Update README.md

README.md

@@ -0,0 +1,3 @@
+This model was trained based on the naver-clova-ix/donut-base model on Russian-language letters to extract the date and number of the letter. 
+Model metrics based on test data:
+Total number of samples: 193, Tree Edit Distance (TED) based accuracy score: 0.781453704848682, F1 accuracy score: 0.6020260492040521