du33169/t5-base-finetuned-GLUE-STSB

STSB

This model is a fine-tuned version of google-t5/t5-base on the GLUE STSB dataset. It achieves the following results on the evaluation set:

• Loss: 0.5496
• Pearson: 0.8882
• Spearmanr: 0.8872
• Combined Score: 0.8877

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: 32
• eval_batch_size: 8
• seed: 42
• optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
• lr_scheduler_type: linear
• num_epochs: 10.0

Training results

Training Loss	Epoch	Step	Combined Score	Validation Loss	Pearson	Spearmanr
No log	1.0	180	0.8180	1.1720	0.8128	0.8233
No log	2.0	360	0.8588	0.7424	0.8585	0.8591
1.0195	3.0	540	0.8756	0.6313	0.8756	0.8756
1.0195	4.0	720	0.8803	0.5849	0.8801	0.8806
1.0195	5.0	900	0.8833	0.6234	0.8838	0.8827
0.315	6.0	1080	0.8859	0.6469	0.8864	0.8854
0.315	7.0	1260	0.8861	0.5571	0.8866	0.8856
0.315	8.0	1440	0.8869	0.5629	0.8877	0.8862
0.2087	9.0	1620	0.8877	0.5569	0.8882	0.8871
0.2087	10.0	1800	0.8877	0.5496	0.8882	0.8872

Framework versions

• Transformers 4.43.3
• Pytorch 1.11.0+cu113
• Datasets 2.20.0
• Tokenizers 0.19.1