openmedlab / USFM

189 stars 10 forks source link

权重文件格式和代码设置冲突 #10

Open fly361 opened 2 months ago

fly361 commented 2 months ago

你好,我按照您最新提供的权重:USFMlatest.pth, 按照您的微调命令:python usfm/train.py tag=seg$dataset(已修改) experiment=ftSeg.yaml model.net.backbone.pretrained=assets/USFMpretrained.ckpt data=$dataset data="{batch_size:40, num_workers:4}" trainer="{devices:[0,1], strategy:ddp}" 进行了运行,

但是却发生了如下段的错误,我经过分析,猜测这是因为您的权重文件是.pth格式的文件,而您编写的项目是针对.ckpt格式的。 请问是否是这样的呢? 那么,我们该如何进行修改,能正确运行命令,提取权重文件呢? 还有一个小问题,请问运行您的项目,大概需要多大的显存呢? 感谢您做出的贡献,如能百忙之中解惑,不胜感激!

错误 [2024-04-26 08:23:06,839][usfm.utils.utils][ERROR] - Traceback (most recent call last): File "/home/hyzheng/anaconda3/envs/USFM/lib/python3.9/site-packages/hydra/_internal/instantiate/_instantiate2.py", line 92, in _call_target return target(*args, kwargs) File "/home/hyzheng/Projects/USFM/USFM-master/usfm/models/beitSegLit.py", line 29, in init self.net = MODELS.build(dict(OmegaConf.to_container(net))) File "/home/hyzheng/anaconda3/envs/USFM/lib/python3.9/site-packages/mmengine/registry/registry.py", line 570, in build return self.build_func(cfg, args, kwargs, registry=self) File "/home/hyzheng/anaconda3/envs/USFM/lib/python3.9/site-packages/mmengine/registry/build_functions.py", line 232, in build_model_from_cfg return build_from_cfg(cfg, registry, default_args) File "/home/hyzheng/anaconda3/envs/USFM/lib/python3.9/site-packages/mmengine/registry/build_functions.py", line 121, in build_from_cfg obj = obj_cls(args) # type: ignore File "/home/hyzheng/anaconda3/envs/USFM/lib/python3.9/site-packages/mmseg/models/segmentors/encoder_decoder.py", line 89, in init self.backbone = MODELS.build(backbone) File "/home/hyzheng/anaconda3/envs/USFM/lib/python3.9/site-packages/mmengine/registry/registry.py", line 570, in build return self.build_func(cfg, args, kwargs, registry=self) File "/home/hyzheng/anaconda3/envs/USFM/lib/python3.9/site-packages/mmengine/registry/build_functions.py", line 232, in build_model_from_cfg return build_from_cfg(cfg, registry, default_args) File "/home/hyzheng/anaconda3/envs/USFM/lib/python3.9/site-packages/mmengine/registry/build_functions.py", line 121, in build_from_cfg obj = obj_cls(args) # type: ignore File "/home/hyzheng/Projects/USFM/USFM-master/usfm/models/components/backbone/beit.py", line 480, in init self.init_weights(pretrained) File "/home/hyzheng/Projects/USFM/USFM-master/usfm/models/components/backbone/beit.py", line 518, in init_weights load_checkpoint(self, pretrained, strict=False) File "/home/hyzheng/Projects/USFM/USFM-master/usfm/models/components/backbone/beit.py", line 29, in load_checkpoint checkpoint = torch.load(filename, map_location=map_location) File "/home/hyzheng/anaconda3/envs/USFM/lib/python3.9/site-packages/torch/serialization.py", line 993, in load with _open_zipfile_reader(opened_file) as opened_zipfile: File "/home/hyzheng/anaconda3/envs/USFM/lib/python3.9/site-packages/torch/serialization.py", line 447, in init super().init(torch._C.PyTorchFileReader(name_or_buffer)) RuntimeError: PytorchStreamReader failed reading zip archive: failed finding central directory**