NER Entity recognition Permission Error (13, 'Access is denied

1

made a config:

why does
python -m prodigy train --ner report_data1 ./tmp_model1 --eval-split 0.2

give me a

:warning: Aborting and saving the final best model. Encountered exception:
PermissionError(13, 'Access is denied')

Full read out below with error at bottom:

:information_source: Using CPU

========================= Generating Prodigy config =========================
:information_source: Auto-generating config with spaCy
:heavy_check_mark: Generated training config

=========================== Initializing pipeline ===========================
[2024-02-15 23:40:11,882] [INFO] Set up nlp object from config
Components: ner
Merging training and evaluation data for 1 components

  • [ner] Training: 82 | Evaluation: 20 (20% split)
    Training: 82 | Evaluation: 20
    Labels: ner (2)
    [2024-02-15 23:40:11,933] [INFO] Pipeline: ['tok2vec', 'ner']
    [2024-02-15 23:40:11,935] [INFO] Created vocabulary
    [2024-02-15 23:40:11,935] [INFO] Finished initializing nlp object
    [2024-02-15 23:40:13,078] [INFO] Initialized pipeline components: ['tok2vec', 'ner']
    :heavy_check_mark: Initialized pipeline

============================= Training pipeline =============================
Components: ner
Merging training and evaluation data for 1 components

  • [ner] Training: 82 | Evaluation: 20 (20% split)
    Training: 82 | Evaluation: 20
    Labels: ner (2)
    :information_source: Pipeline: ['tok2vec', 'ner']
    :information_source: Initial learn rate: 0.001
    E # LOSS TOK2VEC LOSS NER ENTS_F ENTS_P ENTS_R SCORE

0 0 0.00 178.50 0.00 0.00 0.00 0.00
:warning: Aborting and saving the final best model. Encountered exception:
PermissionError(13, 'Access is denied')

2

This seems to suggest you may not have permission to write to the output folder you've specified ./tmp_model1. Can you confirm you writeable permission to this folder? If not, can you retry but pointing to a folder where you do have writeable permission?

Hope this helps!

3

Thanks Ryan,

I am using Jupyterlab/Anaconda on Windows, the folder I am saving to I can make and save files into no problem (i can make directories in it in with python on jupyterlab and write files). Just checked and I have full access on windows. Does Jupyterlab superimpose some new permissions while in an environment?

Dave

4

Unfortunately, I don't know. I'm not an expert on JupyterLab, especially not in Windows. Just curious are you running on a company's network? There could be a lot of background things like a virus protector that's blocking. You may want to talk to an IT admin if they know of anything blocking.

Can you run your Prodigy command in terminal? While we offer the Jupyter extension, Prodigy was really designed to be run in the command line. I've noticed in the past some users get confused thinking they must just the Jupyter notebook when in fact it may add a lot of hidden problems like this.

5

Ok will give that a try. I was using anaconda for the environment management. This is the second problem I have had with Jupyterlab/plugin. Back to basics it is.

6

Thanks Ryan. I have no idea what is going with JupyterLab as the folder I am using I have the required permissions. I can write files, delete files etc etc within JupyterLab within it but it wont play ball with prodigy doing stuff within within a JupyterLab terminal or even a completely separate windows terminal operating on that JupyterLab folder.

Anyway I copied the data into a completely different folder on a completely different drive and it worked on a windows terminal. This must be some weird permissions thing but everything looks fine when I look at the usage rights. Will battle with this when I get bored.

This looks solved from my perspective. Close this from my end. Thanks for your help Ryan.