reduced-3dgs 1.8.19__cp311-cp311-win_amd64.whl → 1.9.1__cp311-cp311-win_amd64.whl

reduced_3dgs/importance/combinations.py

@@ -10,9 +10,17 @@ def BaseImportancePrunerInDensifyTrainer(
         scene_extent: float,
         dataset: List[Camera],
         *args,
-        importance_prune_from_iter=1000,
-        importance_prune_until_iter=15000,
-        importance_prune_interval=100,
+        importance_prune_from_iter=15000,
+        importance_prune_until_iter=20000,
+        importance_prune_interval: int = 1000,
+        importance_prune_type="comprehensive",
+        importance_prune_percent=0.1,
+        importance_prune_thr_important_score=None,
+        importance_prune_thr_v_important_score=1.0,
+        importance_prune_thr_max_v_important_score=None,
+        importance_prune_thr_count=1,
+        importance_prune_thr_T_alpha=0.01,
+        importance_v_pow=0.1,
         **kwargs):
     return DensificationTrainerWrapper(
         lambda model, scene_extent: ImportancePruner(
@@ -21,6 +29,14 @@ def BaseImportancePrunerInDensifyTrainer(
             importance_prune_from_iter=importance_prune_from_iter,
             importance_prune_until_iter=importance_prune_until_iter,
             importance_prune_interval=importance_prune_interval,
+            importance_prune_type=importance_prune_type,
+            importance_prune_percent=importance_prune_percent,
+            importance_prune_thr_important_score=importance_prune_thr_important_score,
+            importance_prune_thr_v_important_score=importance_prune_thr_v_important_score,
+            importance_prune_thr_max_v_important_score=importance_prune_thr_max_v_important_score,
+            importance_prune_thr_count=importance_prune_thr_count,
+            importance_prune_thr_T_alpha=importance_prune_thr_T_alpha,
+            importance_v_pow=importance_v_pow,
         ),
         model,
         scene_extent,

reduced_3dgs/importance/trainer.py

@@ -75,7 +75,7 @@ def count_render(self: GaussianModel, viewpoint_camera: Camera):
     }
 
 
-def prune_gaussians(model: GaussianModel, dataset: CameraDataset):
+def prune_list(model: GaussianModel, dataset: CameraDataset):
     gaussian_count = torch.zeros(model.get_xyz.shape[0], device=model.get_xyz.device, dtype=torch.int)
     opacity_important_score = torch.zeros(model.get_xyz.shape[0], device=model.get_xyz.device, dtype=torch.float)
     T_alpha_important_score = torch.zeros(model.get_xyz.shape[0], device=model.get_xyz.device, dtype=torch.float)
@@ -84,7 +84,80 @@ def prune_gaussians(model: GaussianModel, dataset: CameraDataset):
         gaussian_count += out["gaussians_count"]
         opacity_important_score += out["opacity_important_score"]
         T_alpha_important_score += out["T_alpha_important_score"]
-    return None
+    return gaussian_count, opacity_important_score, T_alpha_important_score
+
+
+# return importance score with adaptive volume measure described in paper
+def calculate_v_imp_score(gaussians: GaussianModel, imp_list, v_pow):
+    """
+    :param gaussians: A data structure containing Gaussian components with a get_scaling method.
+    :param imp_list: The importance scores for each Gaussian component.
+    :param v_pow: The power to which the volume ratios are raised.
+    :return: A list of adjusted values (v_list) used for pruning.
+    """
+    # Calculate the volume of each Gaussian component
+    volume = torch.prod(gaussians.get_scaling, dim=1)
+    # Determine the kth_percent_largest value
+    index = int(len(volume) * 0.9)
+    sorted_volume, _ = torch.sort(volume, descending=True)
+    kth_percent_largest = sorted_volume[index]
+    # Calculate v_list
+    v_list = torch.pow(volume / kth_percent_largest, v_pow)
+    v_list = v_list * imp_list
+    return v_list
+
+
+def score2mask(percent, import_score: list, threshold=None):
+    sorted_tensor, _ = torch.sort(import_score, dim=0)
+    index_nth_percentile = int(percent * (sorted_tensor.shape[0] - 1))
+    value_nth_percentile = sorted_tensor[index_nth_percentile]
+    thr = min(threshold, value_nth_percentile) if threshold is not None else value_nth_percentile
+    prune_mask = (import_score <= thr)
+    return prune_mask
+
+
+def prune_gaussians(
+        gaussians: GaussianModel, dataset: CameraDataset,
+        prune_type="comprehensive",
+        prune_percent=0.1,
+        prune_thr_important_score=None,
+        prune_thr_v_important_score=1.0,
+        prune_thr_max_v_important_score=None,
+        prune_thr_count=1,
+        prune_thr_T_alpha=0.01,
+        v_pow=0.1):
+    gaussian_list, opacity_imp_list, T_alpha_imp_list = prune_list(gaussians, dataset)
+    match prune_type:
+        case "important_score":
+            mask = score2mask(prune_percent, opacity_imp_list, prune_thr_important_score)
+        case "v_important_score":
+            v_list = calculate_v_imp_score(gaussians, opacity_imp_list, v_pow)
+            mask = score2mask(prune_percent, v_list, prune_thr_v_important_score)
+        case "max_v_important_score":
+            v_list = opacity_imp_list * torch.max(gaussians.get_scaling, dim=1)[0]
+            mask = score2mask(prune_percent, v_list, prune_thr_max_v_important_score)
+        case "count":
+            mask = score2mask(prune_percent, gaussian_list, prune_thr_count)
+        case "T_alpha":
+            # new importance score defined by doji
+            mask = score2mask(prune_percent, T_alpha_imp_list, prune_thr_T_alpha)
+        case "comprehensive":
+            mask = torch.zeros_like(gaussian_list, dtype=torch.bool)
+            if prune_thr_important_score is not None:
+                mask |= score2mask(prune_percent, opacity_imp_list, prune_thr_important_score)
+            if prune_thr_v_important_score is not None:
+                v_list = calculate_v_imp_score(gaussians, opacity_imp_list, v_pow)
+                mask |= score2mask(prune_percent, v_list, prune_thr_v_important_score)
+            if prune_thr_max_v_important_score is not None:
+                v_list = opacity_imp_list * torch.max(gaussians.get_scaling, dim=1)[0]
+                mask |= score2mask(prune_percent, v_list, prune_thr_max_v_important_score)
+            if prune_thr_count is not None:
+                mask |= score2mask(prune_percent, gaussian_list, prune_thr_count)
+            if prune_thr_T_alpha is not None:
+                mask |= score2mask(prune_percent, T_alpha_imp_list, prune_thr_T_alpha)
+        case _:
+            raise Exception("Unsupportive prunning method")
+    return mask
 
 
 class ImportancePruner(DensifierWrapper):
@@ -94,17 +167,39 @@ class ImportancePruner(DensifierWrapper):
             importance_prune_from_iter=15000,
             importance_prune_until_iter=20000,
             importance_prune_interval: int = 1000,
+            importance_prune_type="comprehensive",
+            importance_prune_percent=0.1,
+            importance_prune_thr_important_score=None,
+            importance_prune_thr_v_important_score=1.0,
+            importance_prune_thr_max_v_important_score=None,
+            importance_prune_thr_count=1,
+            importance_prune_thr_T_alpha=0.01,
+            importance_v_pow=0.1
     ):
         super().__init__(base_densifier)
         self.dataset = dataset
         self.importance_prune_from_iter = importance_prune_from_iter
         self.importance_prune_until_iter = importance_prune_until_iter
         self.importance_prune_interval = importance_prune_interval
+        self.prune_percent = importance_prune_percent
+        self.prune_thr_important_score = importance_prune_thr_important_score
+        self.prune_thr_v_important_score = importance_prune_thr_v_important_score
+        self.prune_thr_max_v_important_score = importance_prune_thr_max_v_important_score
+        self.prune_thr_count = importance_prune_thr_count
+        self.prune_thr_T_alpha = importance_prune_thr_T_alpha
+        self.v_pow = importance_v_pow
+        self.prune_type = importance_prune_type
 
     def densify_and_prune(self, loss, out, camera, step: int):
         ret = super().densify_and_prune(loss, out, camera, step)
         if self.importance_prune_from_iter <= step <= self.importance_prune_until_iter and step % self.importance_prune_interval == 0:
-            remove_mask = prune_gaussians(self.model, self.dataset)
+            remove_mask = prune_gaussians(
+                self.model, self.dataset,
+                self.prune_type, self.prune_percent,
+                self.prune_thr_important_score, self.prune_thr_v_important_score,
+                self.prune_thr_max_v_important_score, self.prune_thr_count,
+                self.prune_thr_T_alpha, self.v_pow,
+            )
             ret = ret._replace(remove_mask=remove_mask if ret.remove_mask is None else torch.logical_or(ret.remove_mask, remove_mask))
         return ret
 
@@ -114,9 +209,17 @@ def BaseImportancePruningTrainer(
         scene_extent: float,
         dataset: List[Camera],
         *args,
-        importance_prune_from_iter=1000,
-        importance_prune_until_iter=15000,
-        importance_prune_interval: int = 100,
+        importance_prune_from_iter=15000,
+        importance_prune_until_iter=20000,
+        importance_prune_interval: int = 1000,
+        importance_prune_type="comprehensive",
+        importance_prune_percent=0.1,
+        importance_prune_thr_important_score=None,
+        importance_prune_thr_v_important_score=1.0,
+        importance_prune_thr_max_v_important_score=None,
+        importance_prune_thr_count=1,
+        importance_prune_thr_T_alpha=0.01,
+        importance_v_pow=0.1,
         **kwargs):
     return DensificationTrainer(
         model, scene_extent,
@@ -126,5 +229,13 @@ def BaseImportancePruningTrainer(
             importance_prune_from_iter=importance_prune_from_iter,
             importance_prune_until_iter=importance_prune_until_iter,
             importance_prune_interval=importance_prune_interval,
+            importance_prune_type=importance_prune_type,
+            importance_prune_percent=importance_prune_percent,
+            importance_prune_thr_important_score=importance_prune_thr_important_score,
+            importance_prune_thr_v_important_score=importance_prune_thr_v_important_score,
+            importance_prune_thr_max_v_important_score=importance_prune_thr_max_v_important_score,
+            importance_prune_thr_count=importance_prune_thr_count,
+            importance_prune_thr_T_alpha=importance_prune_thr_T_alpha,
+            importance_v_pow=importance_v_pow,
         ), *args, **kwargs
     )

{reduced_3dgs-1.8.19.dist-info → reduced_3dgs-1.9.1.dist-info}/METADATA

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: reduced_3dgs
-Version: 1.8.19
+Version: 1.9.1
 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-1.8.19.dist-info → reduced_3dgs-1.9.1.dist-info}/RECORD

@@ -2,12 +2,12 @@ reduced_3dgs/__init__.py,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMpJWZG3hSuFU,0
 reduced_3dgs/combinations.py,sha256=FrZLxd3AZlHPSq_WJeEdGWH5zh40rAuV5txxr8HsSPY,7031
 reduced_3dgs/quantize.py,sha256=Y44qHyFdOIqke7NoeqXmyKloS43j-al74ZiNsuZZHbM,2527
 reduced_3dgs/train.py,sha256=jXHdXk05o_ebHjx_VBzcY6fRNn9EdKve6Tf5YC5an0o,9803
-reduced_3dgs/diff_gaussian_rasterization/_C.cp311-win_amd64.pyd,sha256=gmzHvGQ0dckkAtNQtEb47B8q0dqqDkVffANtRPUbU_0,1632256
+reduced_3dgs/diff_gaussian_rasterization/_C.cp311-win_amd64.pyd,sha256=RBnGVfTJh1P_C6CDui3eMP84PgFKjRxZZ-YtUE0JMik,1632256
 reduced_3dgs/diff_gaussian_rasterization/__init__.py,sha256=oV6JjTc-50MscX4XHeIWSgLr3l8Y25knBIs-0gRbJr4,7932
 reduced_3dgs/importance/__init__.py,sha256=neJsbY5cLikEGBQGdR4MjwCQ5VWVikT1357DwL0EtWU,289
-reduced_3dgs/importance/combinations.py,sha256=Hxl0Syx2PD3l2bDzJ29_b2YdfC3etm3BH7gbteujGSU,1730
-reduced_3dgs/importance/trainer.py,sha256=AlkQR-dWD5A43ywLgdV2okSwB8Sg9b1u8Zi99u8Gd4A,5250
-reduced_3dgs/importance/diff_gaussian_rasterization/_C.cp311-win_amd64.pyd,sha256=ygopQDBcQr4DoGoA7nf6hr0GXXtTwUja2nDo0f00IDg,1316352
+reduced_3dgs/importance/combinations.py,sha256=Q2WqwXNuclPWBsw15aR14xQ72JewVeZo9igMeaSfuf8,2693
+reduced_3dgs/importance/trainer.py,sha256=ozzjPSXIpHNP4H1mkXlyAv5KJXWt1k4NKCikx9E9S1E,11151
+reduced_3dgs/importance/diff_gaussian_rasterization/_C.cp311-win_amd64.pyd,sha256=1bWY2QSSugSLztOSu9NZjZL5wxRDdHHGeJ7_zBos8qU,1315840
 reduced_3dgs/importance/diff_gaussian_rasterization/__init__.py,sha256=Tix8auyXBb_QFQtXrV3sLE9kdnl5zgHH0BbqcFzDp84,12850
 reduced_3dgs/pruning/__init__.py,sha256=E_YxJ9cDV_B6EJbYUBEcuRYMIht_C72rI1VJUXFCLpM,201
 reduced_3dgs/pruning/combinations.py,sha256=UivTfbSMmaWYVi9E4OF-_AZA-WBWniMiX-wKUftezF8,2331
@@ -20,9 +20,9 @@ reduced_3dgs/quantization/wrapper.py,sha256=cyXqfJgo9b3fS7DYXxOk5LmQudvrEhweOebF
 reduced_3dgs/shculling/__init__.py,sha256=nP2BejDCUdCmJNRbg0hfhHREO6jyZXwIcRiw6ttVgqo,149
 reduced_3dgs/shculling/gaussian_model.py,sha256=f8QWaL09vaV9Tcf6Dngjg_Fmk1wTQPAjWhuhI_N02Y8,2877
 reduced_3dgs/shculling/trainer.py,sha256=9hwR77djhZpyf-URhwKHjnLbe0ZAOS-DIw58RzkcHXQ,6369
-reduced_3dgs/simple_knn/_C.cp311-win_amd64.pyd,sha256=8Wj2LMNDF2ge-kdpiM0bdS0VQ6eS_kqGwetWnfJdroE,1263616
-reduced_3dgs-1.8.19.dist-info/licenses/LICENSE.md,sha256=LQ4_LAqlncGkg_mQy5ykMAFtQDSPB0eKmIEtBut0yjw,4916
-reduced_3dgs-1.8.19.dist-info/METADATA,sha256=j7c3qbf3ewfPBjNgY_toNbiEYmTpTdGZ4g4zoiipjUk,13015
-reduced_3dgs-1.8.19.dist-info/WHEEL,sha256=ZhKniLpVlRfVU2a9ty7N0kTvsYDNee6K-jhM1zk8cMw,101
-reduced_3dgs-1.8.19.dist-info/top_level.txt,sha256=PpU5aT3-baSCdqCtTaZknoB32H93UeKCkYDkRCCZMEI,13
-reduced_3dgs-1.8.19.dist-info/RECORD,,
+reduced_3dgs/simple_knn/_C.cp311-win_amd64.pyd,sha256=d56351BP_34vH4GnQmrZqXxDk9lpa6sw8ocuQYhyH0o,1263616
+reduced_3dgs-1.9.1.dist-info/licenses/LICENSE.md,sha256=LQ4_LAqlncGkg_mQy5ykMAFtQDSPB0eKmIEtBut0yjw,4916
+reduced_3dgs-1.9.1.dist-info/METADATA,sha256=TIrUop8Y3QjsjnVAxNEGADwBMtVArYSbYhhzMYLWMYc,13014
+reduced_3dgs-1.9.1.dist-info/WHEEL,sha256=ZhKniLpVlRfVU2a9ty7N0kTvsYDNee6K-jhM1zk8cMw,101
+reduced_3dgs-1.9.1.dist-info/top_level.txt,sha256=PpU5aT3-baSCdqCtTaZknoB32H93UeKCkYDkRCCZMEI,13
+reduced_3dgs-1.9.1.dist-info/RECORD,,