Loading
¿Tienes un flujo de trabajo de ComfyUI? Obtén un Workflow Doc legible en tu navegador.Haz legible un flujo de trabajo.Pruébalo ahora ↗
This repository contains a ResNet + Transformer decoder checkpoint for Khmer OCR, I don’t have a public paper for this model — everything comes from thousands of experiments across different model architectures and...
Fuente del modelo
Descripción de la fuente
This repository contains a ResNet + Transformer decoder checkpoint for Khmer OCR, I don’t have a public paper for this model — everything comes from thousands of experiments across different model architectures and datasets.
pip install mer
from mer import Mer
model = Mer(markdown=True, device='cuda')
result = model.predict("sample_image.png")
print("Predicted text:", result)
Fuentes
1 fuenteVerificado 17 sept
Artefactos del modelo
1 artefactoExtractos de fuentes
2 extractos--- language: - km license: apache-2.0 tags: - ocr - transformer - vision pipeline_tag: image-to-text --- # Khmer OCR CNN + Transformer This repository contains a ResNet + Transformer decoder checkpoint for Khmer OCR, I don’t have a public paper for this model — everything comes from thousands of experiments across different model architectures and datasets. ## Installation ```python pip install mer ``` ## Usage ```python from mer import Mer model = Mer(markdown=True, device='cuda') result = model.predict("sample_image.png") print("Predicted text:", result) ```
Source context: 24 downloads · 1 likes · Pipeline image-to-text · Repo metythorn/ocr-stn-cnn-transformer-base