Iddoyadlin / hebrew-w2v

a complete reproducible example of training a word2vec model for Hebrew
Apache License 2.0
11 stars 1 forks source link

Unable to load model #1

Closed omrishsu closed 2 years ago

omrishsu commented 2 years ago

when i try to load the model: model_w2v = Word2Vec.load('/home/hebrew-word2vec/wiki_tokenized_model/model.mdl') i get: _pickle.UnpicklingError: invalid load key, '\x00'.

on: return _pickle.load(f, encoding='latin1') in: gensim.utils.unpickle

any idea what is missing here?

Iddoyadlin commented 2 years ago

@omrivm I'm using windows 11, python 3.9.10 and gensim 4.1.2 (installed in venv). what are you using? what does the following output:

  1. pip show gensim
  2. python --version
omrishsu commented 2 years ago

@Iddoyadlin found the problem. For some reason, the rar file is extracted ok on windows, but on ubuntu, it is extracted corrupted.

Iddoyadlin commented 2 years ago

@omrivm were you able to extract it on ubuntu? care to share how you did it?

omrishsu commented 2 years ago

No, i was not able to do that. I've extracted it on windows and transfer the mdl file to ubuntu.