This is the Linux app named Coconut whose latest release can be downloaded as coconutsourcecode.tar.gz. It can be run online in the free hosting provider OnWorks for workstations.
Download and run online this app named Coconut 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.
SCREENSHOTS
Ad
Coconut
DESCRIPTION
Coconut is the official PyTorch implementation of the research paper “Training Large Language Models to Reason in a Continuous Latent Space.” The framework introduces a novel method for enhancing large language models (LLMs) with continuous latent reasoning steps, enabling them to generate and refine reasoning chains within a learned latent space rather than relying solely on discrete symbolic reasoning. It supports training across multiple reasoning paradigms—including standard Chain-of-Thought (CoT), no-thought, and hybrid configurations—using configurable training stages and latent representations. The repository is built with Hugging Face Transformers, PyTorch Distributed, and Weights & Biases (wandb) for logging, supporting large-scale experiments on mathematical and logical reasoning datasets such as GSM8K, ProntoQA, and ProsQA.
Features
- Reproducible experiment scripts matching the paper’s benchmark protocols
- Supports distributed multi-GPU training with torchrun and mixed-precision (bf16)
- Dataset preprocessing tools for GSM8K, ProntoQA, and ProsQA
- Integrated wandb logging and checkpoint management across training stages
- Modular YAML-based configuration for multi-stage training and evaluation
- Implements continuous latent reasoning for LLMs beyond discrete CoT prompting
Programming Language
Python, Unix Shell
Categories
This is an application that can also be fetched from https://sourceforge.net/projects/coconut.mirror/. It has been hosted in OnWorks in order to be run online in an easiest way from one of our free Operative Systems.
