Oh right there are some. Thanks Shree Le lundi 22 juillet 2019 07:12:41 UTC+2, shree a écrit : > > >But there are still a lot of things I do not understand. And one of them > is actually causing me an issue : even with a lot of iterations (475k) I > still do not see any log message with the error on the evaluation set. > At iteration 61235/475300/475526, Mean rms=0.521%, delta=2.073%, char > train=9.379%, word train=9.669%, skip ratio=0.1%, New worst char error = > 9.379 wrote checkpoint. > > Please search in your log file for `Eval Char error rate`. You will see > messages similar to following: > > Line 581: At iteration 11780/22300/22300, Mean rms=0.444%, delta=0.337%, > char train=1.106%, word train=4.349%, skip ratio=0%, New worst char error > = 1.106At iteration 1000, stage 0, Eval Char error rate=96.849627, Word > error rate=99.99418 wrote checkpoint. > Line 695: At iteration 13602/27900/27900, Mean rms=0.402%, delta=0.256%, > char train=0.822%, word train=3.451%, skip ratio=0%, New best char error = > 0.822At iteration 10735, stage 1, Eval Char error rate=98.134002, Word > error rate=99.994139 wrote best > model:./manipuri_layer_Bengali/layer0.822_13602.checkpoint wrote checkpoint. > Line 780: At iteration 14802/31900/31900, Mean rms=0.386%, delta=0.249%, > char train=0.807%, word train=3.359%, skip ratio=0%, New worst char error > = 0.807At iteration 12029, stage 1, Eval Char error rate=1.305587, Word > error rate=4.4057044 wrote checkpoint. > Line 975: At iteration 17515/41300/41300, Mean rms=0.38%, delta=0.237%, > char train=0.769%, word train=3.152%, skip ratio=0%, New worst char error > = 0.769At iteration 13789, stage 1, Eval Char error rate=1.2493521, Word > error rate=4.2435136 wrote checkpoint. > Line 1353: At iteration 22136/59200/59201, Mean rms=0.346%, delta=0.201%, > char train=0.647%, word train=2.701%, skip ratio=0%, New worst char error > = 0.647At iteration 16507, stage 1, Eval Char error rate=1.2372596, Word > error rate=4.2810121 wrote checkpoint. > Line 1480: At iteration 23655/65400/65401, Mean rms=0.338%, delta=0.2%, > char train=0.623%, word train=2.495%, skip ratio=0%, New worst char error > = 0.623At iteration 21133, stage 1, Eval Char error rate=1.0239981, Word > error rate=3.4448664 wrote checkpoint. > Line 1579: At iteration 24812/70200/70201, Mean rms=0.337%, delta=0.196%, > char train=0.641%, word train=2.505%, skip ratio=0%, New worst char error > = 0.641At iteration 22640, stage 1, Eval Char error rate=0.92986846, Word > error rate=3.2734983 wrote checkpoint. > Line 1742: At iteration 26593/78100/78101, Mean rms=0.326%, delta=0.178%, > char train=0.565%, word train=2.307%, skip ratio=0%, New worst char error > = 0.565At iteration 23791, stage 1, Eval Char error rate=0.96613336, Word > error rate=3.3449609 wrote checkpoint. > Line 1967: At iteration 29056/89300/89301, Mean rms=0.332%, delta=0.207%, > char train=0.705%, word train=2.625%, skip ratio=0%, New worst char error > = 0.705At iteration 25585, stage 1, Eval Char error rate=0.9651644, Word > error rate=3.1872666 wrote checkpoint. > Line 2153: At iteration 30993/98500/98501, Mean rms=0.301%, delta=0.158%, > char train=0.475%, word train=2.025%, skip ratio=0%, New best char error = > 0.475At iteration 27017, stage 1, Eval Char error rate=0.86442927, Word > error rate=3.0246162 wrote best > model:./manipuri_layer_Bengali/layer0.475_30993.checkpoint wrote checkpoint. > Line 2246: At iteration 32014/103100/103101, Mean rms=0.321%, delta=0.2%, > char train=0.589%, word train=2.385%, skip ratio=0%, New worst char error > = 0.589At iteration 29079, stage 1, Eval Char error rate=0.91884141, Word > error rate=3.2192128 wrote checkpoint. > Line 2377: At iteration 33395/109600/109601, Mean rms=0.323%, > delta=0.193%, char train=0.617%, word train=2.572%, skip ratio=0%, New > worst char error = 0.617At iteration 30993, stage 1, Eval Char error > rate=0.81090954, Word error rate=2.861102 wrote checkpoint. > Line 2517: At iteration 34767/116500/116501, Mean rms=0.302%, > delta=0.197%, char train=0.641%, word train=2.263%, skip ratio=0%, New > worst char error = 0.641At iteration 32139, stage 1, Eval Char error > rate=0.85321202, Word error rate=3.0337451 > > >
-- You received this message because you are subscribed to the Google Groups "tesseract-ocr" group. To unsubscribe from this group and stop receiving emails from it, send an email to tesseract-ocr+unsubscr...@googlegroups.com. To view this discussion on the web visit https://groups.google.com/d/msgid/tesseract-ocr/173803e2-8f24-4654-846d-935e90af44be%40googlegroups.com.
