GROK-1

Introduction

GROK-1 is an open-weights text generation model developed by XAI-ORG. The model is available on Hugging Face under the Apache 2.0 license, which allows for extensive use and modification.

Architecture

GROK-1 is designed for text generation tasks. It consists of a large-scale architecture with 314 billion parameters, indicating its requirement for substantial computational resources, particularly multi-GPU setups.

Training

Details about the training process for GROK-1 are not provided in the documentation. However, its large size suggests it underwent extensive training to handle various text generation tasks effectively.

Guide: Running Locally

To run GROK-1 locally, follow these steps:

1. Clone the Repository:

   git clone https://github.com/xai-org/grok-1.git && cd grok-1
   

2. Install Dependencies:

   pip install huggingface_hub[hf_transfer]
   

3. Download Checkpoints:

   huggingface-cli download xai-org/grok-1 --repo-type model --include ckpt-0/* --local-dir checkpoints --local-dir-use-symlinks False
   

4. Install Requirements:

   pip install -r requirements.txt
   

5. Run the Model:

   python run.py
   

   Ensure you have a multi-GPU machine due to the model's size.

Cloud GPUs

To efficiently run GROK-1, consider using cloud-based GPU providers such as AWS EC2, Google Cloud Platform, or Azure.

License

GROK-1 is licensed under the Apache 2.0 License, allowing you to freely use, modify, and distribute the model with proper attribution.