bcecba41c0c7bbceae344722efad576d

This model is a fine-tuned version of studio-ousia/luke-japanese-large-lite on the nyu-mll/glue [mnli] dataset. It achieves the following results on the evaluation set:

• Loss: 1.0969
• Data Size: 1.0
• Epoch Runtime: 2049.3016
• Accuracy: 0.3273
• F1 Macro: 0.1644
• Rouge1: 0.3273
• Rouge2: 0.0
• Rougel: 0.3275
• Rougelsum: 0.3277

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	1.0997	0	15.9463	0.3436	0.2517	0.3436	0.0	0.3436	0.3433
1.1336	1	12271	1.1152	0.0078	32.2889	0.3273	0.1644	0.3273	0.0	0.3275	0.3277
1.1114	2	24542	1.1014	0.0156	48.6794	0.3545	0.1745	0.3544	0.0	0.3545	0.3543
1.1097	3	36813	1.0973	0.0312	80.1596	0.3545	0.1745	0.3544	0.0	0.3545	0.3543
1.1033	4	49084	1.0972	0.0625	143.3028	0.3545	0.1745	0.3544	0.0	0.3545	0.3543
1.1069	5	61355	1.0965	0.125	272.7862	0.3545	0.1745	0.3544	0.0	0.3545	0.3543
1.0987	6	73626	1.1065	0.25	525.6871	0.3182	0.1609	0.3184	0.0	0.3182	0.3183
1.1013	7	85897	1.0988	0.5	1036.9807	0.3545	0.1745	0.3544	0.0	0.3545	0.3543
1.0993	8.0	98168	1.0984	1.0	2049.2134	0.3545	0.1745	0.3544	0.0	0.3545	0.3543
1.0993	9.0	110439	1.0969	1.0	2049.3016	0.3273	0.1644	0.3273	0.0	0.3275	0.3277

Framework versions

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