before a while I trained an NER prodigy model and exported it to a directory. previously, I dragged and drop the model to another project directory and then I dragged it back to the original project directory (where prodigy saved it)
Another step what I did (I don/t know if it is worth mention it) I changed the project and directory name..
since these steps, I can't use the model and I am getting this error:
File "spacy/vectors.pyx", line 605, in spacy.vectors.Vectors.from_disk.load_vectors
File "./XnowledgeToOnto/venv/lib/python3.10/site-packages/numpy/lib/npyio.py", line 438, in load
raise ValueError("Cannot load file containing pickled data "
ValueError: Cannot load file containing pickled data when allow_pickle=False
Please help me, I have to fix it as soon as possible for my university project.
Note: I have no more access to prodigy because my 3 month free license has ended..
So it sounds like you accidentally dragged and perhaps messed up files in your model, is that right?
Why can't you rebuild your model using your original code?
Sorry for the confusion, but you can still keep using Prodigy after 3 months with the 3 month research license. The 3 months is just how many months of updates (i.e., newer versions) of Prodigy you can receive for free. So if you received your research license in May and received v1.11.14 at that time, you can still keep using v1.11.14 for however long. In fact, you can now even install up to v1.13.3 as it was released within 3 months.
I just used your email address to lookup your license and double checked that it still works through pypi.
So can you try to reinstall Prodigy with pypi and try to rerun your model?
If you lost your license key, you can use this form to email it back to yourself. You can then install Prodigy through pypi or go to https://download.prodi.gy, put in your license key as the Username, and you can download the wheel file for any Prodigy version your eligible to receive. You may want to hold onto that wheel file so you can keep using that version of Prodigy in the future.
Thank you a lot!
You are right, after going through this trauma I thought that I had to try to retrain my model and check if prodigy still works (because the database of annotated texts was still there) and it happily worked!