Error while training a new entity with large datasets

[HarshavardhanG513]

How to reproduce the behaviour

Your Environment

• Operating System:
• Python Version Used:
• spaCy Version Used:
• Environment Information:

[HarshavardhanG513]

sorry i pressed enter before writing something, so here's the problem. I am training a model to identify job titles as an entity. I used a slightly modified version of the example code on the spacy.io website for training an additional entity.

while running the code with small amounts of data, it works. When tried to improve the accuracy by adding a lot more data, it stopped working and instead gave me the following:

File "getjob.py", line 389, in main() File "getjob.py", line 360, in main nlp.update(texts, annotations, sgd=optimizer, drop=0.35, losses=losses) File "C:\Users\Harsha\Miniconda3\lib\site-packages\spacy\language.py", line 409, in update doc = self.make_doc(doc) File "C:\Users\Harsha\Miniconda3\lib\site-packages\spacy\language.py", line 378, in make_doc return self.tokenizer(text) SystemError: <spacy.tokenizer.Tokenizer object at 0x05B6F270> returned NULL without setting an error

[HarshavardhanG513]

spaCy version 2.0.12 Location C:\Users\Harsha\Miniconda3\lib\site-packages\spacy Platform Windows-10-10.0.17763-SP0 Python version 3.7.1 Models en, en_core_web_md

[ines]

SystemError: <spacy.tokenizer.Tokenizer object at 0x05B6F270> returned NULL without setting an error

This error sounds like you might be using a 32-bit version of Python instead of 64-bit. Also see #3250, where this has come up before.

Could you double-check your Python installation and try again with a 64-bit installation?

[HarshavardhanG513]

doesn't recognize spacy anymore. I used the following conda command: conda install -c conda-forge spacy the requested packages were installed python -m spacy download en no module named spacy.

just trying to resolve the error

[HarshavardhanG513]

i think its working. thank a lot. your advanced spacy course on datacamp was also quite helpful when i started.

[lock[bot]]

This thread has been automatically locked since there has not been any recent activity after it was closed. Please open a new issue for related bugs.