search

imatge-upc / slt_how2sign_wicv2023

Sign Language Translation for Instructional Videos - CVPR WiCV 2023
MIT License
35 stars 9 forks source link

Problem while training the corresponding sentencepiece model #4

Closed rabeya-akter closed 1 year ago

rabeya-akter commented 1 year ago

I am facing a problem while "Training the corresponding sentencepiece model".

As per instructions, I made the .env file like this:

FAIRSEQ_ROOT: /home/rmedu/Music/RabeyaAkter[SK-092-015]/thesis/Codes/slt_how2sign_wicv2023 SAVE_DIR: /home/rmedu/Music/RabeyaAkter[SK-092-015]/thesis/Dataset/tsvFile(cvpr23) VOCAB_SIZE: 7000 FEATS: i3d PARTITION: train

As per the dataset structure, in my SAVE_DIR path I have these files:

├── cvpr23.fairseq.i3d.train.how2sign.tsv ├── cvpr23.fairseq.i3d.val.how2sign.tsv └── cvpr23.fairseq.i3d.test.how2sign.tsv

But I am getting this error everytime:

(slt-how2sign-wicv2023) rmedu@rmedu-desktop:~/Music/RabeyaAkter[SK-092-015]/thesis/Codes/slt_how2sign_wicv2023/examples/sign_language$ task how2sign:train_sentencepiece_lowercased task: Specify the SAVE_DIR in the .env file. task: precondition not met

Can you tell me what am I doing wrong? What exactly should I change to solve this error?

FangSen9000 commented 1 year ago

Can you share your.env file, I'm having these damn problems too！

rabeya-akter commented 1 year ago

for this, you need a copy of the train csv file renamed as cvpr23.fairseq.mediapipe.train.how2sign.tsv and saved it in the slt_how2sign_wicv2023, that is in the root folder. This will work then.