bert_tiny_lda_5_v1_book_stsb

This model is a fine-tuned version of gokulsrinivasagan/bert_tiny_lda_5_v1_book on the GLUE STSB dataset. It achieves the following results on the evaluation set:

• Loss: 0.8189
• Pearson: 0.8027
• Spearmanr: 0.7988
• Combined Score: 0.8007

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: 256
• eval_batch_size: 256
• seed: 10
• optimizer: Use adamw_torch with betas=(0.9,0.999) and epsilon=1e-08 and optimizer_args=No additional optimizer arguments
• lr_scheduler_type: linear
• num_epochs: 50

Training results

Training Loss	Epoch	Step	Validation Loss	Pearson	Spearmanr	Combined Score
2.8706	1.0	23	2.4625	0.0991	0.1000	0.0995
1.7766	2.0	46	1.3387	0.6421	0.6371	0.6396
1.1649	3.0	69	0.9883	0.7530	0.7470	0.7500
0.8446	4.0	92	0.9501	0.7802	0.7827	0.7814
0.6587	5.0	115	0.9120	0.7839	0.7824	0.7831
0.5587	6.0	138	0.8918	0.7876	0.7880	0.7878
0.5104	7.0	161	0.9523	0.7822	0.7828	0.7825
0.4506	8.0	184	0.9421	0.7916	0.7923	0.7919
0.4174	9.0	207	1.0925	0.7876	0.7937	0.7906
0.3229	10.0	230	0.8792	0.7969	0.7946	0.7958
0.3096	11.0	253	0.9665	0.7887	0.7884	0.7886
0.2918	12.0	276	0.8212	0.8017	0.7992	0.8004
0.2479	13.0	299	0.8791	0.8055	0.8037	0.8046
0.2409	14.0	322	0.8217	0.7970	0.7930	0.7950
0.2071	15.0	345	0.8929	0.7998	0.7971	0.7984
0.1907	16.0	368	0.8189	0.8027	0.7988	0.8007
0.1825	17.0	391	0.9393	0.7924	0.7899	0.7912
0.1634	18.0	414	0.8722	0.8046	0.8023	0.8035
0.1531	19.0	437	0.9629	0.8037	0.8023	0.8030
0.153	20.0	460	1.0063	0.7971	0.7947	0.7959
0.1409	21.0	483	0.9757	0.7968	0.7942	0.7955

Framework versions

• Transformers 4.46.3
• Pytorch 2.2.1+cu118
• Datasets 2.17.0
• Tokenizers 0.20.3