Instead of downloading pertained model from hosted, how can I download the pertained models to local

[madhavi1102]

while instantiating ocr_predictor() model with pretrained models 'db_resnet50' and 'crnn_vgg16_bn' which get downloaded from public hosted.
Instead of it, how can I download these pretrained models locally and refer to local path for these models while instantiating ocr_predictor.

Thanks in advance,
Madhavi

[felixdittrich92]

Hi @madhavi1102 👋,

Take a look at https://mindee.github.io/doctr/using_doctr/custom_models_training.html#loading-your-custom-trained-model how to load from local :)

The already downloaded models are stored in .cache/doctr/models you can copy paste it from there

[madhavi1102]

Hi @felixdittrich92 ,

Got it, Thanks for the quick response :)

[yumikim381]

@felixdittrich92 Hi, i couldn't find the folder, is the location still accurate?

[aryakp2302]

while instantiating ocr_predictor() model with pretrained models 'db_resnet50' and 'crnn_vgg16_bn' which get downloaded from public hosted. Instead of it, how can I download these pretrained models locally and refer to local path for these models while instantiating ocr_predictor.

Thanks in advance, Madhavi

Have you find any solution because i also wanted to use pretrained model locally...it would be really great if you can help me Thank you In advance, Arya Patel

[felixT2K]

Hi @madhavi1102 👋🏼,

If you have already downloaded the checkpoints they are stored in /home/.cache/doctr/models otherwise you can also download specific models from the url

for example db_resnet50 pytorch:

doctr/doctr/models/detection/differentiable_binarization/pytorch.py

Line 30 in 0d84915

"url": "https://doctr-static.mindee.com/models?id=v0.7.0/db_resnet50-79bd7d70.pt&src=0",

[yumikim381]

Thank you so much!!

[madhavi1102]

Hi T2K-Felix, Thats great, thanks for letting me know. Regards, Madhavi

…

On Thu, 11 Apr 2024 at 3:24 PM, T2K-Felix ***@***.***> wrote: Hi @madhavi1102 <https://github.com/madhavi1102> 👋🏼, If you have already downloaded the checkpoints they are stored in /home/.cache/doctr/models otherwise you can also download specific models from the url for example db_resnet50 pytorch: https://github.com/mindee/doctr/blob/0d849152a852cb55b3fb0cc0e5e602600349d97d/doctr/models/detection/differentiable_binarization/pytorch.py#L30 — Reply to this email directly, view it on GitHub <#1350 (reply in thread)>, or unsubscribe <https://github.com/notifications/unsubscribe-auth/AKPSPCGNZRK67J5DY3HTL6LY4ZMURAVCNFSM6AAAAAA6A6ACMGVHI2DSMVQWIX3LMV43SRDJONRXK43TNFXW4Q3PNVWWK3TUHM4TAOBRHAYDC> . You are receiving this because you were mentioned.Message ID: ***@***.***>