This is a repo with minimal tooling, modules, models and recipes to get easily get started with deep learning training and experimentation with an emphasis on speech, audio and language modeling.

you need python <3.12

Install package:

pip install slg-nimrod

Install espeak for LM:

brew install espeak #macos

Install Spacy english model

python -m spacy download en_core_web_sm

Download test data on which to run example recipes:

# if not already installed on your system
git lfs install 
# update changes
git lfs fetch --all
# copy the actual data
git lfs checkout
# or just
git lfs pull # combing both steps above into one (like usual git pull)

Check recipes in recipes/ folder. E.g. for a simple digit recognizer on MNIST:

git clone https://github.com/slegroux/nimrod.git
cd nimrod/recipes/images/mnist
python train.py datamodule.num_workers=8 trainer.max_epochs=20 trainer.accelerator='mps' loggers='tensorboard'
head conf/train.yaml

All the parameters of the experiment are editable and read from a .yaml file which details:

• data and logging directory paths
• data module with data source path and batching parameters
• model architecture
• trainer with hardware acceleration and number of epochs
• callbacks for early stopping and automatic logging to Wandb

You might want to use docker containers for reproductible development environment or run your project in the cloud

make container
docker pull slegroux/nimrod
docker run -it --rm -p 8888:8888 slegroux/nimrod /bin/bash

You can also use docker-compose to define services and volumes

cd .devcontainer
docker-compose up
docker-compose down

pip install -e .

to compare training results on different model parameters:

cd nimrod/recipes/images/mnist
python train.py --multirun model.n_h=16,64,256 loggers='tensorboard' trainer.max_epochs=5

Run a simple digit recognizer webapp with GUI

cd server
./run_st_app.sh

2023 Sylvain Le Groux [email protected]