revert "chore(core): update launch backend to subprocess (#158)" (#176)

This reverts commit 9ac889 to fix training bug

tools/train.py

@@ -12,7 +12,7 @@
 
 from yolox.core import Trainer, launch
 from yolox.exp import get_exp
-from yolox.utils import configure_nccl, get_num_devices
+from yolox.utils import configure_nccl
 
 
 def make_parser():
@@ -106,8 +106,8 @@ def main(exp, args):
     exp = get_exp(args.exp_file, args.name)
     exp.merge(args.opts)
 
-    num_gpu = get_num_devices() if args.devices is None else args.devices
-    assert num_gpu <= get_num_devices()
+    num_gpu = torch.cuda.device_count() if args.devices is None else args.devices
+    assert num_gpu <= torch.cuda.device_count()
 
     dist_url = "auto" if args.dist_url is None else args.dist_url
     launch(

yolox/core/launch.py

@@ -55,17 +55,15 @@ def launch(
             port = _find_free_port()
             dist_url = f"tcp://127.0.0.1:{port}"
 
-        processes = []
-        for rank in range(num_gpus_per_machine):
-            p = mp.Process(
-                target=_distributed_worker, 
-                args=(
-                    rank, main_func, world_size, num_gpus_per_machine,
-                    machine_rank, backend, dist_url, args))
-            p.start()
-            processes.append(p)
-        for p in processes:
-            p.join()
+        mp.spawn(
+            _distributed_worker,
+            nprocs=num_gpus_per_machine,
+            args=(
+                main_func, world_size, num_gpus_per_machine,
+                machine_rank, backend, dist_url, args
+            ),
+            daemon=False,
+        )
     else:
         main_func(*args)
 

yolox/data/datasets/mosaicdetection.py

@@ -220,4 +220,4 @@ def mixup(self, origin_img, origin_labels, input_dim):
             origin_img = origin_img.astype(np.float32)
             origin_img = 0.5 * origin_img + 0.5 * padded_cropped_img.astype(np.float32)
 
-        return origin_img, origin_labels
+        return origin_img.astype(np.uint8), origin_labels

yolox/utils/boxes.py

@@ -9,7 +9,7 @@
 
 __all__ = [
     "filter_box", "postprocess", "bboxes_iou", "matrix_iou",
-    "adjust_box_anns", "xyxy2xywh", "xyxy2cxcywh",
+    "adjust_box_anns", "xyxy2xywh",
 ]
 
 
@@ -113,11 +113,3 @@ def xyxy2xywh(bboxes):
     bboxes[:, 2] = bboxes[:, 2] - bboxes[:, 0]
     bboxes[:, 3] = bboxes[:, 3] - bboxes[:, 1]
     return bboxes
-
-
-def xyxy2cxcywh(bboxes):
-    bboxes[:, 2] = bboxes[:, 2] - bboxes[:, 0]
-    bboxes[:, 3] = bboxes[:, 3] - bboxes[:, 1]
-    bboxes[:, 0] = bboxes[:, 0] + bboxes[:, 2] * 0.5
-    bboxes[:, 1] = bboxes[:, 1] + bboxes[:, 3] * 0.5
-    return bboxes

yolox/utils/metric.py

@@ -13,19 +13,12 @@
 __all__ = [
     "AverageMeter",
     "MeterBuffer",
-    "get_num_devices",
     "get_total_and_free_memory_in_Mb",
     "occupy_mem",
     "gpu_mem_usage",
 ]
 
 
-def get_num_devices():
-    devices_list_info = os.popen("nvidia-smi -L")
-    devices_list_info = devices_list_info.read().strip().split("\n")
-    return len(devices_list_info)
-
-
 def get_total_and_free_memory_in_Mb(cuda_device):
     devices_info_str = os.popen(
         "nvidia-smi --query-gpu=memory.total,memory.used --format=csv,nounits,noheader"

yolox/utils/setup_env.py

@@ -48,5 +48,3 @@ def configure_module(ulimit_value=8192):
     except Exception:
         # cv2 version mismatch might rasie exceptions.
         pass
-
-    os.environ["OMP_NUM_THREADS"] = str(1)