feat: add ckpts for DB resnet50 in torch

[charlesmindee]

This PR adds ckpts for DB resnet50 model in torch.

Performances below:

FUNSD
Text Detection - Recall: 79.99%, Precision: 87.22%, Mean IoU: 71.29%
Text Recognition - Accuracy: 85.85% (unicase: 86.76%)
OCR - Recall: 67.72% (unicase: 68.28%), Precision: 73.84% (unicase: 74.45%), Mean IoU: 71.29%

CORD
Text Detection - Recall: 93.00%, Precision: 91.37%, Mean IoU: 76.68%
Text Recognition - Accuracy: 92.53% (unicase: 92.92%)
OCR - Recall: 84.82% (unicase: 85.19%), Precision: 83.33% (unicase: 83.69%), Mean IoU: 76.68%


RECEIPTS
Text Detection - Recall: 86.04%, Precision: 89.94%, Mean IoU: 76.12%
Text Recognition - Accuracy: 91.70% (unicase: 92.45%)
OCR - Recall: 78.22% (unicase: 78.94%), Precision: 81.77% (unicase: 82.52%), Mean IoU: 76.12%

IDS
Text Detection - Recall: 69.46%, Precision: 74.21%, Mean IoU: 61.29%
Text Recognition - Accuracy: 64.81% (unicase: 67.50%)
OCR - Recall: 46.61% (unicase: 49.14%), Precision: 49.80% (unicase: 52.50%), Mean IoU: 61.29%

INVOICES
Text Detection - Recall: 68.28%, Precision: 72.05%, Mean IoU: 62.69%
Text Recognition - Accuracy: 90.12% (unicase: 91.59%)
OCR - Recall: 62.31% (unicase: 63.18%), Precision: 65.75% (unicase: 66.67%), Mean IoU: 62.69%

TAX FORM US
Text Detection - Recall: 77.49%, Precision: 91.13%, Mean IoU: 70.62%
Text Recognition - Accuracy: 84.30% (unicase: 84.83%)
OCR - Recall: 75.46% (unicase: 75.78%), Precision: 88.74% (unicase: 89.11%), Mean IoU: 70.62%

[fg-mindee]

Thanks!

[codecov]

Codecov Report

Merging #489 (c493fb1) into main (8718adf) will increase coverage by 0.02%.
The diff coverage is n/a.

@@            Coverage Diff             @@
##             main     #489      +/-   ##
==========================================
+ Coverage   94.99%   95.02%   +0.02%     
==========================================
  Files         108      108              
  Lines        4177     4177              
==========================================
+ Hits         3968     3969       +1     
+ Misses        209      208       -1     

Flag	Coverage Δ
unittests	95.02% <ø> (+0.02%)	⬆️

Flags with carried forward coverage won't be shown. Click here to find out more.

Impacted Files	Coverage Δ
...s/detection/differentiable_binarization/pytorch.py	97.43% <ø> (ø)
...dels/detection/differentiable_binarization/base.py	91.97% <0.00%> (+0.61%)	⬆️

---

Continue to review full report at Codecov.

Legend - Click here to learn more
Δ = absolute <relative> (impact), ø = not affected, ? = missing data
Powered by Codecov. Last update 8718adf...c493fb1. Read the comment docs.