3049d3cc37cae19539c1bdeeb0e2dab1

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

• Loss: 0.7780
• Data Size: 1.0
• Epoch Runtime: 94.3879
• Accuracy: 0.8550
• F1 Macro: 0.8548
• Rouge1: 0.8553
• Rouge2: 0.0
• Rougel: 0.8553
• Rougelsum: 0.8548

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	0.6939	0	1.9927	0.4910	0.4778	0.4912	0.0	0.4904	0.4912
No log	1	3273	0.6843	0.0078	2.9495	0.6717	0.6711	0.6715	0.0	0.6719	0.6719
0.0107	2	6546	0.4865	0.0156	3.5294	0.7836	0.7835	0.7836	0.0	0.7833	0.7836
0.5169	3	9819	0.4279	0.0312	4.8784	0.8081	0.8081	0.8085	0.0	0.8079	0.8079
0.4564	4	13092	0.4063	0.0625	7.4690	0.8305	0.8305	0.8303	0.0	0.8303	0.8301
0.4036	5	16365	0.3824	0.125	13.7943	0.8318	0.8318	0.8320	0.0	0.8316	0.8316
0.4042	6	19638	0.3622	0.25	23.0014	0.8454	0.8453	0.8454	0.0	0.8456	0.8452
0.3326	7	22911	0.3351	0.5	43.5618	0.8515	0.8508	0.8518	0.0	0.8518	0.8513
0.2792	8.0	26184	0.3220	1.0	88.1046	0.8774	0.8774	0.8776	0.0	0.8776	0.8774
0.1838	9.0	29457	0.3696	1.0	91.1398	0.8680	0.8680	0.8684	0.0	0.8683	0.8679
0.1425	10.0	32730	0.4428	1.0	92.0602	0.8645	0.8645	0.8645	0.0	0.8647	0.8643
0.0967	11.0	36003	0.5786	1.0	92.1231	0.8579	0.8579	0.8579	0.0	0.8581	0.8579
0.0825	12.0	39276	0.7780	1.0	94.3879	0.8550	0.8548	0.8553	0.0	0.8553	0.8548

Framework versions

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