reduced-3dgs 1.9.5__tar.gz → 1.10.1__tar.gz

{reduced_3dgs-1.9.5 → reduced_3dgs-1.10.1}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: reduced_3dgs
-Version: 1.9.5
+Version: 1.10.1
 Summary: Refactored code for the paper "Reducing the Memory Footprint of 3D Gaussian Splatting"
 Home-page: https://github.com/yindaheng98/reduced-3dgs
 Author: yindaheng98

{reduced_3dgs-1.9.5 → reduced_3dgs-1.10.1}/reduced_3dgs/quantization/quantizer.py

@@ -119,10 +119,12 @@ class VectorQuantizer(AbstractQuantizer):
         return self.one_nearst(model._opacity.detach(), codebook)
 
     def produce_clusters_scaling(self, model: GaussianModel, *args, **kwargs):
-        return self.generate_codebook(model.get_scaling.detach(), self.num_clusters_scaling, *args, **kwargs)
+        centers, ids = self.generate_codebook(model.get_scaling.detach(), self.num_clusters_scaling, *args, **kwargs)
+        centers_log = model.scaling_inverse_activation(centers)
+        return centers_log, ids
 
     def find_nearest_cluster_id_scaling(self, model: GaussianModel, codebook: torch.Tensor):
-        return self.one_nearst(model.get_scaling.detach(), codebook)
+        return self.one_nearst(model.get_scaling.detach(), model.scaling_activation(codebook))
 
     def produce_clusters(self, model: GaussianModel, init_codebook_dict={}):
         codebook_dict: Dict[str, torch.Tensor] = {}
@@ -163,7 +165,7 @@ class VectorQuantizer(AbstractQuantizer):
 
     def dequantize(self, model: GaussianModel, ids_dict: Dict[str, torch.Tensor], codebook_dict: Dict[str, torch.Tensor], xyz: torch.Tensor = None, replace=False) -> GaussianModel:
         opacity = codebook_dict["opacity"][ids_dict["opacity"], ...]
-        scaling = model.scaling_inverse_activation(codebook_dict["scaling"][ids_dict["scaling"], ...])
+        scaling = codebook_dict["scaling"][ids_dict["scaling"], ...]
 
         rotation = torch.cat((
             codebook_dict["rotation_re"][ids_dict["rotation_re"], ...],

{reduced_3dgs-1.9.5 → reduced_3dgs-1.10.1}/reduced_3dgs/train.py

@@ -4,13 +4,13 @@ import shutil
 from typing import List, Tuple
 import torch
 from tqdm import tqdm
-from argparse import Namespace
 from gaussian_splatting import GaussianModel
-from gaussian_splatting.dataset import CameraDataset, JSONCameraDataset, TrainableCameraDataset
+from gaussian_splatting.dataset import CameraDataset
 from gaussian_splatting.utils import psnr
-from gaussian_splatting.dataset.colmap import ColmapCameraDataset, ColmapTrainableCameraDataset, colmap_init
+from gaussian_splatting.dataset.colmap import colmap_init
 from gaussian_splatting.trainer import AbstractTrainer
 from gaussian_splatting.trainer.extensions import ScaleRegularizeTrainerWrapper
+from gaussian_splatting.train import prepare_dataset, save_cfg_args
 from reduced_3dgs.quantization import AbstractQuantizer, VectorQuantizeTrainerWrapper
 from reduced_3dgs.shculling import VariableSHGaussianModel, SHCullingTrainer
 from reduced_3dgs.pruning import PruningTrainer
@@ -81,17 +81,20 @@ def prepare_quantizer(
     return trainer, trainer.quantizer
 
 
-def prepare_training(sh_degree: int, source: str, device: str, mode: str, load_ply: str = None, load_camera: str = None, load_depth=False, with_scale_reg=False, quantize: bool = False, load_quantized: str = None, configs={}) -> Tuple[CameraDataset, GaussianModel, AbstractTrainer]:
-    quantizer = None
-    if mode in basemodes:
+def prepare_gaussians(sh_degree: int, source: str, device: str, trainable_camera: bool = False, load_ply: str = None) -> Tuple[CameraDataset, GaussianModel, AbstractTrainer]:
+    if trainable_camera:
+        gaussians = CameraTrainableVariableSHGaussianModel(sh_degree).to(device)
+        gaussians.load_ply(load_ply) if load_ply else colmap_init(gaussians, source)
+    else:
         gaussians = VariableSHGaussianModel(sh_degree).to(device)
         gaussians.load_ply(load_ply) if load_ply else colmap_init(gaussians, source)
-        dataset = (JSONCameraDataset(load_camera, load_depth=load_depth) if load_camera else ColmapCameraDataset(source, load_depth=load_depth)).to(device)
+    return gaussians
+
+
+def prepare_trainer(gaussians: GaussianModel, dataset: CameraDataset, mode: str, with_scale_reg=False, quantize: bool = False, load_quantized: str = None, configs={}) -> AbstractTrainer:
+    if mode in basemodes:
         modes = basemodes
     elif mode in cameramodes:
-        gaussians = CameraTrainableVariableSHGaussianModel(sh_degree).to(device)
-        gaussians.load_ply(load_ply) if load_ply else colmap_init(gaussians, source)
-        dataset = (TrainableCameraDataset.from_json(load_camera, load_depth=load_depth) if load_camera else ColmapTrainableCameraDataset(source, load_depth=load_depth)).to(device)
         modes = cameramodes
     else:
         raise ValueError(f"Unknown mode: {mode}")
@@ -114,13 +117,15 @@ def prepare_training(sh_degree: int, source: str, device: str, mode: str, load_p
             dataset=dataset,
             **configs
         )
-    return dataset, gaussians, trainer, quantizer
+        quantizer = None
+    return trainer, quantizer
 
 
-def save_cfg_args(destination: str, sh_degree: int, source: str):
-    os.makedirs(destination, exist_ok=True)
-    with open(os.path.join(destination, "cfg_args"), 'w') as cfg_log_f:
-        cfg_log_f.write(str(Namespace(sh_degree=sh_degree, source_path=source)))
+def prepare_training(sh_degree: int, source: str, device: str, mode: str, load_ply: str = None, load_camera: str = None, load_depth=False, with_scale_reg=False, quantize: bool = False, load_quantized: str = None, configs={}) -> Tuple[CameraDataset, GaussianModel, AbstractTrainer]:
+    dataset = prepare_dataset(source=source, device=device, trainable_camera=mode in cameramodes, load_camera=load_camera, load_depth=load_depth)
+    gaussians = prepare_gaussians(sh_degree=sh_degree, source=source, device=device, trainable_camera=mode in cameramodes, load_ply=load_ply)
+    trainer, quantizer = prepare_trainer(gaussians=gaussians, dataset=dataset, mode=mode, with_scale_reg=with_scale_reg, quantize=quantize, load_quantized=load_quantized, configs=configs)
+    return dataset, gaussians, trainer, quantizer
 
 
 def training(dataset: CameraDataset, gaussians: GaussianModel, trainer: AbstractTrainer, quantizer: AbstractQuantizer, destination: str, iteration: int, save_iterations: List[int], device: str, empty_cache_every_step=False):

{reduced_3dgs-1.9.5 → reduced_3dgs-1.10.1}/reduced_3dgs.egg-info/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: reduced_3dgs
-Version: 1.9.5
+Version: 1.10.1
 Summary: Refactored code for the paper "Reducing the Memory Footprint of 3D Gaussian Splatting"
 Home-page: https://github.com/yindaheng98/reduced-3dgs
 Author: yindaheng98

{reduced_3dgs-1.9.5 → reduced_3dgs-1.10.1}/setup.py

@@ -60,7 +60,7 @@ if os.name == 'nt':
 
 setup(
     name="reduced_3dgs",
-    version='1.9.5',
+    version='1.10.1',
     author='yindaheng98',
     author_email='yindaheng98@gmail.com',
     url='https://github.com/yindaheng98/reduced-3dgs',