bert_base_lda_20_v1_book_stsb

This model is a fine-tuned version of gokulsrinivasagan/bert_base_lda_20_v1_book on the GLUE STSB dataset. It achieves the following results on the evaluation set:

• Loss: 0.6650
• Pearson: 0.8407
• Spearmanr: 0.8382
• Combined Score: 0.8394

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
2.8738	1.0	23	2.4670	0.1765	0.1748	0.1756
1.4719	2.0	46	1.0280	0.7397	0.7404	0.7401
0.9801	3.0	69	0.8276	0.7956	0.7954	0.7955
0.783	4.0	92	0.7431	0.8197	0.8193	0.8195
0.5677	5.0	115	0.9075	0.8135	0.8152	0.8144
0.4407	6.0	138	0.7474	0.8267	0.8272	0.8269
0.3821	7.0	161	0.6753	0.8391	0.8371	0.8381
0.3036	8.0	184	0.8726	0.8246	0.8260	0.8253
0.269	9.0	207	0.7331	0.8311	0.8293	0.8302
0.2191	10.0	230	0.7562	0.8383	0.8368	0.8375
0.1854	11.0	253	0.7022	0.8365	0.8343	0.8354
0.1718	12.0	276	0.6650	0.8407	0.8382	0.8394
0.1685	13.0	299	0.7270	0.8350	0.8333	0.8342
0.1368	14.0	322	0.7532	0.8392	0.8376	0.8384
0.1351	15.0	345	0.8710	0.8379	0.8379	0.8379
0.1459	16.0	368	0.7801	0.8416	0.8398	0.8407
0.106	17.0	391	0.6833	0.8393	0.8380	0.8387

Framework versions

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