b7262fa149ff03e43e70efb1b2f302f7

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

• Loss: 0.7725
• Data Size: 1.0
• Epoch Runtime: 18.4839
• Accuracy: 0.7308
• F1 Macro: 0.7558
• Rouge1: 0.7315
• Rouge2: 0.0
• Rougel: 0.7315
• Rougelsum: 0.7308

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	1.7408	0	1.8678	0.0185	0.0073	0.0185	0.0	0.0185	0.0185
No log	1	178	1.4953	0.0078	3.4006	0.2401	0.0774	0.2401	0.0	0.2408	0.2393
No log	2	356	1.4353	0.0156	2.3438	0.4268	0.2599	0.4268	0.0	0.4268	0.4261
No log	3	534	1.2380	0.0312	2.7681	0.3743	0.2331	0.3743	0.0	0.3743	0.3736
No log	4	712	1.1336	0.0625	3.2938	0.4609	0.2664	0.4624	0.0	0.4602	0.4602
No log	5	890	1.0869	0.125	4.3544	0.4716	0.2762	0.4723	0.0	0.4716	0.4716
0.0746	6	1068	1.0781	0.25	6.3793	0.5859	0.4217	0.5866	0.0	0.5859	0.5859
0.8633	7	1246	0.9063	0.5	10.2914	0.6335	0.4645	0.6349	0.0	0.6335	0.6342
0.7234	8.0	1424	0.7850	1.0	18.4186	0.6982	0.6855	0.6982	0.0	0.6982	0.6982
0.6556	9.0	1602	0.6991	1.0	18.3814	0.7344	0.7234	0.7351	0.0	0.7344	0.7344
0.5771	10.0	1780	0.7003	1.0	18.1894	0.7273	0.7200	0.7273	0.0	0.7273	0.7273
0.5291	11.0	1958	0.7648	1.0	19.0359	0.7259	0.7324	0.7259	0.0	0.7259	0.7259
0.4595	12.0	2136	0.7981	1.0	18.2395	0.7152	0.7331	0.7152	0.0	0.7159	0.7159
0.391	13.0	2314	0.7725	1.0	18.4839	0.7308	0.7558	0.7315	0.0	0.7315	0.7308

Framework versions

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