Special Character Handling

da03 / Attention-OCR

Visual Attention based OCR

MIT License

1.11k stars 362 forks source link

Special Character Handling #61

Open AdamWP opened 6 years ago

AdamWP commented 6 years ago

How does this project handle special characters such as currency symbols, apostrophes, percent symbols, etc (¢, $, %, ')? Can I just adjust the vocabulary and include these symbols in my dataset?

TARGET_VOCAB_SIZE = 26+10+3 # 0: PADDING, 1: GO, 2: EOS, >2: 0-9, a-z)

balajiwix commented 6 years ago

Please help me on including the special characters