File size: 1,024 Bytes
d5c2739 |
1 2 3 4 5 6 7 8 9 10 11 12 13 14 |
Training dataset length:
18978
Validation dataset length:
4625
Test dataset length:
4566
Current eval performance:
Evaluate results:
{'eval_exact': 37.68648648648649, 'eval_f1': 44.0549722409684, 'eval_total': 4625, 'eval_HasAns_exact': 37.68648648648649, 'eval_HasAns_f1': 44.0549722409684, 'eval_HasAns_total': 4625, 'eval_best_exact': 37.68648648648649, 'eval_best_exact_thresh': 0.0, 'eval_best_f1': 44.0549722409684, 'eval_best_f1_thresh': 0.0, 'eval_samples': 4625}
Predict results:
{'test_exact': 35.129215943933424, 'test_f1': 42.10453725499187, 'test_total': 4566, 'test_HasAns_exact': 35.129215943933424, 'test_HasAns_f1': 42.10453725499187, 'test_HasAns_total': 4566, 'test_best_exact': 35.129215943933424, 'test_best_exact_thresh': 0.0, 'test_best_f1': 42.10453725499187, 'test_best_f1_thresh': 0.0, 'predict_samples': 4566}
Best trial:
BestRun(run_id='1', objective=87.56834325824502, hyperparameters={'learning_rate': 5e-05, 'num_train_epochs': 5, 'per_device_train_batch_size': 16, 'warmup_steps': 0})
|