dbbd1bb7665a15febc12b276080bba67

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

• Loss: 0.4421
• Data Size: 1.0
• Epoch Runtime: 302.0405
• Accuracy: 0.8902
• F1 Macro: 0.8837
• Rouge1: 0.8903
• Rouge2: 0.0
• Rougel: 0.8902
• Rougelsum: 0.8902

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	Accuracy	F1 Macro	Rouge1	Rouge2	Rougel	Rougelsum
No log	0	0	0.6884	0	10.8268	0.5892	0.4365	0.5891	0.0	0.5892	0.5890
0.596	1	11370	0.4708	0.0078	13.3883	0.7725	0.7576	0.7726	0.0	0.7725	0.7724
0.446	2	22740	0.4227	0.0156	15.5708	0.7988	0.7821	0.7989	0.0	0.7988	0.7987
0.4042	3	34110	0.3827	0.0312	20.3000	0.8214	0.8090	0.8214	0.0	0.8213	0.8214
0.3752	4	45480	0.3822	0.0625	29.3705	0.8379	0.8249	0.8380	0.0	0.8378	0.8379
0.3371	5	56850	0.3176	0.125	47.5023	0.8571	0.8491	0.8572	0.0	0.8571	0.8570
0.3098	6	68220	0.3119	0.25	87.9787	0.8602	0.8532	0.8603	0.0	0.8602	0.8603
0.2661	7	79590	0.2734	0.5	155.1950	0.8777	0.8711	0.8777	0.0	0.8778	0.8777
0.2366	8.0	90960	0.2672	1.0	300.2506	0.8896	0.8821	0.8897	0.0	0.8897	0.8896
0.1945	9.0	102330	0.2853	1.0	303.6743	0.8895	0.8835	0.8896	0.0	0.8895	0.8895
0.1287	10.0	113700	0.3267	1.0	306.1309	0.8940	0.8877	0.8941	0.0	0.8941	0.8940
0.1228	11.0	125070	0.3335	1.0	300.5862	0.8916	0.8852	0.8917	0.0	0.8917	0.8915
0.0899	12.0	136440	0.4421	1.0	302.0405	0.8902	0.8837	0.8903	0.0	0.8902	0.8902

Framework versions

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