prediction on new dataset

[Ria312]

Hi
How to use this model for the prediction of labels on a new dataset?

Here is the command I am using :

python run_bluebert_ner.py --do_prepare=false --do_train=false --do_eval=false --do_predict=true --task_name="bc5cdr" --vocab_file=vocab.txt --bert_config_file=bert_config.json --init_checkpoint=model.ckpt-1516 --num_train_epochs=1.0 --do_lower_case=False --data_dir=no_label_dataset --output_dir=no_label_output/

I am just keeping "do_predict" true because I want to see the labels the model "model.ckpt-1516" predict after fine-tuning.

Here is the sample data without tags:

oxacalcitriol
suppresses
secondary
hyperparathyroidism
without
inducing
low
bone
turnover
in
dogs
with
renal
failure

Is it necessary to have labels on new dataset as well?

Thanks

[Ria312]

discovered how to do this.