a3470200b5880cada2b4db84cc65881b

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

• Loss: 1.2736
• Data Size: 1.0
• Epoch Runtime: 6.3421
• Accuracy: 0.7607
• F1 Macro: 0.8046
• Rouge1: 0.7610
• Rouge2: 0.0
• Rougel: 0.7607
• Rougelsum: 0.7607

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	1.6508	0	0.9951	0.1946	0.0928	0.1946	0.0	0.1960	0.1946
No log	1	178	1.5589	0.0078	2.4032	0.2401	0.0774	0.2401	0.0	0.2408	0.2393
No log	2	356	1.4107	0.0156	1.3138	0.4624	0.2648	0.4624	0.0	0.4624	0.4624
No log	3	534	1.1003	0.0312	1.5566	0.6300	0.4925	0.6300	0.0	0.6314	0.6293
No log	4	712	0.8489	0.0625	1.8027	0.6918	0.5311	0.6932	0.0	0.6918	0.6925
No log	5	890	0.7869	0.125	2.0470	0.7031	0.5414	0.7045	0.0	0.7038	0.7031
0.0575	6	1068	0.6957	0.25	2.6734	0.7195	0.5777	0.7195	0.0	0.7202	0.7195
0.5671	7	1246	0.6207	0.5	3.8896	0.7557	0.7406	0.7564	0.0	0.7557	0.7557
0.475	8.0	1424	0.6020	1.0	6.2343	0.7557	0.7659	0.7564	0.0	0.7557	0.7557
0.3048	9.0	1602	0.7315	1.0	5.9812	0.7528	0.7910	0.7536	0.0	0.7536	0.7528
0.1719	10.0	1780	0.8484	1.0	6.3332	0.7557	0.7950	0.7564	0.0	0.7560	0.7557
0.1121	11.0	1958	1.0948	1.0	6.2567	0.7614	0.7879	0.7621	0.0	0.7624	0.7607
0.0738	12.0	2136	1.2736	1.0	6.3421	0.7607	0.8046	0.7610	0.0	0.7607	0.7607

Framework versions

• Transformers 4.57.0
• Pytorch 2.8.0+cu128
• Datasets 4.0.0
• Tokenizers 0.22.1