andi611
/

distilbert-base-uncased-qa-boolq

Text Classification Transformers PyTorch

English distilbert generated_from_trainer Inference Endpoints

Model card Files Files and versions Community

distilbert-base-uncased-qa-boolq / log.log

EC2 Default User

add model

168ef09 almost 3 years ago

raw history blame contribute delete

No virus

845 Bytes

	Training dataset length:
	8484
	Validation dataset length:
	943
	Test dataset length:
	3270
	Current performance:
	Eval:
	{'eval_loss': 0.7018890380859375, 'eval_accuracy': 0.40827147401908803, 'eval_runtime': 3.2526, 'eval_samples_per_second': 289.923, 'eval_steps_per_second': 9.223}
	Test:
	{'eval_loss': 0.7028740644454956, 'eval_accuracy': 0.39847094801223243, 'eval_runtime': 10.3848, 'eval_samples_per_second': 314.882, 'eval_steps_per_second': 9.918}
	Training complete performance:
	Eval:
	{'eval_loss': 1.1437411308288574, 'eval_accuracy': 0.7370095440084835, 'eval_runtime': 3.1329, 'eval_samples_per_second': 301.0, 'eval_steps_per_second': 9.576, 'epoch': 5.0}
	Test:
	{'eval_loss': 1.2070553302764893, 'eval_accuracy': 0.7314984709480122, 'eval_runtime': 10.6939, 'eval_samples_per_second': 305.782, 'eval_steps_per_second': 9.632, 'epoch': 5.0}