7bb248eb0a90ee3032ff727f1a3fb4b6

This model is a fine-tuned version of distilbert/distilbert-base-cased-distilled-squad on the dim/tldr_news dataset. It achieves the following results on the evaluation set:

• Loss: 1.3727
• Data Size: 1.0
• Epoch Runtime: 5.8038
• Accuracy: 0.7450
• F1 Macro: 0.7856
• Rouge1: 0.7450
• Rouge2: 0.0
• Rougel: 0.7457
• Rougelsum: 0.7450

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	1.6380	0	1.0012	0.1626	0.0865	0.1626	0.0	0.1634	0.1619
No log	1	178	1.4706	0.0078	2.4445	0.2891	0.1366	0.2894	0.0	0.2891	0.2891
No log	2	356	1.3269	0.0156	1.1852	0.5163	0.3474	0.5174	0.0	0.5170	0.5163
No log	3	534	1.0773	0.0312	1.2838	0.6179	0.4305	0.6186	0.0	0.6179	0.6179
No log	4	712	0.8334	0.0625	1.5986	0.6783	0.4884	0.6797	0.0	0.6790	0.6783
No log	5	890	0.8016	0.125	1.8825	0.6832	0.5358	0.6839	0.0	0.6847	0.6832
0.058	6	1068	0.7021	0.25	2.3723	0.7358	0.7453	0.7365	0.0	0.7358	0.7358
0.568	7	1246	0.6156	0.5	3.4701	0.7649	0.7901	0.7656	0.0	0.7656	0.7649
0.4685	8.0	1424	0.6116	1.0	5.9947	0.7592	0.7863	0.7599	0.0	0.7592	0.7592
0.3151	9.0	1602	0.7018	1.0	5.8370	0.7507	0.7892	0.7514	0.0	0.7514	0.7507
0.1761	10.0	1780	0.8993	1.0	5.9271	0.75	0.7851	0.75	0.0	0.7507	0.7493
0.0911	11.0	1958	1.1811	1.0	5.8396	0.7521	0.7909	0.7521	0.0	0.7521	0.7521
0.0688	12.0	2136	1.3727	1.0	5.8038	0.7450	0.7856	0.7450	0.0	0.7457	0.7450

Framework versions

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