362e4d5a67883f49d2ade090ba3d8ff8

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

• Loss: 0.4695
• Data Size: 1.0
• Epoch Runtime: 132.6341
• Accuracy: 0.8840
• F1 Macro: 0.8840
• Rouge1: 0.8841
• Rouge2: 0.0
• Rougel: 0.8842
• Rougelsum: 0.8839

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	0.7056	0	2.8677	0.4943	0.3308	0.4947	0.0	0.4943	0.4944
No log	1	3273	0.6287	0.0078	4.1048	0.7654	0.7654	0.7651	0.0	0.7654	0.7653
0.0101	2	6546	0.4770	0.0156	4.9946	0.7925	0.7898	0.7925	0.0	0.7923	0.7921
0.5026	3	9819	0.3873	0.0312	7.0555	0.8290	0.8290	0.8289	0.0	0.8290	0.8289
0.4632	4	13092	0.3882	0.0625	11.1039	0.8279	0.8276	0.8278	0.0	0.8279	0.8279
0.3886	5	16365	0.3644	0.125	19.3711	0.8463	0.8462	0.8461	0.0	0.8467	0.8460
0.3884	6	19638	0.3288	0.25	34.8811	0.8673	0.8673	0.8673	0.0	0.8675	0.8675
0.3411	7	22911	0.3233	0.5	66.3942	0.8695	0.8691	0.8693	0.0	0.8699	0.8695
0.3165	8.0	26184	0.2982	1.0	132.4479	0.8857	0.8857	0.8857	0.0	0.8858	0.8855
0.2251	9.0	29457	0.3222	1.0	133.4428	0.8869	0.8869	0.8869	0.0	0.8873	0.8869
0.1863	10.0	32730	0.3859	1.0	129.6657	0.8800	0.8797	0.8800	0.0	0.8800	0.8801
0.1468	11.0	36003	0.3707	1.0	130.5390	0.8846	0.8845	0.8846	0.0	0.8847	0.8844
0.1151	12.0	39276	0.4695	1.0	132.6341	0.8840	0.8840	0.8841	0.0	0.8842	0.8839

Framework versions

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