I have use quickstart .

[cyjack]

but i don't know how to give sentence score .dose it need word-leval score before do that .I use wmt2020 model from kiwi as the baseline.
http://www.statmt.org/wmt20/quality-estimation-task.html
the model execute but just give me that
{'batch_size': 64,
'config': 'predict_estimator.yaml',
'debug': False,
'experiment_name': 'predict-predest',
'gpu_id': 0,
'load_data': None,
'load_model': 'best_model.torch',
'load_vocab': None,
'log_interval': 100,
'mlflow_always_log_artifacts': False,
'mlflow_tracking_uri': 'mlruns/',
'model': 'estimator',
'output_dir': 'peace&love',
'quiet': False,
'run_name': None,
'run_uuid': None,
'save_config': None,
'save_data': None,
'seed': 42}
Local output directory is: peace&love
Predict with the PredEst (Predictor-Estimator) model
{'batch_size': 64,
'config': 'predict_estimator.yaml',
'debug': False,
'experiment_name': 'predict-predest',
'gpu_id': -1,
'load_data': None,
'load_model': 'best_model.torch',
'load_vocab': None,
'log_interval': 100,
'mlflow_always_log_artifacts': False,
'mlflow_tracking_uri': 'mlruns/',
'model': 'estimator',
'output_dir': 'peace&love',
'quiet': False,
'run_name': None,
'run_uuid': None,
'save_config': None,
'save_data': None,
'seed': 42}
Local output directory is: peace&love
Predict with the PredEst (Predictor-Estimator) model
Loaded vocabularies from best_model.torch
Is it wrong ,so stop?

[captainvera]

Sorry, I didn't see this issue earlier.

Do you mind sharing your predict_estimator.yaml file?

[daandouwe]

Closed due to inactivity.