juruomp/rat-sql-tc

Code for our paper "Pay More Attention to History: A Context Modeling Strategy for Conversational Text-to-SQL"

RAT-SQL-TC

Code for our paper "Pay More Attention to History: A Context Modeling Strategy for Conversational Text-to-SQL"

Please refer to gap-text2sql rep for environment and requirements.

Preprocess dataset

python run.py preprocess experiments/sparc-tcs-configs/gap-run.jsonnet

Inference

python run.py eval experiments/sparc-tcs-configs/gap-run.jsonnet

Training

python run.py train experiments/sparc-tcs-configs/gap-run.jsonnet

rat-sql-tc's Issues

Problem with starting training

Hi developer of RAT-SQL-TC
These days I've been trying to run your code on my machine but encountered a problem when training, here is the traceback in my PyCharm:

Traceback (most recent call last):
  File "/home/xfy/NLP/RAT-SQL-TC/rat-sql-gap/run.py", line 107, in <module>
    main()
  File "/home/xfy/NLP/RAT-SQL-TC/rat-sql-gap/run.py", line 69, in main
    train.main(train_config)
  File "/home/xfy/NLP/RAT-SQL-TC/rat-sql-gap/seq2struct/commands/train.py", line 324, in main
    trainer = Trainer(logger, config)
  File "/home/xfy/NLP/RAT-SQL-TC/rat-sql-gap/seq2struct/commands/train.py", line 101, in __init__
    self.model_preproc.load()
  File "/home/xfy/NLP/RAT-SQL-TC/rat-sql-gap/seq2struct/models/enc_dec.py", line 56, in load
    self.enc_preproc.load()
  File "/home/xfy/NLP/RAT-SQL-TC/rat-sql-gap/seq2struct/models/spider/spider_enc.py", line 1646, in load
    self.tokenizer = BartTokenizer.from_pretrained(self.data_dir if enc_dir is None else enc_dir)
  File "/home/xfy/CondaEnvs/text2sql/lib/python3.8/site-packages/transformers/tokenization_utils_base.py", line 1768, in from_pretrained
    raise EnvironmentError(
OSError: Can't load tokenizer for 'data/sparc-bart-tcs/sparc,nl2code-1115,output_from=true,fs=2,emb=bart,cvlink/enc'. If you were trying to load it from 'https://huggingface.co/models', make sure you don't have a local directory with the same name. Otherwise, make sure 'data/sparc-bart-tcs/sparc,nl2code-1115,output_from=true,fs=2,emb=bart,cvlink/enc' is the correct path to a directory containing all relevant files for a BartTokenizer tokenizer.

The problem seems to have something to do with the BartTokenizer file missing on huggingface because PuDB returns something else as shown in the picture below.

Any idea on how to fix this?