WTQ fine-tune - Githubissues

google-research / tapas

End-to-end neural table-text understanding models.

Apache License 2.0

1.15k stars 217 forks source link

WTQ fine-tune #42

Closed lairikeqiA closed 4 years ago

lairikeqiA commented 4 years ago

I plan to fine-tune WTQ on pre-trained SQA model. How do I create the correct format of WTQ?

ghost commented 4 years ago

Hi!

This should be explained in the README, basically the following command should create the data in interaction and TF example format:

python3 tapas/run_task_main.py \
  --task="WTQ" \
  --input_dir="${wtq_data_dir}" \
  --output_dir="${output_dir}" \
  --bert_vocab_file="${tapas_data_dir}/vocab.txt" \
  --mode="create_data"