search

hustvl / SparseTrack

Official PyTorch implementation of SparseTrack (the new version of code will come soon)
MIT License
133 stars 12 forks source link

When using the v8 version for training, I added the pre-training weight file of yolov8l.pt to mot17_train_cfg.py and encountered a problem. #26

Open luke110 opened 8 months ago

luke110 commented 8 months ago

I rewrote the build trainer in mot17_train_config.py

Like this:

train = dict( output_dir="./yolov8_mot17_train-all", init_checkpoint="/HOME/scw6492/run/luke/yolov8/pretrain/yolov8l.pt",

I only change this.

Next I encountered the following problem： [01/19 14:05:56 d2.checkpoint.detection_checkpoint]: [DetectionCheckpointer] Loading from /HOME/scw6492/run/luke/yolov8/pretrain/yolov8l.pt ... [01/19 14:05:56 fvcore.common.checkpoint]: [Checkpointer] Loading from /HOME/scw6492/run/luke/yolov8/pretrain/yolov8l.pt ... Traceback (most recent call last): File "/HOME/scw6492/run/luke/yolov8/train.py", line 181, in launch( File "/data/run01/scw6492/luke/yolov8/detectron2/detectron2/engine/launch.py", line 69, in launch mp.start_processes( File "/HOME/scw6492/.conda/envs/sparse/lib/python3.9/site-packages/torch/multiprocessing/spawn.py", line 197, in start_processes while not context.join(): File "/HOME/scw6492/.conda/envs/sparse/lib/python3.9/site-packages/torch/multiprocessing/spawn.py", line 160, in join raise ProcessRaisedException(msg, error_index, failed_process.pid) torch.multiprocessing.spawn.ProcessRaisedException:

-- Process 0 terminated with the following error: Traceback (most recent call last): File "/HOME/scw6492/.conda/envs/sparse/lib/python3.9/site-packages/torch/multiprocessing/spawn.py", line 69, in _wrap fn(i, args) File "/data/run01/scw6492/luke/yolov8/detectron2/detectron2/engine/launch.py", line 123, in _distributed_worker main_func(args) File "/HOME/scw6492/run/luke/yolov8/train.py", line 176, in main do_train(args, cfg) File "/HOME/scw6492/run/luke/yolov8/train.py", line 144, in do_train checkpointer.resume_or_load(cfg.train.init_checkpoint, resume=args.resume) File "/HOME/scw6492/.conda/envs/sparse/lib/python3.9/site-packages/fvcore/common/checkpoint.py", line 227, in resume_or_load return self.load(path, checkpointables=[]) File "/data/run01/scw6492/luke/yolov8/detectron2/detectron2/checkpoint/detection_checkpoint.py", line 62, in load ret = super().load(path, *args, **kwargs) File "/HOME/scw6492/.conda/envs/sparse/lib/python3.9/site-packages/fvcore/common/checkpoint.py", line 156, in load incompatible = self._load_model(checkpoint) File "/data/run01/scw6492/luke/yolov8/detectron2/detectron2/checkpoint/detection_checkpoint.py", line 126, in _load_model incompatible = super()._load_model(checkpoint) File "/HOME/scw6492/.conda/envs/sparse/lib/python3.9/site-packages/fvcore/common/checkpoint.py", line 273, in _load_model self._convert_ndarray_to_tensor(checkpoint_state_dict) File "/HOME/scw6492/.conda/envs/sparse/lib/python3.9/site-packages/fvcore/common/checkpoint.py", line 366, in _convert_ndarray_to_tensor for k in list(state_dict.keys()): File "/HOME/scw6492/.conda/envs/sparse/lib/python3.9/site-packages/torch/nn/modules/module.py", line 1614, in getattr raise AttributeError("'{}' object has no attribute '{}'".format( AttributeError: 'DetectionModel' object has no attribute 'keys'