4f87dc3b31ede0d9acd709dac534be6a

This model is a fine-tuned version of distilbert/distilbert-base-cased on the nyu-mll/glue [stsb] dataset. It achieves the following results on the evaluation set:

• Loss: 0.6816
• Data Size: 1.0
• Epoch Runtime: 6.3926
• Mse: 0.6818
• Mae: 0.6268
• R2: 0.6950

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: 8
• eval_batch_size: 8
• seed: 42
• distributed_type: multi-GPU
• num_devices: 4
• total_train_batch_size: 32
• total_eval_batch_size: 32
• optimizer: Use adamw_torch with betas=(0.9,0.999) and epsilon=1e-08 and optimizer_args=No additional optimizer arguments
• lr_scheduler_type: constant
• num_epochs: 50

Training results

Training Loss	Epoch	Step	Validation Loss	Data Size	Epoch Runtime	Mse	Mae	R2
No log	0	0	7.2287	0	0.9784	7.2300	2.2572	-2.2342
No log	1	179	5.6197	0.0078	1.2760	5.6208	1.9630	-1.5144
No log	2	358	3.4773	0.0156	1.3161	3.4783	1.5637	-0.5560
No log	3	537	2.3218	0.0312	1.4039	2.3226	1.2771	-0.0390
No log	4	716	2.5156	0.0625	1.6278	2.5163	1.3010	-0.1256
No log	5	895	1.0082	0.125	2.0720	1.0086	0.8143	0.5488
0.1301	6	1074	0.9477	0.25	2.5712	0.9480	0.7495	0.5759
0.8179	7	1253	0.8762	0.5	3.8962	0.8765	0.7367	0.6079
0.5774	8.0	1432	0.8607	1.0	6.3075	0.8608	0.7323	0.6149
0.3812	9.0	1611	0.6917	1.0	6.2085	0.6919	0.6394	0.6905
0.2557	10.0	1790	0.6580	1.0	6.1720	0.6583	0.6149	0.7055
0.2068	11.0	1969	0.6559	1.0	6.0549	0.6562	0.6173	0.7065
0.1521	12.0	2148	0.6466	1.0	6.1622	0.6468	0.6148	0.7107
0.134	13.0	2327	0.6678	1.0	6.2170	0.6681	0.6328	0.7011
0.1287	14.0	2506	0.6951	1.0	6.2653	0.6953	0.6328	0.6890
0.1025	15.0	2685	0.7079	1.0	6.4028	0.7082	0.6440	0.6832
0.1012	16.0	2864	0.6816	1.0	6.3926	0.6818	0.6268	0.6950

Framework versions

• Transformers 4.57.0
• Pytorch 2.8.0+cu128
• Datasets 4.3.0
• Tokenizers 0.22.1