Error in easyocr.Reader with urlretrieve(model_url[model_file][0], MODEL_PATH)

[z-aliakseyeu]

Hello!
Thanks for that amazing library first of all!
Could someone please help to resolve the issue i encountered today only (yesterday and before it was working smoothly).

in my code i have let's say:

import easyocr
reader = easyocr.Reader(['id', 'en'])

When i run it - i am getting the following error:

CUDA not available - defaulting to CPU. Note: This module is much faster with a GPU.
MD5 hash mismatch, possible file corruption
Re-downloading the recognition model, please wait
Traceback (most recent call last):
  File "tryout_easyocr.py", line 5, in <module>
    reader = easyocr.Reader(['id', 'en'])
  File "/usr/local/lib/python3.7/site-packages/easyocr/easyocr.py", line 194, in __init__
    urlretrieve(model_url[model_file][0], MODEL_PATH)
  File "/usr/local/Cellar/python/3.7.8/Frameworks/Python.framework/Versions/3.7/lib/python3.7/urllib/request.py", line 288, in urlretrieve
    % (read, size), result)

Regardless what language i choose - i face this error in all the environments:

• in mac os runtime
• in docker
• in ubuntu
• in colab https://colab.research.google.com/github/vistec-AI/colab/blob/master/easyocr.ipynb#scrollTo=lIYdn1woOS1n

Diving deeper it tries to download the following file:
https://www.jaided.ai/read_download/latin.pth
which i wasn't able to download with wget, curl or browser as well for the same issue.

Seems https://www.jaided.ai/ resets the connection during download

[Proteusiq]

I am running into the same issue
[RESOLVED]

It is Download issues that happens due to network overload(multiple people trying to download at the same time)
Solved by directly downloading models and place them to %USERPROFILE%/.EasyOCR/model from

detector weight

latin

See #11

[z-aliakseyeu]

Thanks @Proteusiq !

Now the only question i have is how to download Indonesian lang model 😅

[rkcosmos]

Indonesian language use latin model. If latin model is downloaded and placed in model directory. You should be able to call
reader = easyocr.Reader(['id', 'en']).

[z-aliakseyeu]

Thanks @rkcosmos!
Is there some plans/ideas how to solve it without manual download?
Maybe i can contribute to make it work smoother?

[rkcosmos]

I plan to move download part behind Cloudflare and let them handle the traffic.

[imvladikon]

The manual download also doesn't work. I checked it (e.g. wget https://www.jaided.ai/read_download/cyrillic.pth -O /content/models/cyrillic.pth and just using a browser download) after several seconds, downloading is interrupted

[rkcosmos]

cyrillic model

[rkcosmos]

too many people trying to get model, I will move it behind Cloudflare tomorrow, otherwise I'm going to broke because of traffic bill.

[SiYuan-Lee]

I encounter the same issue, where can I download the simple Chinese model instead of the Latin model.

[rkcosmos]

All links are on README.

[xhimalaya]

I am running into the same issue
[RESOLVED]

It is Download issues that happens due to network overload(multiple people trying to download at the same time)
Solved by directly downloading models and place them to %USERPROFILE%/.EasyOCR/model from

detector weight

latin

See #11

I am running into the same issue
[RESOLVED]

It is Download issues that happens due to network overload(multiple people trying to download at the same time)
Solved by directly downloading models and place them to %USERPROFILE%/.EasyOCR/model from

detector weight

latin

See #11

can you please tell me where to put those files after downloading???

[alexpm94]

@xhimalaya Place them in ~/.EasyOCR/model. That directory is created when easyocr is installed.

[Prasang-Biyani]

While downloading the model, I am facing this issue.

[rkcosmos]

sorry, my website is down

[rkcosmos]

it should be up now.

[Prasang-Biyani]

nope, still not working

[rkcosmos]

not sure what's going on. At least website is up. You can also try manual download links from google drive. The links are on README.

[Vijayabhaskar96]

A Suggestion to @rkcosmos You can host the models on https://gofile.io and add the direct download link as a mirror link to it so that if the model fails to download from your website it can retry to download from Gofile.io for the meantime till you host the models on Cloudflare.

[rkcosmos]

thanks for the suggestion, it's already behind CDN. Good to know the alternative.

[imvladikon]

@Vijayabhaskar96 @rkcosmos I think there is a way to do it like in the spacy package and model for spacy. They just provide models through assets on the GitHub and module for downloading it e.g. https://github.com/explosion/spaCy/releases/tag/v2.3.0 and another option is just google drive (retrieving using https://pypi.org/project/gdown/)

[rkcosmos]

@imvladikon @Vijayabhaskar96 At the end, cloudflare doesn't allow heavy download traffic. I end up with downloading from google drive. thanks for ur suggestion.