Skip to content

About

An upgrade framework for train and validate compare with icefall using Lightning.

Topics

Resources

Stars

16 stars

Watchers

3 watching

Forks

Latest commit

 

History

13 Commits

Folders and files

Repository files navigation

Introduction

This repository builds upon the incredible work of icefall, integrating and extending its features to provide a convenient and controllable framework for model training and testing. The base line code use Zipformer with Eden scheduler and ScaledAdam optimizer. Feel free to customize the provided configuration file and layer modules to suit your specific model and training requirements.

How to use

Installation

Step 1: Install this repo

git clone https://github.com/ZQuang2202/Zipformer_Lightning.git
cd Zipformer_Lightning

Step 2: Create environment

conda create -n zipformer_lightning python=3.10
export CUDA_HOME=/usr/local/cuda
export PATH=$CUDA_HOME/bin:$PATH
export LD_LIBRARY_PATH=$CUDA_HOME/lib64:$LD_LIBRARY_PATH
export LD_LIBRARY_PATH=$CUDA_HOME/lib:$LD_LIBRARY_PATH
export LD_LIBRARY_PATH=$CUDA_HOME/extras/CUPTI/lib64:$LD_LIBRARY_PATH
export CUDAToolkit_ROOT_DIR=$CUDA_HOME
export CUDAToolkit_ROOT=$CUDA_HOME
export CMAKE_CUDA_ARCHITECTURES=75
conda activate zipformer_lightning
pip install -r requirements.txt

Step 3: Install Warp RNNT

bash scripts/install_rnnt.sh

Step 4: Install the repo to an editable package

pip install -e .

Training and Test

Customize your configuration at the config file. Customize your model and modules at here.

To train and test:

cd egs/librispeech
export CUDA_VISIBLE_DEVICES=0,1 & python3 run.py --config configs/zipformer.yaml

About

An upgrade framework for train and validate compare with icefall using Lightning.

Topics

Resources

Stars

16 stars

Watchers

3 watching

Forks

Releases

Packages

Used by

Contributors

Languages