Flair: AttributeError: 'GRU' object has no attribute '_flat_weights_names'

[carlodavid012]

There is error when loading a text classifier,

Code:

from flair.models import TextClassifier

classifier = TextClassifier.load('en-sentiment')

Error:

AttributeError                            Traceback (most recent call last)
<ipython-input-4-f517a501f1cf> in <module>
----> 1 classifier = TextClassifier.load('en-sentiment')

~/Desktop/flair/flair/nn.py in load(cls, model)
     86             state = torch.load(f, map_location=flair.device)
     87 
---> 88         model = cls._init_model_with_state_dict(state)
     89 
     90         model.eval()

~/Desktop/flair/flair/models/text_classification_model.py in _init_model_with_state_dict(state)
    129             multi_label=state["multi_label"],
    130             beta=beta,
--> 131             loss_weights=weights,
    132         )
    133 

~/Desktop/flair/flair/models/text_classification_model.py in __init__(self, document_embeddings, label_dictionary, multi_label, multi_label_threshold, beta, loss_weights)
     90 
     91         # auto-spawn on GPU if available
---> 92         self.to(flair.device)
     93 
     94     def _init_weights(self):

~/.local/lib/python3.6/site-packages/torch/nn/modules/module.py in to(self, *args, **kwargs)
    423             return t.to(device, dtype if t.is_floating_point() else None, non_blocking)
    424 
--> 425         return self._apply(convert)
    426 
    427     def register_backward_hook(self, hook):

~/.local/lib/python3.6/site-packages/torch/nn/modules/module.py in _apply(self, fn)
    199     def _apply(self, fn):
    200         for module in self.children():
--> 201             module._apply(fn)
    202 
    203         def compute_should_use_set_data(tensor, tensor_applied):

~/.local/lib/python3.6/site-packages/torch/nn/modules/module.py in _apply(self, fn)
    199     def _apply(self, fn):
    200         for module in self.children():
--> 201             module._apply(fn)
    202 
    203         def compute_should_use_set_data(tensor, tensor_applied):

~/.local/lib/python3.6/site-packages/torch/nn/modules/rnn.py in _apply(self, fn)
    135         # Note: be v. careful before removing this, as 3rd party device types
    136         # likely rely on this behavior to properly .to() modules like LSTM.
--> 137         self._flat_weights = [getattr(self, weight) for weight in self._flat_weights_names]
    138 
    139         # Flattens params (on CUDA)

~/.local/lib/python3.6/site-packages/torch/nn/modules/module.py in __getattr__(self, name)
    574                 return modules[name]
    575         raise AttributeError("'{}' object has no attribute '{}'".format(
--> 576             type(self).__name__, name))
    577 
    578     def __setattr__(self, name, value):

AttributeError: 'GRU' object has no attribute '_flat_weights_names'

Environment

• Ubuntu 18.04
• Version: 0.4.4

[djstrong]

It is probably fixed in master: https://github.com/flairNLP/flair/pull/1360

[alanakbik]

@carlodavid012 thanks for reporting this - the error persists even in master. I will push a fix in a moment. We will also do a release probably today so it will then also be fixed if you install the new version through pip!

[chandu188]

@alanakbik Could you please confirm if it is fixed? We still face this issue in the 0.4.5 release.

[alanakbik]

@chandu188 in our tests, yes. To reproduce:

• open a new Python3 notebook on Google Colab

• update libraries

pip install --upgrade flair torch

• code snippet

from flair.models import TextClassifier
from flair.data import Sentence

classifier = TextClassifier.load('en-sentiment')

classifier.predict(Sentence('This movie is great'))