nerfbaselines.train

class nerfbaselines.train.Trainer(*, train_dataset: str | ~pathlib.Path | ~typing.Callable[[], ~nerfbaselines.types.Dataset], test_dataset: None | str | ~pathlib.Path | ~typing.Callable[[], ~nerfbaselines.types.Dataset] = None, method: ~typing.Type[~nerfbaselines.types.Method], output: ~pathlib.Path = PosixPath('.'), num_iterations: int | None = None, save_iters: ~nerfbaselines.utils.Indices = <nerfbaselines.utils.Indices object>, eval_single_iters: ~nerfbaselines.utils.Indices = <nerfbaselines.utils.Indices object>, eval_all_iters: ~nerfbaselines.utils.Indices = <nerfbaselines.utils.Indices object>, use_wandb: bool = True, color_space: ~typing.Literal['srgb', 'linear'] | None = None)

Bases: object

close()

ensure_loggers_initialized()

eval_all()

eval_single()

install()

log_metrics(metrics, prefix: str = '')

save()

setup_data()

train()

train_iteration()

nerfbaselines.train.compute_exponential_gamma(num_iters: int, initial_lr: float, final_lr: float) → float

nerfbaselines.train.compute_image_metrics(pred, gt)

nerfbaselines.train.make_grid(*images, ncol=None, padding=2, max_width=1920, background=1.0)