for i in range(opt.genCamera):
cam = next(camSampler)
cameras.append(cam)
splatter.initCameras(cameras=cameras)
writeCameras(scene, os.path.join(opt.output, 'cameras.ply'))
else:
for i in range(len(scene.cameras)):
scene.cameras[i].width = opt.width
scene.cameras[i].height = opt.height
splatter.initCameras(cameras=scene.cameras, genSunMode="triColor")
for pointPath, pointRelPath in zip(points_paths, points_relpaths):
keyName = os.path.join(os.path.join(opt.output, pointRelPath[:-4]))
points = readCloud(pointPath, device="cpu")
scene.loadPoints(points)
fileName = getBasename(pointPath)
splatter.setCloud(scene.cloud)
successful_views= []
rendered = []
for i, cam in enumerate(scene.cameras):
splatter.setCamera(i)
result = splatter.render()
if result is None:
print("splatted a None")
continue
result = result.detach()[0]
rendered.append(result)
print(pointRelPath)
for i, gt in enumerate(rendered):
reg_str = ",".join(["%.3f" % (100*v) for v in trainer.loss_reg])
entries = [trainer.modifier] + [loss_str] + [reg_str]
loss_log.write(",".join(entries)+"\n")
print("{:03d} {}: lr {} loss ({}) \n : reg ({})".format(
t, trainer.modifier, trainer.lr, loss_str, reg_str))
trainer.finish_cycle()
writeScene(scene, os.path.join(expr_dir, 'final_scene.json'),
os.path.join(expr_dir, 'final_cloud.ply'))
if __name__ == "__main__":
opt = FilterOptions().parse()
torch.manual_seed(24)
torch.backends.cudnn.deterministic = True
torch.backends.cudnn.benchmark = False
np.random.seed(24)
# Create ground truth
scene = readScene(opt.source, device="cpu")
if opt.cloud:
points = readCloud(opt.cloud, device="cpu")
points_coords, _, _ = normalize_point_batch(
points[:, :3].unsqueeze(0), NCHW=False)
points[:, :3] = points_coords.squeeze(0)*2
scene.loadPoints(points)
trainImageFilter(scene, benchmark=opt.benchmark)