Spaces:
Running
on
Zero
Running
on
Zero
File size: 4,799 Bytes
9e15541 |
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130 131 132 133 134 135 136 137 138 139 140 141 142 143 144 145 146 147 148 149 150 151 |
import logging
from pathlib import Path
import torch
from torch import nn
from torch.utils.data import DataLoader, Subset
from scenedino.common.io.configs import load_model_config
from scenedino.models import make_model
from scenedino.datasets import make_test_dataset
from scenedino.common.geometry import distance_to_z
from scenedino.renderer import NeRFRenderer
from scenedino.common.ray_sampler import ImageRaySampler, get_ray_sampler
from scenedino.evaluation.base_evaluator import base_evaluation
from scenedino.training.trainer_downstream import BTSDownstreamWrapper
IDX = 0
logger = logging.getLogger("evaluation")
class BTSWrapper(nn.Module):
def __init__(
self,
renderer,
config,
# evaluation_fns
) -> None:
super().__init__()
self.renderer = renderer
# TODO: have a consitent sampling range
self.z_near = config.get("z_near", 3.0)
self.z_far = config.get("z_far", 80.0)
self.sampler = ImageRaySampler(self.z_near, self.z_far)
# self.evaluation_fns = evaluation_fns
@staticmethod
def get_loss_metric_names():
return ["loss", "loss_l2", "loss_mask", "loss_temporal"]
def forward(self, data):
data = dict(data)
images = torch.stack(data["imgs"], dim=1) # n, v, c, h, w
poses = torch.stack(data["poses"], dim=1) # n, v, 4, 4 w2c
projs = torch.stack(data["projs"], dim=1) # n, v, 4, 4 (-1, 1)
B, n_frames, c, h, w = images.shape
device = images.device
# Use first frame as keyframe
to_base_pose = torch.inverse(poses[:, :1, :, :])
poses = to_base_pose.expand(-1, n_frames, -1, -1) @ poses
# TODO: make configurable
ids_encoder = [0]
self.renderer.net.compute_grid_transforms(
projs[:, ids_encoder], poses[:, ids_encoder]
)
self.renderer.net.encode(
images,
projs,
poses,
ids_encoder=ids_encoder,
ids_render=ids_encoder,
images_alt=images * 0.5 + 0.5,
)
all_rays, all_rgb_gt = self.sampler.sample(images * 0.5 + 0.5, poses, projs)
data["fine"] = []
data["coarse"] = []
self.renderer.net.set_scale(0)
render_dict = self.renderer(all_rays, want_weights=True, want_alphas=True)
if "fine" not in render_dict:
render_dict["fine"] = dict(render_dict["coarse"])
render_dict["rgb_gt"] = all_rgb_gt
render_dict["rays"] = all_rays
# TODO: check if distance to z is needed
render_dict = self.sampler.reconstruct(render_dict)
render_dict["coarse"]["depth"] = distance_to_z(
render_dict["coarse"]["depth"], projs
)
render_dict["fine"]["depth"] = distance_to_z(
render_dict["fine"]["depth"], projs
)
data["fine"].append(render_dict["fine"])
data["coarse"].append(render_dict["coarse"])
data["rgb_gt"] = render_dict["rgb_gt"]
data["rays"] = render_dict["rays"]
data["z_near"] = torch.tensor(self.z_near, device=images.device)
data["z_far"] = torch.tensor(self.z_far, device=images.device)
# for eval_fn in self.evaluation_fns:
# data["metrics"].update(eval_fn(data, model=self.renderer.net))
return data
def evaluation(local_rank, config):
return base_evaluation(local_rank, config, get_dataflow, initialize)
def get_dataflow(config):
test_dataset = make_test_dataset(config["dataset"])
test_loader = DataLoader(
test_dataset, # Subset(test_dataset, torch.randperm(test_dataset.length)[:1000]),
batch_size=config.get("batch_size", 1),
num_workers=config["num_workers"],
shuffle=False,
drop_last=False,
)
return test_loader
def initialize(config: dict):
checkpoint = Path(config["checkpoint"])
logger.info(f"Loading model config from {checkpoint.parent}")
load_model_config(checkpoint.parent, config)
net = make_model(config["model"], config["downstream"])
# net = make_model(config["model"])
renderer = NeRFRenderer.from_conf(config["renderer"])
renderer = renderer.bind_parallel(net, gpus=None).eval()
# TODO: attach evaluation functions rather that add them to the wrapper
# eval_fns = []
# for eval_conf in config["evaluations"]:
# eval_fn = make_eval_fn(eval_conf)
# if eval_fn is not None:
# eval_fns.append(eval_fn)
ray_sampler = get_ray_sampler(config["training"]["ray_sampler"])
model = BTSDownstreamWrapper(renderer, ray_sampler, config["model"])
# model = BTSWrapper(renderer, config["model"])
# model = BTSWrapper(renderer, config["model"], eval_fns)
return model
|