ColossalAI/examples/tutorial/hybrid_parallel

Multi-dimensional Parallelism with Colossal-AI

Table of contents

• Overview
• Quick Start

📚 Overview

This example lets you to quickly try out the hybrid parallelism provided by Colossal-AI. You can change the parameters below to try out different settings in the config.py.

# parallel setting
TENSOR_PARALLEL_SIZE = 2
TENSOR_PARALLEL_MODE = '1d'

parallel = dict(
    pipeline=2,
    tensor=dict(mode=TENSOR_PARALLEL_MODE, size=TENSOR_PARALLEL_SIZE),
)

🚀 Quick Start

1. Install PyTorch

2. Install the dependencies.

pip install -r requirements.txt

3. Run the training scripts with synthetic data.

colossalai run --nproc_per_node 4 train.py --config config.py

4. Modify the config file to play with different types of tensor parallelism, for example, change tensor parallel size to be 4 and mode to be 2d and run on 8 GPUs.