cvrunner package
Subpackages
- cvrunner.experiment package
- Submodules
- cvrunner.experiment.experiment module
BaseExperimentBaseExperiment.batch_sizeBaseExperiment.build_criterion()BaseExperiment.build_dataloader()BaseExperiment.build_dataset()BaseExperiment.build_loss_function()BaseExperiment.build_model()BaseExperiment.build_optimizer_scheduler()BaseExperiment.load_checkpoint()BaseExperiment.num_epochsBaseExperiment.random_seedBaseExperiment.runner_cls()BaseExperiment.save_checkpoint()BaseExperiment.train_epoch_end()BaseExperiment.train_epoch_start()BaseExperiment.train_step()BaseExperiment.val_epoch_end()BaseExperiment.val_epoch_start()BaseExperiment.val_freqBaseExperiment.val_step()BaseExperiment.wandb_projectBaseExperiment.wandb_runname
- Module contents
- cvrunner.runner package
Submodules
cvrunner.cli module
- cvrunner.cli.build_and_push_image(source_image, target_image, context)
- Return type:
None
- cvrunner.cli.build_docker_image(source_image, target_image, env_dir)
Build Docker image using docker-compose with temporary override files.
- Return type:
None
- cvrunner.cli.docker_image_exists(image)
Check if a Docker image exists locally.
- Return type:
bool
- cvrunner.cli.get_compose_cmd()
Get the docker compose command.
- cvrunner.cli.main()
- Return type:
None
- cvrunner.cli.run_in_docker(args)
Run experiment inside Docker container.
- Return type:
None
- cvrunner.cli.run_local(args)
Run experiment locally using torchrun. This handles both Single-GPU and Multi-GPU transparently.
- Return type:
None
- cvrunner.cli.run_on_k8s(args)
Submit Job to K8s.
- Return type:
None
- cvrunner.cli.sync_env_secret(context)
- Return type:
None