to run this app:
- 1. Downloaded this application in your PC.
- 2. Enter in our file manager https://www.onworks.net/myfiles.php?username=XXXXX with the username that you want.
- 3. Upload this application in such filemanager.
- 4. Start the OnWorks Linux online or Windows online emulator or MACOS online emulator from this website.
- 5. From the OnWorks Linux OS you have just started, goto our file manager https://www.onworks.net/myfiles.php?username=XXXXX with the username that you want.
- 6. Download the application, install it and run it.
Retro
Ad
DESCRIPTION
RETRO (Retrieval-Enhanced Transformer) is a large language model architecture developed by OpenAI that augments transformer models with a retrieval mechanism. Instead of relying solely on learned parameters, RETRO retrieves relevant documents from a large external database during inference, allowing it to ground responses in external knowledge. This design improves factual accuracy, reduces hallucinations, and enables smaller models to perform comparably to much larger ones by leveraging retrieval. The repository provides code and resources for training and evaluating RETRO models, along with infrastructure for integrating retrieval into the transformer pipeline. It includes example configurations, datasets, and utilities for building retrieval-augmented generation systems. RETRO represents an important step toward combining large-scale language modeling with information retrieval, offering researchers a foundation to study hybrid approaches for scaling AI responsibly.
Features
- Implements the Retrieval-Enhanced Transformer (RETRO) architecture
- Integrates external document retrieval into model inference
- Improves factual grounding and reduces hallucinations
- Enables smaller models to match performance of larger ones
- Provides code and configurations for training and evaluation
- Supports research into retrieval-augmented generation systems
Programming Language
C, C++, Python
Categories
This is an application that can also be fetched from https://sourceforge.net/projects/retro.mirror/. It has been hosted in OnWorks in order to be run online in an easiest way from one of our free Operative Systems.
