908ac3876256a23d7215469f4dc3e128

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

• Loss: 1.2793
• Data Size: 1.0
• Epoch Runtime: 19.3505
• Accuracy: 0.7692
• F1 Macro: 0.8090
• Rouge1: 0.7699
• Rouge2: 0.0
• Rougel: 0.7692
• Rougelsum: 0.7685

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	1.6040	0	1.8148	0.2266	0.1813	0.2266	0.0	0.2266	0.2266
No log	1	178	1.4686	0.0078	2.9580	0.3246	0.1691	0.3253	0.0	0.3260	0.3246
No log	2	356	1.2141	0.0156	3.0739	0.6357	0.4354	0.6357	0.0	0.6364	0.6357
No log	3	534	0.9243	0.0312	3.4456	0.6662	0.5176	0.6662	0.0	0.6669	0.6662
No log	4	712	0.8206	0.0625	4.1527	0.7109	0.5406	0.7124	0.0	0.7116	0.7109
No log	5	890	0.7427	0.125	5.1000	0.7102	0.5405	0.7109	0.0	0.7109	0.7102
0.0531	6	1068	0.7274	0.25	7.1081	0.7209	0.5787	0.7216	0.0	0.7216	0.7209
0.5836	7	1246	0.6455	0.5	11.1110	0.7557	0.7376	0.7564	0.0	0.7557	0.7557
0.4986	8.0	1424	0.6025	1.0	19.5262	0.7614	0.7763	0.7621	0.0	0.7621	0.7614
0.3578	9.0	1602	0.6960	1.0	18.8277	0.7614	0.7845	0.7621	0.0	0.7614	0.7614
0.2094	10.0	1780	0.7427	1.0	19.0056	0.7706	0.8069	0.7706	0.0	0.7713	0.7706
0.1496	11.0	1958	0.9641	1.0	19.2184	0.7592	0.7989	0.7596	0.0	0.7599	0.7589
0.0997	12.0	2136	1.2793	1.0	19.3505	0.7692	0.8090	0.7699	0.0	0.7692	0.7685

Framework versions

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