Issue in importing Apex

[Alamelu1812]

I am having issues while importing apex. I get an error similar to the ones posted in run_classifier repository.

in
----> 1 import apex
2 import pandas as pd
3 import numpy as np
4 import torch
5

/anaconda3/lib/python3.7/site-packages/apex/init.py in
16 from apex.exceptions import (ApexAuthSecret,
17 ApexSessionSecret)
---> 18 from apex.interfaces import (ApexImplementation,
19 IApex)
20 from apex.lib.libapex import (groupfinder,

/anaconda3/lib/python3.7/site-packages/apex/interfaces.py in
8 pass
9
---> 10 class ApexImplementation(object):
11 """ Class so that we can tell if Apex is installed from other
12 applications

/anaconda3/lib/python3.7/site-packages/apex/interfaces.py in ApexImplementation()
12 applications
13 """
---> 14 implements(IApex)

/anaconda3/lib/python3.7/site-packages/zope/interface/declarations.py in implements(*interfaces)
481 # the coverage for this block there. :(
482 if PYTHON3:
--> 483 raise TypeError(_ADVICE_ERROR % 'implementer')
484 _implements("implements", interfaces, classImplements)
485

TypeError: Class advice impossible in Python3. Use the @Implementer class decorator instead.

Could anyone help me with this?

[negacy]

I had the same issue. If you didn't import apex, it will work fine, but I think that is used for faster computation.

[negacy]

Also, you can install it using their GitHub page.

[Alamelu1812]

If I don't import apex, the following statement fails.

from pytorch_pretrained_bert.modeling import BertForPreTraining, BertPreTrainedModel, BertModel, BertConfig, BertForMaskedLM, BertForSequenceClassification

I would like to know whether the above import worked fine without import apex.

I am getting the following error without importing apex. Any suggestions would be really helpful.

TypeError Traceback (most recent call last)
in
24
25 from pytorch_pretrained_bert.tokenization import BertTokenizer, WordpieceTokenizer
---> 26 from pytorch_pretrained_bert.modeling import BertForPreTraining, BertPreTrainedModel, BertModel, BertConfig, BertForMaskedLM, BertForSequenceClassification

/anaconda3/lib/python3.7/site-packages/pytorch_pretrained_bert/init.py in
5 from .tokenization_gpt2 import GPT2Tokenizer
6
----> 7 from .modeling import (BertConfig, BertModel, BertForPreTraining,
8 BertForMaskedLM, BertForNextSentencePrediction,
9 BertForSequenceClassification, BertForMultipleChoice,

/anaconda3/lib/python3.7/site-packages/pytorch_pretrained_bert/modeling.py in
216
217 try:
--> 218 from apex.normalization.fused_layer_norm import FusedLayerNorm as BertLayerNorm
219 except ImportError:
220 print("Better speed can be achieved with apex installed from https://www.github.com/nvidia/apex.")

/anaconda3/lib/python3.7/site-packages/apex/init.py in
16 from apex.exceptions import (ApexAuthSecret,
17 ApexSessionSecret)
---> 18 from apex.interfaces import (ApexImplementation,
19 IApex)
20 from apex.lib.libapex import (groupfinder,

/anaconda3/lib/python3.7/site-packages/apex/interfaces.py in
8 pass
9
---> 10 class ApexImplementation(object):
11 """ Class so that we can tell if Apex is installed from other
12 applications

/anaconda3/lib/python3.7/site-packages/apex/interfaces.py in ApexImplementation()
12 applications
13 """
---> 14 implements(IApex)

/anaconda3/lib/python3.7/site-packages/zope/interface/declarations.py in implements(*interfaces)
481 # the coverage for this block there. :(
482 if PYTHON3:
--> 483 raise TypeError(_ADVICE_ERROR % 'implementer')
484 _implements("implements", interfaces, classImplements)
485

TypeError: Class advice impossible in Python3. Use the @Implementer class decorator instead.

[negacy]

What Python version are you using? Try it with 3.6

[Alamelu1812]

I am using Python 3.7. I read in the forums that moving back to 3.6 should solve the problem. I'll try and let you know. Thanks for your help

[bhoomit]

Was able to solve it with

git clone https://github.com/NVIDIA/apex.git && cd apex && python setup.py install --cuda_ext --cpp_ext