Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Jump to bottom

AttributeError: BaseRecord has no attribute segmentation #1081

Closed
DrAle85 opened this issue Mar 30, 2022 · 1 comment
Closed

AttributeError: BaseRecord has no attribute segmentation #1081

DrAle85 opened this issue Mar 30, 2022 · 1 comment
Labels
bug Something isn't working

Comments

@DrAle85
Copy link

DrAle85 commented Mar 30, 2022

🐛 Bug

When trying to use instance segmentation with unet, the learn.lr_find() throws this error.
AttributeError: BaseRecord has no attribute segmentation

To Reproduce
Steps to reproduce the behavior:

  1. Segmented data are exported in COCO Format from CVAT.
    The code below was adapted from the UNET Tutorial:
# create a parser   
parser=parsers.COCOMaskParser(annotations_filepath=annotation_file, img_dir=images_dir)   
class_map=parser.class_map

Output is 100%

# Transforms
# size is set to 384 because EfficientDet requires its inputs to be divisible by 128
image_size = 256
train_tfms = tfms.A.Adapter([*tfms.A.aug_tfms(size=image_size, presize=256), tfms.A.Normalize()])
valid_tfms = tfms.A.Adapter([*tfms.A.resize_and_pad(image_size), tfms.A.Normalize()])

# Datasets
train_ds = Dataset(train_records, train_tfms)
valid_ds = Dataset(valid_records, valid_tfms)

# Show an element of the train_ds with augmentation transformations applied
samples = [train_ds[0] for _ in range(3)]
show_samples(samples=samples,display_label=False, ncols=3, figsize=(25,10))

Output as expected!

model_type=models.fastai.unet
train_dl = model_type.train_dl(train_ds, batch_size=8, num_workers=0, shuffle=True)
valid_dl = model_type.valid_dl(valid_ds, batch_size=8, num_workers=0, shuffle=False)

backbone = model_type.backbones.resnet34()
model = model_type.model(backbone=backbone, num_classes=class_map.num_classes, img_size=image_size)
# def accuracy_cavid copied into code and was left out here
learn = model_type.fastai.learner(dls=[train_dl, valid_dl], model=model, metrics=[accuracy_camvid])

learn.lr_find()

Results with:

---------------------------------------------------------------------------
AttributeError                            Traceback (most recent call last)
/home/user/Documents/IceVision/icevision_unet_class.ipynb Cell 15' in <cell line: 1>()
----> 1 learn.lr_find()

File ~/anaconda3/envs/icevision/lib/python3.8/site-packages/fastai/callback/schedule.py:285, in lr_find(self, start_lr, end_lr, num_it, stop_div, show_plot, suggest_funcs)
    [283](file:///~/anaconda3/envs/icevision/lib/python3.8/site-packages/fastai/callback/schedule.py?line=282) n_epoch = num_it//len(self.dls.train) + 1
    [284](file:///~/anaconda3/envs/icevision/lib/python3.8/site-packages/fastai/callback/schedule.py?line=283) cb=LRFinder(start_lr=start_lr, end_lr=end_lr, num_it=num_it, stop_div=stop_div)
--> [285](file:///~/anaconda3/envs/icevision/lib/python3.8/site-packages/fastai/callback/schedule.py?line=284) with self.no_logging(): self.fit(n_epoch, cbs=cb)
    [286](file:///~/anaconda3/envs/icevision/lib/python3.8/site-packages/fastai/callback/schedule.py?line=285) if suggest_funcs is not None:
    [287](file:///~/anaconda3/envs/icevision/lib/python3.8/site-packages/fastai/callback/schedule.py?line=286)     lrs, losses = tensor(self.recorder.lrs[num_it//10:-5]), tensor(self.recorder.losses[num_it//10:-5])

File ~/anaconda3/envs/icevision/lib/python3.8/site-packages/fastai/learner.py:221, in Learner.fit(self, n_epoch, lr, wd, cbs, reset_opt)
    [219](file:///~/anaconda3/envs/icevision/lib/python3.8/site-packages/fastai/learner.py?line=218) self.opt.set_hypers(lr=self.lr if lr is None else lr)
    [220](file:///~/anaconda3/envs/icevision/lib/python3.8/site-packages/fastai/learner.py?line=219) self.n_epoch = n_epoch
--> [221](file:///~/anaconda3/envs/icevision/lib/python3.8/site-packages/fastai/learner.py?line=220) self._with_events(self._do_fit, 'fit', CancelFitException, self._end_cleanup)

File ~/anaconda3/envs/icevision/lib/python3.8/site-packages/fastai/learner.py:163, in Learner._with_events(self, f, event_type, ex, final)
    [162](file:///~/anaconda3/envs/icevision/lib/python3.8/site-packages/fastai/learner.py?line=161) def _with_events(self, f, event_type, ex, final=noop):
--> [163](file:///~/anaconda3/envs/icevision/lib/python3.8/site-packages/fastai/learner.py?line=162)     try: self(f'before_{event_type}');  f()
    [164](file:///~/anaconda3/envs/icevision/lib/python3.8/site-packages/fastai/learner.py?line=163)     except ex: self(f'after_cancel_{event_type}')
    [165](file:///~/anaconda3/envs/icevision/lib/python3.8/site-packages/fastai/learner.py?line=164)     self(f'after_{event_type}');  final()

File ~/anaconda3/envs/icevision/lib/python3.8/site-packages/fastai/learner.py:212, in Learner._do_fit(self)
    [210](file:///~/anaconda3/envs/icevision/lib/python3.8/site-packages/fastai/learner.py?line=209) for epoch in range(self.n_epoch):
    [211](file:///~/anaconda3/envs/icevision/lib/python3.8/site-packages/fastai/learner.py?line=210)     self.epoch=epoch
--> [212](file:///~/anaconda3/envs/icevision/lib/python3.8/site-packages/fastai/learner.py?line=211)     self._with_events(self._do_epoch, 'epoch', CancelEpochException)

File ~/anaconda3/envs/icevision/lib/python3.8/site-packages/fastai/learner.py:163, in Learner._with_events(self, f, event_type, ex, final)
    [162](file:///~/anaconda3/envs/icevision/lib/python3.8/site-packages/fastai/learner.py?line=161) def _with_events(self, f, event_type, ex, final=noop):
--> [163](file:///~/anaconda3/envs/icevision/lib/python3.8/site-packages/fastai/learner.py?line=162)     try: self(f'before_{event_type}');  f()
    [164](file:///~/anaconda3/envs/icevision/lib/python3.8/site-packages/fastai/learner.py?line=163)     except ex: self(f'after_cancel_{event_type}')
    [165](file:///~/anaconda3/envs/icevision/lib/python3.8/site-packages/fastai/learner.py?line=164)     self(f'after_{event_type}');  final()

File ~/anaconda3/envs/icevision/lib/python3.8/site-packages/fastai/learner.py:206, in Learner._do_epoch(self)
    [205](file:///~/anaconda3/envs/icevision/lib/python3.8/site-packages/fastai/learner.py?line=204) def _do_epoch(self):
--> [206](file:///~/anaconda3/envs/icevision/lib/python3.8/site-packages/fastai/learner.py?line=205)     self._do_epoch_train()
    [207](file:///~/anaconda3/envs/icevision/lib/python3.8/site-packages/fastai/learner.py?line=206)     self._do_epoch_validate()

File ~/anaconda3/envs/icevision/lib/python3.8/site-packages/fastai/learner.py:198, in Learner._do_epoch_train(self)
    [196](file:///~/anaconda3/envs/icevision/lib/python3.8/site-packages/fastai/learner.py?line=195) def _do_epoch_train(self):
    [197](file:///~/anaconda3/envs/icevision/lib/python3.8/site-packages/fastai/learner.py?line=196)     self.dl = self.dls.train
--> [198](file:///~/anaconda3/envs/icevision/lib/python3.8/site-packages/fastai/learner.py?line=197)     self._with_events(self.all_batches, 'train', CancelTrainException)

File ~/anaconda3/envs/icevision/lib/python3.8/site-packages/fastai/learner.py:163, in Learner._with_events(self, f, event_type, ex, final)
    [162](file:///~/anaconda3/envs/icevision/lib/python3.8/site-packages/fastai/learner.py?line=161) def _with_events(self, f, event_type, ex, final=noop):
--> [163](file:///~/anaconda3/envs/icevision/lib/python3.8/site-packages/fastai/learner.py?line=162)     try: self(f'before_{event_type}');  f()
    [164](file:///~/anaconda3/envs/icevision/lib/python3.8/site-packages/fastai/learner.py?line=163)     except ex: self(f'after_cancel_{event_type}')
    [165](file:///~/anaconda3/envs/icevision/lib/python3.8/site-packages/fastai/learner.py?line=164)     self(f'after_{event_type}');  final()

File ~/anaconda3/envs/icevision/lib/python3.8/site-packages/fastai/learner.py:169, in Learner.all_batches(self)
    [167](file:///~/anaconda3/envs/icevision/lib/python3.8/site-packages/fastai/learner.py?line=166) def all_batches(self):
    [168](file:///~/anaconda3/envs/icevision/lib/python3.8/site-packages/fastai/learner.py?line=167)     self.n_iter = len(self.dl)
--> [169](file:///~/anaconda3/envs/icevision/lib/python3.8/site-packages/fastai/learner.py?line=168)     for o in enumerate(self.dl): self.one_batch(*o)

File ~/anaconda3/envs/icevision/lib/python3.8/site-packages/fastai/data/load.py:109, in DataLoader.__iter__(self)
    [107](file:///~/anaconda3/envs/icevision/lib/python3.8/site-packages/fastai/data/load.py?line=106) self.before_iter()
    [108](file:///~/anaconda3/envs/icevision/lib/python3.8/site-packages/fastai/data/load.py?line=107) self.__idxs=self.get_idxs() # called in context of main process (not workers/subprocesses)
--> [109](file:///~/anaconda3/envs/icevision/lib/python3.8/site-packages/fastai/data/load.py?line=108) for b in _loaders[self.fake_l.num_workers==0](self.fake_l):
    [110](file:///~/anaconda3/envs/icevision/lib/python3.8/site-packages/fastai/data/load.py?line=109)     if self.device is not None: b = to_device(b, self.device)
    [111](file:///~/anaconda3/envs/icevision/lib/python3.8/site-packages/fastai/data/load.py?line=110)     yield self.after_batch(b)

File ~/anaconda3/envs/icevision/lib/python3.8/site-packages/torch/utils/data/dataloader.py:521, in _BaseDataLoaderIter.__next__(self)
    [519](file:///~/anaconda3/envs/icevision/lib/python3.8/site-packages/torch/utils/data/dataloader.py?line=518) if self._sampler_iter is None:
    [520](file:///~/anaconda3/envs/icevision/lib/python3.8/site-packages/torch/utils/data/dataloader.py?line=519)     self._reset()
--> [521](file:///~/anaconda3/envs/icevision/lib/python3.8/site-packages/torch/utils/data/dataloader.py?line=520) data = self._next_data()
    [522](file:///~/anaconda3/envs/icevision/lib/python3.8/site-packages/torch/utils/data/dataloader.py?line=521) self._num_yielded += 1
    [523](file:///~/anaconda3/envs/icevision/lib/python3.8/site-packages/torch/utils/data/dataloader.py?line=522) if self._dataset_kind == _DatasetKind.Iterable and \
    [524](file:///~/anaconda3/envs/icevision/lib/python3.8/site-packages/torch/utils/data/dataloader.py?line=523)         self._IterableDataset_len_called is not None and \
    [525](file:///~/anaconda3/envs/icevision/lib/python3.8/site-packages/torch/utils/data/dataloader.py?line=524)         self._num_yielded > self._IterableDataset_len_called:

File ~/anaconda3/envs/icevision/lib/python3.8/site-packages/torch/utils/data/dataloader.py:561, in _SingleProcessDataLoaderIter._next_data(self)
    [559](file:///~/anaconda3/envs/icevision/lib/python3.8/site-packages/torch/utils/data/dataloader.py?line=558) def _next_data(self):
    [560](file:///~/anaconda3/envs/icevision/lib/python3.8/site-packages/torch/utils/data/dataloader.py?line=559)     index = self._next_index()  # may raise StopIteration
--> [561](file:///~/anaconda3/envs/icevision/lib/python3.8/site-packages/torch/utils/data/dataloader.py?line=560)     data = self._dataset_fetcher.fetch(index)  # may raise StopIteration
    [562](file:///~/anaconda3/envs/icevision/lib/python3.8/site-packages/torch/utils/data/dataloader.py?line=561)     if self._pin_memory:
    [563](file:///~/anaconda3/envs/icevision/lib/python3.8/site-packages/torch/utils/data/dataloader.py?line=562)         data = _utils.pin_memory.pin_memory(data)

File ~/anaconda3/envs/icevision/lib/python3.8/site-packages/torch/utils/data/_utils/fetch.py:39, in _IterableDatasetFetcher.fetch(self, possibly_batched_index)
     [37](file:///~/anaconda3/envs/icevision/lib/python3.8/site-packages/torch/utils/data/_utils/fetch.py?line=36)         raise StopIteration
     [38](file:///~/anaconda3/envs/icevision/lib/python3.8/site-packages/torch/utils/data/_utils/fetch.py?line=37) else:
---> [39](file:///~/anaconda3/envs/icevision/lib/python3.8/site-packages/torch/utils/data/_utils/fetch.py?line=38)     data = next(self.dataset_iter)
     [40](file:///~/anaconda3/envs/icevision/lib/python3.8/site-packages/torch/utils/data/_utils/fetch.py?line=39) return self.collate_fn(data)

File ~/anaconda3/envs/icevision/lib/python3.8/site-packages/fastai/data/load.py:118, in DataLoader.create_batches(self, samps)
    [116](file:///~/anaconda3/envs/icevision/lib/python3.8/site-packages/fastai/data/load.py?line=115) if self.dataset is not None: self.it = iter(self.dataset)
    [117](file:///~/anaconda3/envs/icevision/lib/python3.8/site-packages/fastai/data/load.py?line=116) res = filter(lambda o:o is not None, map(self.do_item, samps))
--> [118](file:///~/anaconda3/envs/icevision/lib/python3.8/site-packages/fastai/data/load.py?line=117) yield from map(self.do_batch, self.chunkify(res))

File ~/anaconda3/envs/icevision/lib/python3.8/site-packages/fastai/data/load.py:[144](file:///~/anaconda3/envs/icevision/lib/python3.8/site-packages/fastai/data/load.py?line=143), in DataLoader.do_batch(self, b)
--> 144 def do_batch(self, b): return self.retain(self.create_batch(self.before_batch(b)), b)

File ~/anaconda3/envs/icevision/lib/python3.8/site-packages/icevision/engines/fastai/adapters/convert_dataloader_to_fastai.py:18, in convert_dataloader_to_fastai.<locals>.FastaiDataLoaderWithCollate.create_batch(self, b)
     [17](file:///~/anaconda3/envs/icevision/lib/python3.8/site-packages/icevision/engines/fastai/adapters/convert_dataloader_to_fastai.py?line=16) def create_batch(self, b):
---> [18](file:///~/anaconda3/envs/icevision/lib/python3.8/site-packages/icevision/engines/fastai/adapters/convert_dataloader_to_fastai.py?line=17)     return (dataloader.collate_fn, raise_error_convert)[self.prebatched](b)

File ~/anaconda3/envs/icevision/lib/python3.8/site-packages/icevision/models/utils.py:88, in unload_records.<locals>.inner(records)
     [87](file:///~/anaconda3/envs/icevision/lib/python3.8/site-packages/icevision/models/utils.py?line=86) def inner(records):
---> [88](file:///~/anaconda3/envs/icevision/lib/python3.8/site-packages/icevision/models/utils.py?line=87)     tupled_output, records = build_batch(records)
     [89](file:///~/anaconda3/envs/icevision/lib/python3.8/site-packages/icevision/models/utils.py?line=88)     for record in records:
     [90](file:///~/anaconda3/envs/icevision/lib/python3.8/site-packages/icevision/models/utils.py?line=89)         record.unload()

File ~/anaconda3/envs/icevision/lib/python3.8/site-packages/icevision/models/utils.py:79, in apply_batch_tfms.<locals>.inner(records)
     [77](file:///~/anaconda3/envs/icevision/lib/python3.8/site-packages/icevision/models/utils.py?line=76) if batch_tfms is not None:
     [78](file:///~/anaconda3/envs/icevision/lib/python3.8/site-packages/icevision/models/utils.py?line=77)     records = batch_tfms(records)
---> [79](file:///~/anaconda3/envs/icevision/lib/python3.8/site-packages/icevision/models/utils.py?line=78) return build_batch(records, **build_batch_kwargs)

File ~/anaconda3/envs/icevision/lib/python3.8/site-packages/icevision/models/fastai/unet/dataloaders.py:59, in build_train_batch(records)
     [55](file:///~/anaconda3/envs/icevision/lib/python3.8/site-packages/icevision/models/fastai/unet/dataloaders.py?line=54) for record in records:
     [56](file:///~/anaconda3/envs/icevision/lib/python3.8/site-packages/icevision/models/fastai/unet/dataloaders.py?line=55)     # can be optimzed to be converted to tensor once at the end
     [57](file:///~/anaconda3/envs/icevision/lib/python3.8/site-packages/icevision/models/fastai/unet/dataloaders.py?line=56)     tensor_images.append(im2tensor(record.img))
     [58](file:///~/anaconda3/envs/icevision/lib/python3.8/site-packages/icevision/models/fastai/unet/dataloaders.py?line=57)     tensor_masks.append(
---> [59](file:///~/anaconda3/envs/icevision/lib/python3.8/site-packages/icevision/models/fastai/unet/dataloaders.py?line=58)         tensor(record.segmentation.mask_array.data).long().squeeze()
     [60](file:///~/anaconda3/envs/icevision/lib/python3.8/site-packages/icevision/models/fastai/unet/dataloaders.py?line=59)     )
     [62](file:///~/anaconda3/envs/icevision/lib/python3.8/site-packages/icevision/models/fastai/unet/dataloaders.py?line=61) tensor_images = torch.stack(tensor_images)
     [63](file:///~/anaconda3/envs/icevision/lib/python3.8/site-packages/icevision/models/fastai/unet/dataloaders.py?line=62) tensor_masks = torch.stack(tensor_masks)

File ~/anaconda3/envs/icevision/lib/python3.8/site-packages/icevision/core/components/composite.py:46, in TaskComposite.__getattr__(self, name)
     [43](file:///~/anaconda3/envs/icevision/lib/python3.8/site-packages/icevision/core/components/composite.py?line=42) except KeyError:
     [44](file:///~/anaconda3/envs/icevision/lib/python3.8/site-packages/icevision/core/components/composite.py?line=43)     pass
---> [46](file:///~/anaconda3/envs/icevision/lib/python3.8/site-packages/icevision/core/components/composite.py?line=45) raise AttributeError(f"{self.__class__.__name__} has no attribute {name}")

AttributeError: BaseRecord has no attribute segmentation

Expected behavior
For it to work...

Screenshots
No applicable screenshots.

Desktop (please complete the following information):

  • OS: CentOS8
@DrAle85 DrAle85 added the bug Something isn't working label Mar 30, 2022
@FraPochetti
Copy link
Contributor

You are confusing Instance with Semantic Segmentation.
UNET is for Semantic Segmentation.
You can check how to parse data appropriately for it here

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Assignees
No one assigned
Labels
bug Something isn't working
Projects
None yet
Milestone
No milestone
Development

No branches or pull requests
2 participants
@FraPochetti @DrAle85