valor-lite 0.33.17__py3-none-any.whl → 0.33.19__py3-none-any.whl

valor_lite/semantic_segmentation/computation.py

@@ -31,9 +31,6 @@ def compute_intermediate_confusion_matrices(
         A 2-D confusion matrix with shape (n_labels + 1, n_labels + 1).
     """
 
-    n_gt_labels = groundtruth_labels.size
-    n_pd_labels = prediction_labels.size
-
     groundtruth_counts = groundtruths.sum(axis=1)
     prediction_counts = predictions.sum(axis=1)
 
@@ -42,33 +39,23 @@ def compute_intermediate_confusion_matrices(
     ).sum()
 
     intersection_counts = np.logical_and(
-        groundtruths.reshape(n_gt_labels, 1, -1),
-        predictions.reshape(1, n_pd_labels, -1),
+        groundtruths[:, None, :],
+        predictions[None, :, :],
     ).sum(axis=2)
-
     intersected_groundtruth_counts = intersection_counts.sum(axis=1)
     intersected_prediction_counts = intersection_counts.sum(axis=0)
 
     confusion_matrix = np.zeros((n_labels + 1, n_labels + 1), dtype=np.int32)
     confusion_matrix[0, 0] = background_counts
-    for gidx in range(n_gt_labels):
-        gt_label_idx = groundtruth_labels[gidx]
-        for pidx in range(n_pd_labels):
-            pd_label_idx = prediction_labels[pidx]
-            confusion_matrix[
-                gt_label_idx + 1,
-                pd_label_idx + 1,
-            ] = intersection_counts[gidx, pidx]
-
-            if gidx == 0:
-                confusion_matrix[0, pd_label_idx + 1] = (
-                    prediction_counts[pidx]
-                    - intersected_prediction_counts[pidx]
-                )
-
-        confusion_matrix[gt_label_idx + 1, 0] = (
-            groundtruth_counts[gidx] - intersected_groundtruth_counts[gidx]
-        )
+    confusion_matrix[
+        np.ix_(groundtruth_labels + 1, prediction_labels + 1)
+    ] = intersection_counts
+    confusion_matrix[0, prediction_labels + 1] = (
+        prediction_counts - intersected_prediction_counts
+    )
+    confusion_matrix[groundtruth_labels + 1, 0] = (
+        groundtruth_counts - intersected_groundtruth_counts
+    )
 
     return confusion_matrix
 
@@ -111,9 +98,9 @@ def compute_metrics(
     NDArray[np.float64]
         Confusion matrix containing IOU values.
     NDArray[np.float64]
-        Hallucination ratios.
+        Unmatched prediction ratios.
     NDArray[np.float64]
-        Missing prediction ratios.
+        Unmatched ground truth ratios.
     """
     n_labels = label_metadata.shape[0]
     gt_counts = label_metadata[:, 0]
@@ -121,7 +108,7 @@ def compute_metrics(
 
     counts = data.sum(axis=0)
 
-    # compute iou, missing_predictions and hallucinations
+    # compute iou, unmatched_ground_truth and unmatched predictions
     intersection_ = counts[1:, 1:]
     union_ = (
         gt_counts[:, np.newaxis] + pd_counts[np.newaxis, :] - intersection_
@@ -135,20 +122,20 @@ def compute_metrics(
         out=ious,
     )
 
-    hallucination_ratio = np.zeros((n_labels), dtype=np.float64)
+    unmatched_prediction_ratio = np.zeros((n_labels), dtype=np.float64)
     np.divide(
         counts[0, 1:],
         pd_counts,
         where=pd_counts > 1e-9,
-        out=hallucination_ratio,
+        out=unmatched_prediction_ratio,
     )
 
-    missing_prediction_ratio = np.zeros((n_labels), dtype=np.float64)
+    unmatched_ground_truth_ratio = np.zeros((n_labels), dtype=np.float64)
     np.divide(
         counts[1:, 0],
         gt_counts,
         where=gt_counts > 1e-9,
-        out=missing_prediction_ratio,
+        out=unmatched_ground_truth_ratio,
     )
 
     # compute precision, recall, f1
@@ -181,6 +168,6 @@ def compute_metrics(
         f1_score,
         accuracy,
         ious,
-        hallucination_ratio,
-        missing_prediction_ratio,
+        unmatched_prediction_ratio,
+        unmatched_ground_truth_ratio,
     )

valor_lite/semantic_segmentation/manager.py

@@ -243,6 +243,10 @@ class Evaluator:
         return self.compute_precision_recall_iou(filter_=filter_)
 
 
+def defaultdict_int():
+    return defaultdict(int)
+
+
 class DataLoader:
     """
     Segmentation DataLoader.
@@ -250,8 +254,8 @@ class DataLoader:
 
     def __init__(self):
         self._evaluator = Evaluator()
-        self.groundtruth_count = defaultdict(lambda: defaultdict(int))
-        self.prediction_count = defaultdict(lambda: defaultdict(int))
+        self.groundtruth_count = defaultdict(defaultdict_int)
+        self.prediction_count = defaultdict(defaultdict_int)
         self.matrices = list()
         self.pixel_count = list()
 

valor_lite/semantic_segmentation/metric.py

@@ -209,11 +209,11 @@ class Metric(BaseMetric):
                 dict[str, float],  # iou
             ],
         ],
-        hallucinations: dict[
+        unmatched_predictions: dict[
             str,  # prediction label value
             dict[str, float],  # pixel ratio
         ],
-        missing_predictions: dict[
+        unmatched_ground_truths: dict[
             str,  # ground truth label value
             dict[str, float],  # pixel ratio
         ],
@@ -222,8 +222,8 @@ class Metric(BaseMetric):
         The confusion matrix and related metrics for semantic segmentation tasks.
 
         This class encapsulates detailed information about the model's performance, including correct
-        predictions, misclassifications, hallucinations (false positives), and missing predictions
-        (false negatives). It provides counts for each category to facilitate in-depth analysis.
+        predictions, misclassifications, unmatched_predictions (subset of false positives), and unmatched ground truths
+        (subset of false negatives). It provides counts for each category to facilitate in-depth analysis.
 
         Confusion Matrix Format:
         {
@@ -234,14 +234,14 @@ class Metric(BaseMetric):
             },
         }
 
-        Hallucinations Format:
+        Unmatched Predictions Format:
         {
             <prediction label>: {
                 'iou': <float>,
             },
         }
 
-        Missing Predictions Format:
+        Unmatched Ground Truths Format:
         {
             <ground truth label>: {
                 'iou': <float>,
@@ -253,10 +253,10 @@ class Metric(BaseMetric):
         confusion_matrix : dict
             Nested dictionaries representing the Intersection over Union (IOU) scores for each
             ground truth label and prediction label pair.
-        hallucinations : dict
+        unmatched_predictions : dict
             Dictionary representing the pixel ratios for predicted labels that do not correspond
             to any ground truth labels (false positives).
-        missing_predictions : dict
+        unmatched_ground_truths : dict
             Dictionary representing the pixel ratios for ground truth labels that were not predicted
             (false negatives).
 
@@ -268,8 +268,8 @@ class Metric(BaseMetric):
             type=MetricType.ConfusionMatrix.value,
             value={
                 "confusion_matrix": confusion_matrix,
-                "hallucinations": hallucinations,
-                "missing_predictions": missing_predictions,
+                "unmatched_predictions": unmatched_predictions,
+                "unmatched_ground_truths": unmatched_ground_truths,
             },
             parameters={},
         )

valor_lite/semantic_segmentation/utilities.py

@@ -18,8 +18,8 @@ def unpack_precision_recall_iou_into_metric_lists(
         f1_score,
         accuracy,
         ious,
-        hallucination_ratios,
-        missing_prediction_ratios,
+        unmatched_prediction_ratios,
+        unmatched_ground_truth_ratios,
     ) = results
 
     metrics = defaultdict(list)
@@ -43,16 +43,16 @@ def unpack_precision_recall_iou_into_metric_lists(
                 for gt_label_idx in range(n_labels)
                 if label_metadata[gt_label_idx, 0] > 0
             },
-            hallucinations={
+            unmatched_predictions={
                 index_to_label[pd_label_idx]: {
-                    "ratio": float(hallucination_ratios[pd_label_idx])
+                    "ratio": float(unmatched_prediction_ratios[pd_label_idx])
                 }
                 for pd_label_idx in range(n_labels)
                 if label_metadata[pd_label_idx, 0] > 0
             },
-            missing_predictions={
+            unmatched_ground_truths={
                 index_to_label[gt_label_idx]: {
-                    "ratio": float(missing_prediction_ratios[gt_label_idx])
+                    "ratio": float(unmatched_ground_truth_ratios[gt_label_idx])
                 }
                 for gt_label_idx in range(n_labels)
                 if label_metadata[gt_label_idx, 0] > 0

{valor_lite-0.33.17.dist-info → valor_lite-0.33.19.dist-info}/METADATA

@@ -1,6 +1,6 @@
 Metadata-Version: 2.1
 Name: valor-lite
-Version: 0.33.17
+Version: 0.33.19
 Summary: Compute valor metrics locally.
 License: MIT License
         
@@ -29,22 +29,22 @@ Requires-Python: >=3.10
 Description-Content-Type: text/markdown
 License-File: LICENSE
 Requires-Dist: evaluate
+Requires-Dist: importlib_metadata; python_version < "3.8"
 Requires-Dist: nltk
 Requires-Dist: numpy
-Requires-Dist: Pillow >=9.1.0
+Requires-Dist: Pillow>=9.1.0
 Requires-Dist: requests
-Requires-Dist: rouge-score
+Requires-Dist: rouge_score
 Requires-Dist: shapely
 Requires-Dist: tqdm
-Requires-Dist: importlib-metadata ; python_version < "3.8"
 Provides-Extra: mistral
-Requires-Dist: mistralai >=1.0 ; extra == 'mistral'
+Requires-Dist: mistralai>=1.0; extra == "mistral"
 Provides-Extra: openai
-Requires-Dist: openai ; extra == 'openai'
+Requires-Dist: openai; extra == "openai"
 Provides-Extra: test
-Requires-Dist: pytest ; extra == 'test'
-Requires-Dist: coverage ; extra == 'test'
-Requires-Dist: pre-commit ; extra == 'test'
+Requires-Dist: pytest; extra == "test"
+Requires-Dist: coverage; extra == "test"
+Requires-Dist: pre-commit; extra == "test"
 
 # valor-lite: Fast, local machine learning evaluation.
 

{valor_lite-0.33.17.dist-info → valor_lite-0.33.19.dist-info}/RECORD

@@ -1,24 +1,26 @@
 valor_lite/LICENSE,sha256=M0L53VuwfEEqezhHb7NPeYcO_glw7-k4DMLZQ3eRN64,1068
 valor_lite/__init__.py,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMpJWZG3hSuFU,0
+valor_lite/profiling.py,sha256=TLIROA1qccFw9NoEkMeQcrvvGGO75c4K5yTIWoCUix8,11746
 valor_lite/schemas.py,sha256=pB0MrPx5qFLbwBWDiOUUm-vmXdWvbJLFCBmKgbcbI5g,198
 valor_lite/classification/__init__.py,sha256=8MI8bGwCxYGqRP7KxG7ezhYv4qQ5947XGvvlF8WPM5g,392
 valor_lite/classification/annotation.py,sha256=0aUOvcwBAZgiNOJuyh-pXyNTG7vP7r8CUfnU3OmpUwQ,1113
-valor_lite/classification/computation.py,sha256=qfBhhuDYCiY8h2RdBG3shzgJbHLXDVNujkYFg9xZa6U,12116
+valor_lite/classification/computation.py,sha256=XU-55bzR2JPKDRr8CvCbHHavM5vjBw1TrUX1mOIm_TY,12121
 valor_lite/classification/manager.py,sha256=8GXZECSx4CBbG5NfPrA19BPENqmrjo-wZBmaulWHY20,16676
-valor_lite/classification/metric.py,sha256=fkAo-_3s4EIRSkyn3owBSf4_Gp6lBK9xdToDYMWmT8A,12236
-valor_lite/classification/utilities.py,sha256=PmQar06Vt-ew4Jvnn0IM63mq730QVTsdRtFdVu1HMFU,6885
+valor_lite/classification/metric.py,sha256=_mW3zynmpW8jUIhK2OeX4usdftHgHM9_l7EAbEe2N3w,12288
+valor_lite/classification/utilities.py,sha256=7P3H2wsUR_qmK4WJBFPkViOvhhKVMkCHvDKJbeentMM,6963
 valor_lite/object_detection/__init__.py,sha256=Ql8rju2q7y0Zd9zFvtBJDRhgQFDm1RSYkTsyH3ZE6pA,648
 valor_lite/object_detection/annotation.py,sha256=x9bsl8b75yvkMByXXiIYI9d9T03olDqtykSvKJc3aFw,7729
-valor_lite/object_detection/computation.py,sha256=P5ijxEBuZ3mxYjBQy24TiQpGxRmPuS40Gwn44uv0J7M,28064
-valor_lite/object_detection/manager.py,sha256=I1AwelhxeOA7GJ31eCw8ubNvmprIEGRmbxONozQMsC4,22998
-valor_lite/object_detection/metric.py,sha256=8QhdauuaRrzE39idetkFYTPxA12wrBalQDIR4IUzEbg,24794
-valor_lite/object_detection/utilities.py,sha256=98VSW-g8EYI8Cdd9KHLHdm6F4fI89jaX5I4z99zny4s,16271
-valor_lite/semantic_segmentation/__init__.py,sha256=HQQkr3iBPQfdUrsu0uvx-Uyv9SYmumU1B3slbWOnpNY,245
-valor_lite/semantic_segmentation/annotation.py,sha256=CujYFdHS3fgr4Y7mEDs_u1XBmbPJzNU2CdqvjCT_d_A,2938
-valor_lite/semantic_segmentation/computation.py,sha256=471Pl-0TCFBdkgZMvYDFs4aa6Ak5kv31xarK_USl3pU,5122
-valor_lite/semantic_segmentation/manager.py,sha256=pMepH3zk_fApyFtC9tLrmEYuCbg1n5TLh1J8QRadE44,14287
-valor_lite/semantic_segmentation/metric.py,sha256=aJv3wPEl6USLhZ3c4yz6prnBU-EaG4Kz16f0BXcodd4,7046
-valor_lite/semantic_segmentation/utilities.py,sha256=vZM66YNMz9VJclhuKvcWp74nF65s6bscnnD5U9iDW7Q,2925
+valor_lite/object_detection/computation.py,sha256=70XkgyxAbvZC8FcTKqdAATD5Muebw_jGjoTKPaecwjE,28141
+valor_lite/object_detection/manager.py,sha256=utdILUUCx04EWC0_bHGpEPaxcCOhmsOx5lxT9qU1a9s,23033
+valor_lite/object_detection/metric.py,sha256=npK2sxiwCUTKlRlFym1AlZTvP9herf9lakbsBDwljGM,24901
+valor_lite/object_detection/utilities.py,sha256=F0sfhkledX3h3jXLDpLro7zHbUF9ig7r35J7QbogQVY,16437
+valor_lite/semantic_segmentation/__init__.py,sha256=BhTUbwbdJa1FdS4ZA3QSIZ8TuJmdGGLGCd5hX6SzKa4,297
+valor_lite/semantic_segmentation/annotation.py,sha256=xd2qJyIeTW8CT_Goyu3Kvl_51b9b6D3WvUfqwShR0Sk,4990
+valor_lite/semantic_segmentation/benchmark.py,sha256=iVdxUo9LgDbbXUa6eRhZ49LOYw-yyr2W4p9FP3KHg0k,3848
+valor_lite/semantic_segmentation/computation.py,sha256=l98h8s9RTWQOB_eg2rconqGL1ZbTS4GMtz69vbyEdQ0,4741
+valor_lite/semantic_segmentation/manager.py,sha256=TtwJI7Bsn3zHL2ECOqCmymG-JqREo7I6qxYtycbz54Y,14322
+valor_lite/semantic_segmentation/metric.py,sha256=T9RfPJf4WgqGQTXYvSy08vJG5bjXXJnyYZeW0mlxMa8,7132
+valor_lite/semantic_segmentation/utilities.py,sha256=y9SgArhrk-u_r54SgQxGEKrDjow5KvPG8Zj1GeZefqY,2958
 valor_lite/text_generation/__init__.py,sha256=pGhpWCSZjLM0pPHCtPykAfos55B8ie3mi9EzbNxfj-U,356
 valor_lite/text_generation/annotation.py,sha256=O5aXiwCS4WjA-fqn4ly-O0MsTHoIOmqxqCaAp9IeI3M,1270
 valor_lite/text_generation/computation.py,sha256=cG35qMpxNPEYHXN2fz8wcanESriSHoWMl1idpm9-ous,18638
@@ -31,8 +33,8 @@ valor_lite/text_generation/llm/instructions.py,sha256=fz2onBZZWcl5W8iy7zEWkPGU9N
 valor_lite/text_generation/llm/integrations.py,sha256=-rTfdAjq1zH-4ixwYuMQEOQ80pIFzMTe0BYfroVx3Pg,6974
 valor_lite/text_generation/llm/utilities.py,sha256=bjqatGgtVTcl1PrMwiDKTYPGJXKrBrx7PDtzIblGSys,1178
 valor_lite/text_generation/llm/validators.py,sha256=Wzr5RlfF58_2wOU-uTw7C8skan_fYdhy4Gfn0jSJ8HM,2700
-valor_lite-0.33.17.dist-info/LICENSE,sha256=M0L53VuwfEEqezhHb7NPeYcO_glw7-k4DMLZQ3eRN64,1068
-valor_lite-0.33.17.dist-info/METADATA,sha256=7YPVo6unLWsA7zNV7LcKnwBJm6AqCWjfaeM5nVyVKho,5888
-valor_lite-0.33.17.dist-info/WHEEL,sha256=P9jw-gEje8ByB7_hXoICnHtVCrEwMQh-630tKvQWehc,91
-valor_lite-0.33.17.dist-info/top_level.txt,sha256=9ujykxSwpl2Hu0_R95UQTR_l07k9UUTSdrpiqmq6zc4,11
-valor_lite-0.33.17.dist-info/RECORD,,
+valor_lite-0.33.19.dist-info/LICENSE,sha256=M0L53VuwfEEqezhHb7NPeYcO_glw7-k4DMLZQ3eRN64,1068
+valor_lite-0.33.19.dist-info/METADATA,sha256=lwmxYzLPlKkCZM0mVkepQYYy4mjbg34UEOeeqM7ZE2k,5880
+valor_lite-0.33.19.dist-info/WHEEL,sha256=PZUExdf71Ui_so67QXpySuHtCi3-J3wvF4ORK6k_S8U,91
+valor_lite-0.33.19.dist-info/top_level.txt,sha256=9ujykxSwpl2Hu0_R95UQTR_l07k9UUTSdrpiqmq6zc4,11
+valor_lite-0.33.19.dist-info/RECORD,,

{valor_lite-0.33.17.dist-info → valor_lite-0.33.19.dist-info}/WHEEL

@@ -1,5 +1,5 @@
 Wheel-Version: 1.0
-Generator: setuptools (75.3.0)
+Generator: setuptools (75.6.0)
 Root-Is-Purelib: true
 Tag: py3-none-any