tinybert_base_train_book_ent_15p_s_init_book_stsb

This model is a fine-tuned version of gokulsrinivasagan/tinybert_base_train_book_ent_15p_s_init_book on the GLUE STSB dataset. It achieves the following results on the evaluation set:

• Loss: 1.0337
• Pearson: 0.7643
• Spearmanr: 0.7639
• Combined Score: 0.7641

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
2.9343	1.0	23	2.5829	0.0657	0.0433	0.0545
1.8406	2.0	46	1.9884	0.4642	0.4337	0.4489
1.1799	3.0	69	1.5395	0.6493	0.6670	0.6581
0.8711	4.0	92	1.4130	0.7134	0.7290	0.7212
0.7099	5.0	115	1.8762	0.7019	0.7462	0.7240
0.5987	6.0	138	1.6049	0.7365	0.7621	0.7493
0.5083	7.0	161	1.2573	0.7443	0.7553	0.7498
0.4407	8.0	184	1.4610	0.7501	0.7657	0.7579
0.3764	9.0	207	1.1276	0.7477	0.7498	0.7487
0.3617	10.0	230	1.1506	0.7639	0.7664	0.7652
0.3201	11.0	253	1.1589	0.7611	0.7652	0.7632
0.2911	12.0	276	1.2315	0.7604	0.7661	0.7633
0.2563	13.0	299	1.0337	0.7643	0.7639	0.7641
0.2677	14.0	322	1.1795	0.7602	0.7636	0.7619
0.2379	15.0	345	1.2357	0.7590	0.7641	0.7615
0.2167	16.0	368	1.3301	0.7558	0.7614	0.7586
0.2085	17.0	391	1.2452	0.7679	0.7707	0.7693
0.1924	18.0	414	1.2505	0.7645	0.7672	0.7659

Framework versions

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