spoken-to-signed-translation icon indicating copy to clipboard operation
spoken-to-signed-translation copied to clipboard

Enhancing Pipeline with Vietnamese Language Support

Open nqkhanh2002 opened this issue 1 year ago • 4 comments
trafficstars

Hi @AmitMY Continuing the problem from previous issues and PRs, After I can create a dummy lexicon and get a csv file like the image below image I have pip install all . and run the command text_to_gloss_to_pose --text "Xin Chao." --glosser "simple" --lexicon "spoken-to-signed-translation/assets/vn_sign" --spoken-language "vi" --signed-language "vn_sign" --pose "quick_test.pose"

and get the error, am I getting an error that does not support any vietnamese language

usage: text_to_gloss_to_pose [-h] --text TEXT --glosser {simple,spacylemma,rules,nmt} --spoken-language {de,fr,it,en} --signed-language {sgg,gsg,bfi} --lexicon LEXICON --pose POSE
text_to_gloss_to_pose: error: argument --spoken-language: invalid choice: 'vi' (choose from 'de', 'fr', 'it', 'en')

Thank you for your enthusiastic support @AmitMY

nqkhanh2002 avatar Jun 29 '24 18:06 nqkhanh2002

the error is because we don't support this language, since we did not test it.

if you want to add it, you need to add it https://github.com/sign-language-processing/spoken-to-signed-translation/blob/main/spoken_to_signed/bin.py#L65

Also, check out your paths: image some back slashes, some forward slashes

AmitMY avatar Jun 29 '24 20:06 AmitMY

Thanks @AmitMY ,

After fixing many errors, I got stuck at the step (as shown below). I went to the lookup_sequence error function, so I wanted to ask.

  • This function is looking for the glosses column in index.csv, right? Because as shown in the image above, I have that column blank.
  • And when I edit thelookup_sequencefunction and python -m pip install . The program will automatically rebuild the code as before, losing my changes.

image image

nqkhanh2002 avatar Jun 30 '24 05:06 nqkhanh2002

Hi @AmitMY , Do you have an update on this bug, how about I spend some time re-reading the project wiki. Thank you for your support

nqkhanh2002 avatar Jul 06 '24 04:07 nqkhanh2002

Hey there. I recently broke both hands so I'll be a lot slower to reply. What I see in your example is that you're using the simple lemmatizer that breaks your word into two. You should try using the spacy lemmatizer

When you reinstall, it should install with your changes unless you're in the wrong directory.

AmitMY avatar Jul 06 '24 10:07 AmitMY