opensportslib 0.1.3__tar.gz → 0.1.3.dev2__tar.gz

{opensportslib-0.1.3/opensportslib.egg-info → opensportslib-0.1.3.dev2}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: opensportslib
-Version: 0.1.3
+Version: 0.1.3.dev2
 Summary: OpenSportsLib is the professional library, designed for advanced video understanding in sports. It provides state-of-the-art tools for action recognition, spotting, retrieval, and captioning, making it ideal for researchers, analysts, and developers working with sports video data.
 Author: Jeet Vora
 Requires-Python: >=3.12
@@ -27,6 +27,7 @@ Requires-Dist: pytorch-lightning
 Requires-Dist: pandas
 Requires-Dist: pyarrow
 Requires-Dist: huggingface_hub
+Requires-Dist: easydict
 Provides-Extra: test
 Requires-Dist: pytest; extra == "test"
 Requires-Dist: pytest-cov; extra == "test"
@@ -112,6 +113,9 @@ Use it as the main entry point to find:
 - extracted features
 - pretrained models and checkpoints
 
+See the [Model Zoo](docs/model-zoo.md) for available pretrained models,
+reported scores, datasets, and loading snippets.
+
 --
 
 ## Quickstart
@@ -130,7 +134,7 @@ from opensportslib.apis import ClassificationModel
 
 my_model = ClassificationModel(
     config="/path/to/classification.yaml",
-    weights="/path/to/weights.pt",  # optional
+    weights=None,  # optional: path or Hugging Face model ID
 )
 
 my_model.train(
@@ -146,7 +150,7 @@ from opensportslib.apis import ClassificationModel
 
 my_model = ClassificationModel(
     config="/path/to/classification.yaml",
-    weights="/path/to/weights.pt",  # optional
+    weights=None,  # optional: path or Hugging Face model ID
 )
 
 predictions = my_model.infer(
@@ -177,7 +181,7 @@ from opensportslib.apis import LocalizationModel
 
 my_model = LocalizationModel(
     config="/path/to/localization.yaml",
-    weights="/path/to/weights.pt",  # optional
+    weights=None,  # optional: path or Hugging Face model ID
 )
 
 predictions = my_model.infer(

{opensportslib-0.1.3 → opensportslib-0.1.3.dev2}/README.md

@@ -78,6 +78,9 @@ Use it as the main entry point to find:
 - extracted features
 - pretrained models and checkpoints
 
+See the [Model Zoo](docs/model-zoo.md) for available pretrained models,
+reported scores, datasets, and loading snippets.
+
 --
 
 ## Quickstart
@@ -96,7 +99,7 @@ from opensportslib.apis import ClassificationModel
 
 my_model = ClassificationModel(
     config="/path/to/classification.yaml",
-    weights="/path/to/weights.pt",  # optional
+    weights=None,  # optional: path or Hugging Face model ID
 )
 
 my_model.train(
@@ -112,7 +115,7 @@ from opensportslib.apis import ClassificationModel
 
 my_model = ClassificationModel(
     config="/path/to/classification.yaml",
-    weights="/path/to/weights.pt",  # optional
+    weights=None,  # optional: path or Hugging Face model ID
 )
 
 predictions = my_model.infer(
@@ -143,7 +146,7 @@ from opensportslib.apis import LocalizationModel
 
 my_model = LocalizationModel(
     config="/path/to/localization.yaml",
-    weights="/path/to/weights.pt",  # optional
+    weights=None,  # optional: path or Hugging Face model ID
 )
 
 predictions = my_model.infer(

{opensportslib-0.1.3 → opensportslib-0.1.3.dev2}/examples/quickstart/basic_classification.py

@@ -9,7 +9,7 @@ def main():
 
     my_model = ClassificationModel(
         config="examples/configs/classification_video.yaml",
-        weights="/path/to/weights.pt",  # optional
+        weights=None,  # optional: path or Hugging Face model ID
     )
 
     my_model.train(

{opensportslib-0.1.3 → opensportslib-0.1.3.dev2}/examples/quickstart/basic_localization.py

@@ -9,7 +9,7 @@ def main():
 
     my_model = LocalizationModel(
         config="examples/configs/localization.yaml",
-        weights="/path/to/weights.pt",  # optional
+        weights=None,  # optional: path or Hugging Face model ID
     )
 
     my_model.train(

{opensportslib-0.1.3 → opensportslib-0.1.3.dev2}/opensportslib/config/classification.yaml

@@ -1,35 +1,44 @@
 TASK: classification
 DATA:
   dataset_name: mvfouls
-  data_dir: /home/vorajv/opensportslib/SoccerNet/mvfouls
+  data_dir: /home/giancos/datasets/OpenSportsLab/OSL-XFoul/224p
   data_modality: video
   view_type: multi  # multi or single
   num_classes: 8 # mvfoul
   train: 
     type: annotations_train.json
     video_path: ${DATA.data_dir}/train
-    path: ${DATA.train.video_path}/annotations-train.json
+    path: ${DATA.train.video_path}/train.json
     dataloader:
       batch_size: 8
       shuffle: true
       num_workers: 4
       pin_memory: true
+      mp_context: spawn
+      persistent_workers: true
+      prefetch_factor: 4  
   valid:
     type: annotations_valid.json
     video_path: ${DATA.data_dir}/valid
-    path: ${DATA.valid.video_path}/annotations-valid.json
+    path: ${DATA.valid.video_path}/valid.json
     dataloader:
       batch_size: 1
       num_workers: 1
       shuffle: false
+      mp_context: spawn
+      persistent_workers: true
+      prefetch_factor: 4  
   test:
     type: annotations_test.json
     video_path: ${DATA.data_dir}/test
-    path: ${DATA.test.video_path}/annotations-test.json
+    path: ${DATA.test.video_path}/test.json
     dataloader:
       batch_size: 1
-      num_workers: 1
+      num_workers: 0
       shuffle: false
+      mp_context: spawn
+      persistent_workers: true
+      prefetch_factor: 4  
   num_frames: 16               # 8 before + 8 after the foul
   input_fps: 25                # Original FPS of video
   target_fps: 17               # Temporal downsampling to 1s clip (approx)

{opensportslib-0.1.3 → opensportslib-0.1.3.dev2}/opensportslib/config/localization.yaml

@@ -4,7 +4,7 @@ dali: True
 
 DATA:
   dataset_name: SoccerNet
-  data_dir: /home/vorajv/opensportslib/SoccerNet/annotations/
+  data_dir: /home/giancos/datasets/OpenSportsLab/OSL-SNBAS/224p-2024/
   classes:
     - PASS
     - DRIVE
@@ -37,7 +37,7 @@ DATA:
     classes: ${DATA.classes}
     output_map: [data, label]
     video_path: ${DATA.data_dir}/train/
-    path: ${DATA.train.video_path}/annotations-2024-224p-train.json
+    path: ${DATA.train.video_path}/train.json
     dataloader:
       batch_size: 8
       shuffle: true
@@ -49,10 +49,12 @@ DATA:
     classes: ${DATA.classes}
     output_map: [data, label]
     video_path: ${DATA.data_dir}/valid/
-    path: ${DATA.valid.video_path}/annotations-2024-224p-valid.json
+    path: ${DATA.valid.video_path}/valid.json
     dataloader:
       batch_size: 8
       shuffle: true
+      num_workers: 4
+      pin_memory: true
 
   valid_data_frames:
     type: VideoGameWithDaliVideo
@@ -64,13 +66,15 @@ DATA:
     dataloader:
       batch_size: 4
       shuffle: false
+      num_workers: 4
+      pin_memory: true
 
   test:
     type: VideoGameWithDaliVideo
     classes: ${DATA.classes}
     output_map: [data, label]
     video_path: ${DATA.data_dir}/test/
-    path: ${DATA.test.video_path}/annotations-2024-224p-test.json
+    path: ${DATA.test.video_path}/test.json
     results: results_spotting_test
     nms_window: 2 
     metric: tight

{opensportslib-0.1.3 → opensportslib-0.1.3.dev2}/opensportslib/config/sngar-frames.yaml

@@ -8,13 +8,14 @@ TASK: classification
 
 DATA:
   dataset_name: sngar
-  data_dir: /home/spark_user1/opensportslib/sngar-frames
+  data_dir: /home/giancos/datasets/OpenSportsLab/soccernetpro-classification-GAR/frames-parquet
   data_modality: frames_npy
   # max_samples: 100 # only used for quick testing
   num_frames: 16
   frame_size: [224, 224]
   train:
-    path: ${DATA.data_dir}/annotations_train.json
+    video_path: ${DATA.data_dir}/train
+    path: ${DATA.data_dir}/train.json
     dataloader:
       batch_size: 8 # for frozen backbone, use 64
       # for unfrozen backbone, use 32-16-8 depending on the memory available
@@ -22,13 +23,15 @@ DATA:
       num_workers: 8
       pin_memory: true
   valid:
-    path: ${DATA.data_dir}/annotations_valid.json
+    video_path: ${DATA.data_dir}/valid
+    path: ${DATA.data_dir}/valid.json
     dataloader:
       batch_size: 8
       num_workers: 8
       shuffle: false
   test:
-    path: ${DATA.data_dir}/annotations_test.json
+    video_path: ${DATA.data_dir}/test
+    path: ${DATA.data_dir}/test.json
     dataloader:
       batch_size: 8
       num_workers: 8

{opensportslib-0.1.3 → opensportslib-0.1.3.dev2}/opensportslib/config/sngar-tracking.yaml

@@ -9,7 +9,7 @@ TASK: classification
 DATA:
   dataset_name: sngar
   data_modality: tracking_parquet
-  data_dir: /home/karkid/opensportslib/tracking-dataset
+  data_dir: /home/giancos/datasets/OpenSportsLab/soccernetpro-classification-GAR/tracking-parquet
   preload_data: false
   train: 
     type: annotations_train.json
@@ -103,10 +103,10 @@ TRAIN:
     type: CrossEntropyLoss
 
 SYSTEM:
- log_dir: ./logs
- save_dir: ./checkpoints_tracking
- use_seed: true
- seed: 42
- GPU: 4
- device: cuda   # auto | cuda | cpu
- gpu_id: 0
+  log_dir: ./logs
+  save_dir: ./checkpoints_tracking
+  use_seed: true
+  seed: 42
+  GPU: 1
+  device: cuda   # auto | cuda | cpu
+  gpu_id: 0

{opensportslib-0.1.3 → opensportslib-0.1.3.dev2}/opensportslib/core/trainer/classification_trainer.py

@@ -14,6 +14,7 @@ import gc
 import json
 import time
 import logging
+import multiprocessing as mp
 
 import torch
 import tqdm
@@ -865,36 +866,53 @@ class Trainer_Classification:
         else:
             val_sampler = None
 
-        num_train_workers = self.config.DATA.train.dataloader.num_workers
-        num_val_workers = self.config.DATA.valid.dataloader.num_workers
+        train_num_workers = getattr(self.config.DATA.train.dataloader, "num_workers", 0)
+        train_pin_memory = getattr(self.config.DATA.train.dataloader, "pin_memory", self.device.type == "cuda")
+        train_mp_context = getattr(self.config.DATA.train.dataloader, "mp_context", None)
+        train_persistent_workers = getattr(self.config.DATA.train.dataloader, "persistent_workers", train_num_workers > 0)
+        train_prefetch_factor = getattr(self.config.DATA.train.dataloader, "prefetch_factor", 4 if train_num_workers > 0 else None)
+        
+        if train_mp_context is not None:
+            train_mp_context = mp.get_context(train_mp_context)
 
         train_loader = DataLoader(
             train_dataset,
             batch_size=self.config.DATA.train.dataloader.batch_size,
             shuffle=(train_sampler is None and shuffle),
             sampler=train_sampler,
-            num_workers=num_train_workers,
-            pin_memory=True,
+            num_workers=train_num_workers,
+            pin_memory=train_pin_memory,
             collate_fn=collate_fn,
             worker_init_fn=seed_worker,
             generator=g,
             drop_last=True,
-            persistent_workers=num_train_workers > 0,
-            prefetch_factor=4 if num_train_workers > 0 else None,
+            multiprocessing_context=train_mp_context,
+            persistent_workers=train_persistent_workers,
+            prefetch_factor=train_prefetch_factor,
         )
 
+        valid_num_workers = getattr(self.config.DATA.valid.dataloader, "num_workers", 0)
+        valid_pin_memory = getattr(self.config.DATA.valid.dataloader, "pin_memory", self.device.type == "cuda")
+        valid_mp_context = getattr(self.config.DATA.valid.dataloader, "mp_context", None)
+        valid_persistent_workers = getattr(self.config.DATA.valid.dataloader, "persistent_workers", valid_num_workers > 0)
+        valid_prefetch_factor = getattr(self.config.DATA.valid.dataloader, "prefetch_factor", 4 if valid_num_workers > 0 else None)
+
+        if valid_mp_context is not None:
+            valid_mp_context = mp.get_context(valid_mp_context)
+            
         val_loader = DataLoader(
             val_dataset,
             batch_size=self.config.DATA.valid.dataloader.batch_size,
             shuffle=False,
             sampler=val_sampler,   
-            num_workers=num_val_workers,
-            pin_memory=True,
+            num_workers=valid_num_workers,
+            pin_memory=valid_pin_memory,
             collate_fn=collate_fn,
             worker_init_fn=seed_worker,
             generator=g,
-            persistent_workers=num_val_workers > 0,
-            prefetch_factor=4 if num_val_workers > 0 else None,
+            multiprocessing_context=valid_mp_context,
+            persistent_workers=valid_persistent_workers,
+            prefetch_factor=valid_prefetch_factor,
         )
 
         # select the modality-specific trainer.

{opensportslib-0.1.3 → opensportslib-0.1.3.dev2}/opensportslib/datasets/classification_dataset.py

@@ -73,7 +73,7 @@ class ClassificationDataset(Dataset):
         self.config = config
         self.split = split
         self.exclude_labels = ["Unknown", "Dont know"]
-        self.data_dir = config.DATA.data_dir
+        self.video_path = getattr(config.DATA, split).video_path #config.DATA.data_dir
         self.processor = None
 
         # view_type is optional; only MVFoul uses it as of now
@@ -197,7 +197,7 @@ class VideoDataset(ClassificationDataset):
     """
 
     def __init__(self, config, annotations_path, processor, split="train"):
-        super().__init__(config, annotations_path, split)
+        super().__init__(config, annotations_path, processor, split=split)
 
         self.processor = processor
         self.view_type = getattr(config.DATA, "view_type", "single")
@@ -227,15 +227,13 @@ class VideoDataset(ClassificationDataset):
         """read a video file, temporally sub-sample, and apply transforms.
 
         Args:
-            path: realtive path (under data_dir) to the video file.
+            path: realtive path (under video_path) to the video file.
 
         Returns:
             numpy.ndarray of shape (T, H, W, C).
         """
-        full_path = os.path.join(self.config.DATA.data_dir, path)
-
-        if full_path.endswith(".npy"):
-            frames = np.load(full_path).astype(np.float32) / 255.0
+        if path.endswith(".npy"):
+            frames = np.load(os.path.join(self.video_path, path)).astype(np.float32) / 255.0
             if self.transform is not None:
                 frames = self.transform(frames)
             mean = np.array([0.485, 0.456, 0.406], dtype=np.float32)
@@ -243,7 +241,7 @@ class VideoDataset(ClassificationDataset):
             frames = (frames - mean) / std
             return frames
 
-        v = read_video(os.path.join(self.config.DATA.data_dir, path))
+        v = read_video(os.path.join(self.video_path, path))
 
         v = process_frames(
             v,
@@ -473,7 +471,7 @@ class TrackingDataset(ClassificationDataset):
         """read a single parquet tracking clip.
         
         Args:
-            path: Relative path (under ``data_dir``) to the parquet
+            path: Relative path (under ``video_path``) to the parquet
                 file.
 
         Returns:
@@ -481,7 +479,7 @@ class TrackingDataset(ClassificationDataset):
         """
         import pandas as pd
 
-        full_path = os.path.join(self.data_dir, path)
+        full_path = os.path.join(self.video_path, path)
         return pd.read_parquet(full_path)
     
     def __getitem__(self, idx):