tinybert_base_train_book_ent_15p_s_init_kd_a_in_stsb

This model is a fine-tuned version of gokulsrinivasagan/tinybert_base_train_book_ent_15p_s_init_kd_a_in on the GLUE STSB dataset. It achieves the following results on the evaluation set:

• Loss: 0.7920
• Pearson: 0.8137
• Spearmanr: 0.8098
• Combined Score: 0.8117

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
2.7558	1.0	23	2.5348	0.0801	0.0885	0.0843
1.7861	2.0	46	1.4064	0.6507	0.6311	0.6409
1.1688	3.0	69	1.0797	0.7300	0.7220	0.7260
0.9278	4.0	92	1.3977	0.7547	0.7668	0.7607
0.7682	5.0	115	0.9325	0.7896	0.7847	0.7872
0.6375	6.0	138	0.9133	0.7935	0.7949	0.7942
0.5372	7.0	161	0.9057	0.8036	0.8019	0.8027
0.4744	8.0	184	1.0945	0.8039	0.8066	0.8052
0.4393	9.0	207	0.8419	0.8062	0.8037	0.8050
0.3847	10.0	230	0.8400	0.8115	0.8085	0.8100
0.3565	11.0	253	0.8999	0.8135	0.8099	0.8117
0.3359	12.0	276	0.9316	0.8143	0.8113	0.8128
0.2988	13.0	299	0.7920	0.8137	0.8098	0.8117
0.2798	14.0	322	0.9671	0.8085	0.8075	0.8080
0.2582	15.0	345	0.9492	0.8141	0.8103	0.8122
0.2469	16.0	368	0.8195	0.8165	0.8136	0.8151
0.2384	17.0	391	0.8370	0.8149	0.8103	0.8126
0.2041	18.0	414	0.8979	0.8135	0.8086	0.8111

Framework versions

• Transformers 4.51.2
• Pytorch 2.6.0+cu126
• Datasets 3.5.0
• Tokenizers 0.21.1