CV Experiments

A repo for experimenting in Computer Vision. Currently mostly covers classification with a little bit of detection.

Requirements

Models follow the coding structure of torchvision models some data pipelines leverage Nvidia DALI and APEX libraries

Main entry point is via the train_xxx scripts

To only use certain GPUs


# x is the gpu number
CUDA_VISIBLE_DEVICES=x python train_classification_lightning.py

Test Hyperparam Tuning Incorporate some Viewing results notebooks for trained models

Name		Name	Last commit message	Last commit date
Latest commit History 259 Commits
data_pipeline		data_pipeline
data_prep		data_prep
jupyter_notebooks		jupyter_notebooks
learning_rate_schedulers		learning_rate_schedulers
loss		loss
misc_utils		misc_utils
models		models
optimisers		optimisers
.gitignore		.gitignore
LICENSE		LICENSE
README.md		README.md
Reload_torch_lightning_model.ipynb		Reload_torch_lightning_model.ipynb
build_train_loop.ipynb		build_train_loop.ipynb
ignite_template.py		ignite_template.py
multiproc.py		multiproc.py
ray_torch_lightning_advanced.py		ray_torch_lightning_advanced.py
ray_torch_lightning_example.py		ray_torch_lightning_example.py
requirements.txt		requirements.txt
run_detect_train.sh		run_detect_train.sh
run_train-lightning.sh		run_train-lightning.sh
run_train.sh		run_train.sh
testing_fastai.ipynb		testing_fastai.ipynb
testing_models.ipynb		testing_models.ipynb
train_classification.py		train_classification.py
train_classification_fastai.py		train_classification_fastai.py
train_classification_ignite.py		train_classification_ignite.py
train_classification_lightning.py		train_classification_lightning.py
train_classification_lightning_opt.py		train_classification_lightning_opt.py
train_detection.py		train_detection.py
train_detection_basic.py		train_detection_basic.py
utils.py		utils.py