distilhubert-finetuned-minds14

This model is a fine-tuned version of ntu-spml/distilhubert on the MINDS14 dataset. It achieves the following results on the evaluation set:

• Loss: 3.4050
• Accuracy: 0.5263

Model description

More information needed

Intended uses & limitations

More information needed

Training and evaluation data

More information needed

Training procedure

Training hyperparameters

The following hyperparameters were used during training:

• learning_rate: 5e-05
• train_batch_size: 2
• eval_batch_size: 2
• seed: 42
• optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
• lr_scheduler_type: linear
• lr_scheduler_warmup_ratio: 0.1
• num_epochs: 50
• mixed_precision_training: Native AMP

Training results

Training Loss	Epoch	Step	Validation Loss	Accuracy
2.6434	1.0	253	2.6398	0.0702
2.5451	2.0	506	2.6434	0.0877
2.5303	3.0	759	2.6009	0.0702
2.5841	4.0	1012	2.5143	0.1053
2.1201	5.0	1265	2.3397	0.1579
1.6421	6.0	1518	2.1145	0.3509
1.5245	7.0	1771	1.9039	0.4386
1.1713	8.0	2024	1.6985	0.3684
1.1484	9.0	2277	1.5798	0.5088
0.6891	10.0	2530	1.7365	0.4912
0.3124	11.0	2783	1.8768	0.4561
0.077	12.0	3036	2.2095	0.4386
0.0167	13.0	3289	2.3894	0.4912
0.1748	14.0	3542	2.1305	0.5789
0.0366	15.0	3795	2.2102	0.5614
0.0021	16.0	4048	2.2237	0.5614
0.0012	17.0	4301	2.3768	0.5263
0.0009	18.0	4554	2.6185	0.4912
0.0006	19.0	4807	2.5854	0.5263
0.0005	20.0	5060	2.6191	0.5965
0.0004	21.0	5313	2.6767	0.5789
0.0004	22.0	5566	2.7203	0.5965
0.0003	23.0	5819	2.6451	0.5965
0.0003	24.0	6072	2.6883	0.5965
0.0002	25.0	6325	2.7872	0.5789
0.0002	26.0	6578	2.8503	0.5789
0.0002	27.0	6831	2.8895	0.5789
0.0001	28.0	7084	2.8882	0.5789
0.0001	29.0	7337	2.8726	0.5439
0.0001	30.0	7590	2.8971	0.5614
0.0001	31.0	7843	2.9427	0.5614
0.0001	32.0	8096	3.0154	0.5439
0.0001	33.0	8349	3.0109	0.5439
0.0001	34.0	8602	3.0281	0.5439
0.0001	35.0	8855	3.0510	0.5439
0.0001	36.0	9108	3.1110	0.5439
0.0001	37.0	9361	3.1634	0.5439
0.0	38.0	9614	3.1704	0.5263
0.0	39.0	9867	3.2145	0.5263
0.0	40.0	10120	3.2405	0.5439
0.0	41.0	10373	3.2725	0.5263
0.0	42.0	10626	3.2861	0.5263
0.0	43.0	10879	3.3515	0.5263
0.0	44.0	11132	3.3364	0.5263
0.0	45.0	11385	3.3570	0.5263
0.0	46.0	11638	3.3776	0.5263
0.0	47.0	11891	3.3857	0.5263
0.0	48.0	12144	3.3985	0.5263
0.0	49.0	12397	3.4012	0.5263
0.0	50.0	12650	3.4050	0.5263

Framework versions

• Transformers 4.35.2
• Pytorch 2.1.0+cu118
• Datasets 2.15.0
• Tokenizers 0.15.0