ultralytics 8.3.31 add max_num_obj factor for AutoBatch (#17514)

Co-authored-by: UltralyticsAssistant <web@ultralytics.com>
Co-authored-by: Glenn Jocher <glenn.jocher@ultralytics.com>

ultralytics/__init__.py

@@ -1,6 +1,6 @@
 # Ultralytics YOLO 🚀, AGPL-3.0 license
 
-__version__ = "8.3.30"
+__version__ = "8.3.31"
 
 import os
 

ultralytics/engine/trainer.py

@@ -279,12 +279,7 @@ class BaseTrainer:
 
         # Batch size
         if self.batch_size < 1 and RANK == -1:  # single-GPU only, estimate best batch size
-            self.args.batch = self.batch_size = check_train_batch_size(
+            self.args.batch = self.batch_size = self.auto_batch()
-                model=self.model,
-                imgsz=self.args.imgsz,
-                amp=self.amp,
-                batch=self.batch_size,
-            )
 
         # Dataloaders
         batch_size = self.batch_size // max(world_size, 1)
@@ -478,6 +473,16 @@ class BaseTrainer:
         self._clear_memory()
         self.run_callbacks("teardown")
 
+    def auto_batch(self, max_num_obj=0):
+        """Get batch size by calculating memory occupation of model."""
+        return check_train_batch_size(
+            model=self.model,
+            imgsz=self.args.imgsz,
+            amp=self.amp,
+            batch=self.batch_size,
+            max_num_obj=max_num_obj,
+        )  # returns batch size
+
     def _get_memory(self):
         """Get accelerator memory utilization in GB."""
         if self.device.type == "mps":

ultralytics/models/yolo/detect/train.py

@@ -141,3 +141,10 @@ class DetectionTrainer(BaseTrainer):
         boxes = np.concatenate([lb["bboxes"] for lb in self.train_loader.dataset.labels], 0)
         cls = np.concatenate([lb["cls"] for lb in self.train_loader.dataset.labels], 0)
         plot_labels(boxes, cls.squeeze(), names=self.data["names"], save_dir=self.save_dir, on_plot=self.on_plot)
+
+    def auto_batch(self):
+        """Get batch size by calculating memory occupation of model."""
+        train_dataset = self.build_dataset(self.trainset, mode="train", batch=16)
+        # 4 for mosaic augmentation
+        max_num_obj = max(len(l["cls"]) for l in train_dataset.labels) * 4
+        return super().auto_batch(max_num_obj)

ultralytics/utils/autobatch.py

@@ -11,7 +11,7 @@ from ultralytics.utils import DEFAULT_CFG, LOGGER, colorstr
 from ultralytics.utils.torch_utils import autocast, profile
 
 
-def check_train_batch_size(model, imgsz=640, amp=True, batch=-1):
+def check_train_batch_size(model, imgsz=640, amp=True, batch=-1, max_num_obj=1):
     """
     Compute optimal YOLO training batch size using the autobatch() function.
 
@@ -20,6 +20,7 @@ def check_train_batch_size(model, imgsz=640, amp=True, batch=-1):
         imgsz (int, optional): Image size used for training.
         amp (bool, optional): Use automatic mixed precision if True.
         batch (float, optional): Fraction of GPU memory to use. If -1, use default.
+        max_num_obj (int, optional): The maximum number of objects from dataset.
 
     Returns:
         (int): Optimal batch size computed using the autobatch() function.
@@ -29,10 +30,12 @@ def check_train_batch_size(model, imgsz=640, amp=True, batch=-1):
         Otherwise, a default fraction of 0.6 is used.
     """
     with autocast(enabled=amp):
-        return autobatch(deepcopy(model).train(), imgsz, fraction=batch if 0.0 < batch < 1.0 else 0.6)
+        return autobatch(
+            deepcopy(model).train(), imgsz, fraction=batch if 0.0 < batch < 1.0 else 0.6, max_num_obj=max_num_obj
+        )
 
 
-def autobatch(model, imgsz=640, fraction=0.60, batch_size=DEFAULT_CFG.batch):
+def autobatch(model, imgsz=640, fraction=0.60, batch_size=DEFAULT_CFG.batch, max_num_obj=1):
     """
     Automatically estimate the best YOLO batch size to use a fraction of the available CUDA memory.
 
@@ -41,6 +44,7 @@ def autobatch(model, imgsz=640, fraction=0.60, batch_size=DEFAULT_CFG.batch):
         imgsz (int, optional): The image size used as input for the YOLO model. Defaults to 640.
         fraction (float, optional): The fraction of available CUDA memory to use. Defaults to 0.60.
         batch_size (int, optional): The default batch size to use if an error is detected. Defaults to 16.
+        max_num_obj (int, optional): The maximum number of objects from dataset.
 
     Returns:
         (int): The optimal batch size.
@@ -70,7 +74,7 @@ def autobatch(model, imgsz=640, fraction=0.60, batch_size=DEFAULT_CFG.batch):
     batch_sizes = [1, 2, 4, 8, 16] if t < 16 else [1, 2, 4, 8, 16, 32, 64]
     try:
         img = [torch.empty(b, 3, imgsz, imgsz) for b in batch_sizes]
-        results = profile(img, model, n=1, device=device)
+        results = profile(img, model, n=1, device=device, max_num_obj=max_num_obj)
 
         # Fit a solution
         y = [x[2] for x in results if x]  # memory [2]

ultralytics/utils/tal.py

@@ -3,6 +3,7 @@
 import torch
 import torch.nn as nn
 
+from . import LOGGER
 from .checks import check_version
 from .metrics import bbox_iou, probiou
 from .ops import xywhr2xyxyxyxy
@@ -73,6 +74,7 @@ class TaskAlignedAssigner(nn.Module):
             return self._forward(pd_scores, pd_bboxes, anc_points, gt_labels, gt_bboxes, mask_gt)
         except torch.OutOfMemoryError:
             # Move tensors to CPU, compute, then move back to original device
+            LOGGER.warning("WARNING: CUDA OutOfMemoryError in TaskAlignedAssigner, using CPU")
             cpu_tensors = [t.cpu() for t in (pd_scores, pd_bboxes, anc_points, gt_labels, gt_bboxes, mask_gt)]
             result = self._forward(*cpu_tensors)
             return tuple(t.to(device) for t in result)

ultralytics/utils/torch_utils.py

@@ -623,7 +623,7 @@ def convert_optimizer_state_dict_to_fp16(state_dict):
     return state_dict
 
 
-def profile(input, ops, n=10, device=None):
+def profile(input, ops, n=10, device=None, max_num_obj=0):
     """
     Ultralytics speed, memory and FLOPs profiler.
 
@@ -671,6 +671,14 @@ def profile(input, ops, n=10, device=None):
                         t[2] = float("nan")
                     tf += (t[1] - t[0]) * 1000 / n  # ms per op forward
                     tb += (t[2] - t[1]) * 1000 / n  # ms per op backward
+                    if max_num_obj:  # simulate training with predictions per image grid (for AutoBatch)
+                        torch.randn(
+                            x.shape[0],
+                            max_num_obj,
+                            int(sum([(x.shape[-1] / s) * (x.shape[-2] / s) for s in m.stride.tolist()])),
+                            device=device,
+                            dtype=torch.float32,
+                        )
                 mem = torch.cuda.memory_reserved() / 1e9 if torch.cuda.is_available() else 0  # (GB)
                 s_in, s_out = (tuple(x.shape) if isinstance(x, torch.Tensor) else "list" for x in (x, y))  # shapes
                 p = sum(x.numel() for x in m.parameters()) if isinstance(m, nn.Module) else 0  # parameters