/IITD-OCR

Inference scripts for IITD OCR docker images.

Primary LanguageShell

IITD - Scene Text Recognition and Optical Character Recognition

Hindi - Scene Text Recognition (Word Level)

Docker-Image

  • Pull the docker image: docker pull iitdnltm/str:iitd_hin
  • Download weights_hin.ckpt in mounted_volume/model
  • Download the demo.zip and extract in mounted_volume/data
  • run usage_scrips/load.sh to load the docker image into a container
./usage_scripts <modality> <language> <version> <full path to image dir>

# example`
./usage_scripts scene_text hindi v1 /home/username/data/demo

English - Scene Text Recognition (Word Level)

Docker-Image

    • Pull the docker image: docker pull iitdnltm/str:iitd_eng
  • Download weights_eng.ckpt in mounted_volume/model
  • Download the demo.zip and extract in mounted_volume/data
  • run usage_scrips/load.sh to load the docker image into a container
./usage_scripts <modality> <language> <version> <full path to image dir>

# example`
./usage_scripts scene_text english v1 /home/username/data/demo

Urdu - Optical Character Recognition (Line level)

Docker-Image

  • Pull the docker image: docker pull iitdnltm/ocr:iitd_urdu
  • Download weights_urdu.ckpt in mounted_volume/model
  • Download the demo.zip and extract in mounted_volume/data
  • run usage_scrips/load.sh to load the docker image into a container
./usage_scripts <modality> <language> <version> <full path to image dir>

# example`
./usage_scripts printed urdu v1 /home/username/data/demo