mirror of
https://github.com/nerfstudio-project/nerfstudio.git
synced 2026-11-14 03:44:02 +08:00
229 lines
9.6 KiB
Python
229 lines
9.6 KiB
Python
# Copyright 2022 The Plenoptix Team. All rights reserved.
|
|
#
|
|
# Licensed under the Apache License, Version 2.0 (the "License");
|
|
# you may not use this file except in compliance with the License.
|
|
# You may obtain a copy of the License at
|
|
#
|
|
# http://www.apache.org/licenses/LICENSE-2.0
|
|
#
|
|
# Unless required by applicable law or agreed to in writing, software
|
|
# distributed under the License is distributed on an "AS IS" BASIS,
|
|
# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
|
|
# See the License for the specific language governing permissions and
|
|
# limitations under the License.
|
|
|
|
"""
|
|
Abstracts for the Pipeline class.
|
|
"""
|
|
from __future__ import annotations
|
|
|
|
import typing
|
|
from abc import abstractmethod
|
|
from typing import Any, Dict, List, Optional
|
|
|
|
import torch
|
|
import torch.distributed as dist
|
|
from torch import nn
|
|
from torch.nn import Parameter
|
|
from torch.nn.parallel import DistributedDataParallel as DDP
|
|
|
|
from nerfactory.configs import base as cfg
|
|
from nerfactory.datamanagers.base import DataManager
|
|
from nerfactory.models.base import Model
|
|
from nerfactory.utils import profiler, writer
|
|
from nerfactory.utils.callbacks import TrainingCallback, TrainingCallbackAttributes
|
|
from nerfactory.utils.writer import EventName, TimeWriter
|
|
|
|
|
|
def module_wrapper(module: nn.Module) -> nn.Module:
|
|
"""
|
|
If DDP, then return the .module. Otherwise, return the model.
|
|
"""
|
|
if isinstance(module, DDP):
|
|
return module.module
|
|
return module
|
|
|
|
|
|
class Pipeline(nn.Module):
|
|
"""The intent of this class is to provide a higher level interface for the Model
|
|
that will be easy to use for our Trainer class.
|
|
|
|
This class will contain high level functions for the model like getting the loss
|
|
dictionaries and visualization code. It should have ways to get the next iterations
|
|
training loss, evaluation loss, and generate whole images for visualization. Each model
|
|
class should be 1:1 with a pipeline that can act as a standardized interface and hide
|
|
differences in how each model takes in and outputs data.
|
|
|
|
This class's function is to hide the data manager and model classes from the trainer,
|
|
worrying about:
|
|
1) Fetching data with the data manager
|
|
2) Feeding the model the data and fetching the loss
|
|
Hopefully this provides a higher level interface for the trainer to use, and
|
|
simplifying the model classes, which each may have different forward() methods
|
|
and so on.
|
|
|
|
|
|
TODO: For viewer functionality to be added down the line, we should make sure
|
|
that we are still abstracting away the Model from the end user. The viewer function
|
|
should probably be done by adding a new iterator on the base data manager that will
|
|
talk to the actual viewer. This is probably ideal to have the viewer be
|
|
primarily located in the data manager (first because it makes sense as the
|
|
viewers main job in this context is to feed in data for the model to load)
|
|
so that we can have an easier time ensuring that the viewer is always
|
|
returning the same formatted data as for in train / eval. All this is pending changes to
|
|
be done in the future... but just bear in mind that if learned parameters are in the data manager,
|
|
the viewer may have to use those parameters as well.
|
|
|
|
|
|
Args:
|
|
config: configuration to instantiate pipeline
|
|
device: location to place model and data
|
|
test_mode: if True, loads test datset. if False, loads train/eval datasets
|
|
world_size: total number of machines available
|
|
local_rank: rank of current machine
|
|
|
|
Attributes:
|
|
self.datamanager: The data manager that will be used
|
|
self.model: The model that will be used
|
|
"""
|
|
|
|
def __init__(
|
|
self, config: cfg.PipelineConfig, device: str, test_mode: bool = False, world_size: int = 1, local_rank: int = 0
|
|
):
|
|
super().__init__()
|
|
self.datamanager: DataManager = config.datamanager.setup(
|
|
device=device, test_mode=test_mode, world_size=world_size, local_rank=local_rank
|
|
)
|
|
self.datamanager.to(device)
|
|
# TODO(ethan): get rid of scene_bounds from the model
|
|
assert self.datamanager.train_input_dataset is not None, "Missing input dataset"
|
|
self.model: Model = config.model.setup(
|
|
scene_bounds=self.datamanager.train_input_dataset.dataset_inputs.scene_bounds,
|
|
num_train_data=len(self.datamanager.train_input_dataset),
|
|
)
|
|
self.model.to(device)
|
|
|
|
self.world_size = world_size
|
|
if world_size > 1:
|
|
self.model = typing.cast(
|
|
Model, typing.cast(Model, DDP(self.model, device_ids=[local_rank], find_unused_parameters=True))
|
|
)
|
|
dist.barrier(device_ids=[local_rank])
|
|
|
|
@property
|
|
def device(self):
|
|
"""Returns the device that the model is on."""
|
|
return module_wrapper(self.model).device
|
|
|
|
@abstractmethod
|
|
@profiler.time_function
|
|
def get_train_loss_dict(self, step: int):
|
|
"""This function gets your training loss dict. This will be responsible for
|
|
getting the next batch of data from the DataManager and interfacing with the
|
|
Model class, feeding the data to the model's forward function.
|
|
|
|
Args:
|
|
step: current iteration step to update sampler if using DDP (distributed)
|
|
"""
|
|
if self.world_size > 1 and step:
|
|
assert self.datamanager.train_sampler is not None
|
|
self.datamanager.train_sampler.set_epoch(step)
|
|
ray_bundle, batch = self.datamanager.next_train(step)
|
|
model_outputs, loss_dict, metrics_dict = self.model(ray_bundle, batch)
|
|
return model_outputs, loss_dict, metrics_dict
|
|
|
|
@abstractmethod
|
|
@profiler.time_function
|
|
def get_eval_loss_dict(self, step: int):
|
|
"""This function gets your evaluation loss dict. It needs to get the data
|
|
from the DataManager and feed it to the model's forward function
|
|
|
|
Args:
|
|
step: current iteration step
|
|
"""
|
|
self.eval()
|
|
if self.world_size > 1:
|
|
assert self.datamanager.eval_sampler is not None
|
|
self.datamanager.eval_sampler.set_epoch(step)
|
|
ray_bundle, batch = self.datamanager.next_eval(step)
|
|
model_outputs, loss_dict, metrics_dict = self.model(ray_bundle, batch)
|
|
self.train()
|
|
return model_outputs, loss_dict, metrics_dict
|
|
|
|
@abstractmethod
|
|
@profiler.time_function
|
|
def get_eval_image_metrics_and_images(self, step: int):
|
|
"""This function gets your evaluation loss dict. It needs to get the data
|
|
from the DataManager and feed it to the model's forward function
|
|
|
|
Args:
|
|
step: current iteration step
|
|
"""
|
|
self.eval()
|
|
image_idx, camera_ray_bundle, batch = self.datamanager.next_eval_image(step)
|
|
with TimeWriter(None, None, write=False) as test_t:
|
|
outputs = module_wrapper(self.model).get_outputs_for_camera_ray_bundle(camera_ray_bundle)
|
|
metrics_dict, images_dict = module_wrapper(self.model).get_image_metrics_and_images(outputs, batch)
|
|
writer.put_time(
|
|
name=EventName.TEST_RAYS_PER_SEC,
|
|
duration=len(camera_ray_bundle) / test_t.duration,
|
|
step=step,
|
|
avg_over_steps=True,
|
|
)
|
|
metrics_dict["image_idx"] = image_idx
|
|
self.train()
|
|
return metrics_dict, images_dict
|
|
|
|
@abstractmethod
|
|
@profiler.time_function
|
|
def get_average_eval_image_metrics(self, step: Optional[int] = None):
|
|
"""Iterate over all the images in the eval dataset and get the average."""
|
|
self.eval()
|
|
metrics_dict_list = []
|
|
# TODO: add something like tqdm but so that it doesn't interfere with logging
|
|
for camera_ray_bundle, batch in self.datamanager.fixed_indices_eval_dataloader:
|
|
outputs = module_wrapper(self.model).get_outputs_for_camera_ray_bundle(camera_ray_bundle)
|
|
metrics_dict, _ = module_wrapper(self.model).get_image_metrics_and_images(outputs, batch)
|
|
metrics_dict_list.append(metrics_dict)
|
|
# average the metrics list
|
|
metrics_dict = {}
|
|
for key in metrics_dict_list[0].keys():
|
|
metrics_dict[key] = float(
|
|
torch.mean(torch.tensor([metrics_dict[key] for metrics_dict in metrics_dict_list]))
|
|
)
|
|
self.train()
|
|
return metrics_dict
|
|
|
|
@abstractmethod
|
|
def log_test_image_outputs(self) -> None:
|
|
"""Log the test image outputs"""
|
|
|
|
def load_pipeline(self, loaded_state: Dict[str, Any]) -> None:
|
|
"""Load the checkpoint from the given path
|
|
|
|
Args:
|
|
loaded_state: pre-trained model state dict
|
|
"""
|
|
state = {key.replace("module.", ""): value for key, value in loaded_state.items()}
|
|
self.load_state_dict(state) # type: ignore
|
|
|
|
def get_training_callbacks(
|
|
self, training_callback_attributes: TrainingCallbackAttributes
|
|
) -> List[TrainingCallback]:
|
|
"""Returns the training callbacks from both the Dataloader and the Model."""
|
|
datamanager_callbacks = self.datamanager.get_training_callbacks(training_callback_attributes)
|
|
model_callbacks = module_wrapper(self.model).get_training_callbacks(training_callback_attributes)
|
|
callbacks = datamanager_callbacks + model_callbacks
|
|
return callbacks
|
|
|
|
def get_param_groups(self) -> Dict[str, List[Parameter]]:
|
|
"""Get the param groups for the pipeline.
|
|
|
|
Returns:
|
|
A list of dictionaries containing the pipeline's param groups.
|
|
"""
|
|
datamanager_params = self.datamanager.get_param_groups()
|
|
model_params = module_wrapper(self.model).get_param_groups()
|
|
# TODO(ethan): assert that key names don't overlap
|
|
return {**datamanager_params, **model_params}
|