donut
donut copied to clipboard
clovaai
Official Implementation of OCR-free Document Understanding Transformer (Donut) and Synthetic Document Generator (SynthDoG), ECCV 2022
I'm trying to train a model on a custom dataset. The training begins, but stops at the moment of generating the dataset. The last lines that are printed to the...
I also hope to see the config/.yaml file, as I can't get any results when I train it myself.
Previously, when I trained on Japanese data, the predictions only displayed or the same character. Then, I downgraded the transformers version to 4.25.1, which resulted in a "NotImplementedError: Make sure...
During the training process, the generated Japanese documents can only recognize . Here is the error report; how can I resolve this?Some weights of the model checkpoint at donut-base-finetuned-docvqa-main were...
I recently fine-tuned a model Donut for DocVQA. The fine-tuning process completed successfully, but I encountered an issue during inference. When I ask a question that should correspond to a...
I don't know if there was a problem with the data processing or the metadata.jsonl file was created incorrectly. I found that the input_ids input to the donut model contained...
Please find the sample document. I want to use Donut to convert the document into text. Please provide the sample code or example function for converting the Japanese document into...
Hi I was trying to generate Hindi Synthdog. I am facing some issues with the generation, even tried with the en config . I was hoping if anybody can help
We use that format to organize the data, but we get an error： dataset_name ├── test │ ├── metadata.jsonl │ ├── {image_path0} │ ├── {image_path1} │ . │ . ├──...
When I train cord-v2（the pre-trained model used to train cord-v2 is the donut-base model.）, the validation process shows: ：Prediction: 17, 17, 17, 17, 17, 17, 17, 17, 17, 17, 17,...
