a65d9c4adf27447664d82dde48cc24a9

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

• Loss: 1.0307
• Data Size: 1.0
• Epoch Runtime: 41.2299
• Accuracy: 0.7777
• F1 Macro: 0.8141
• Rouge1: 0.7784
• Rouge2: 0.0
• Rougel: 0.7784
• Rougelsum: 0.7784

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	1.5650	0	3.2521	0.2777	0.1115	0.2770	0.0	0.2770	0.2777
No log	1	178	1.5858	0.0078	4.1440	0.2862	0.1376	0.2869	0.0	0.2862	0.2862
No log	2	356	1.2355	0.0156	4.7329	0.5185	0.3133	0.5192	0.0	0.5192	0.5185
No log	3	534	0.8361	0.0312	5.9672	0.6612	0.5153	0.6612	0.0	0.6619	0.6605
No log	4	712	0.9015	0.0625	7.7237	0.7145	0.5549	0.7152	0.0	0.7152	0.7145
No log	5	890	0.8214	0.125	10.9602	0.7322	0.5678	0.7330	0.0	0.7330	0.7322
0.0544	6	1068	0.6937	0.25	15.6494	0.7415	0.7021	0.7422	0.0	0.7422	0.7415
0.5963	7	1246	0.7073	0.5	23.6016	0.7244	0.6791	0.7251	0.0	0.7251	0.7244
0.5197	8.0	1424	0.6662	1.0	43.3845	0.7635	0.7399	0.7642	0.0	0.7642	0.7635
0.4196	9.0	1602	0.6533	1.0	40.7272	0.7756	0.7870	0.7770	0.0	0.7763	0.7756
0.328	10.0	1780	0.7369	1.0	41.2215	0.7607	0.7993	0.7614	0.0	0.7607	0.7607
0.2367	11.0	1958	1.0076	1.0	41.9694	0.7031	0.7347	0.7031	0.0	0.7031	0.7038
0.1884	12.0	2136	1.0821	1.0	40.9372	0.7678	0.8077	0.7678	0.0	0.7685	0.7685
0.2027	13.0	2314	1.0307	1.0	41.2299	0.7777	0.8141	0.7784	0.0	0.7784	0.7784

Framework versions

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