With this file, I got the error in tr file making step.
Tesseract Open Source OCR Engine v3.05.01 with Leptonica
Page 1
row xheight=10.1471, but median xheight = 12.5
row xheight=10.1471, but median xheight = 12.5
row xheight=10, but median xheight = 12.5
row xheight=20, but median xheight = 12.5
row xheight=27, but median xheight = 12.5
row xheight=35, but median xheight = 12.5
row xheight=42, but median xheight = 12.5
FAIL!
APPLY_BOXES: boxfile line 112/y ((249,333),(257,345)): FAILURE! Couldn't find a matching blob
APPLY_BOXES:
Boxes read from boxfile: 335
Boxes failed resegmentation: 1
Found 334 good blobs.
Generated training data for 70 words
How can I fix it ?