bead8f5cd997d20d419d9d217c4a2ef7

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

• Loss: 0.6215
• Data Size: 1.0
• Epoch Runtime: 50.8919
• Accuracy: 0.6885
• F1 Macro: 0.4078
• Rouge1: 0.6895
• Rouge2: 0.0
• Rougel: 0.6885
• Rougelsum: 0.6885

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	0.7178	0	2.3216	0.3711	0.3519	0.3711	0.0	0.3701	0.3711
No log	1	267	0.6417	0.0078	2.9584	0.6885	0.4078	0.6895	0.0	0.6885	0.6885
No log	2	534	0.6363	0.0156	3.8269	0.6885	0.4078	0.6895	0.0	0.6885	0.6885
No log	3	801	0.6202	0.0312	5.2654	0.6885	0.4078	0.6895	0.0	0.6885	0.6885
No log	4	1068	0.6201	0.0625	7.0664	0.6885	0.4078	0.6895	0.0	0.6885	0.6885
0.0381	5	1335	0.6603	0.125	10.5214	0.6885	0.4078	0.6895	0.0	0.6885	0.6885
0.6168	6	1602	0.6563	0.25	17.4156	0.6885	0.4078	0.6895	0.0	0.6885	0.6885
0.6144	7	1869	0.6708	0.5	28.3239	0.6885	0.4078	0.6895	0.0	0.6885	0.6885
0.604	8.0	2136	0.6195	1.0	52.1667	0.6885	0.4078	0.6895	0.0	0.6885	0.6885
0.6096	9.0	2403	0.6451	1.0	51.0585	0.6885	0.4078	0.6895	0.0	0.6885	0.6885
0.6176	10.0	2670	0.6254	1.0	51.8813	0.6885	0.4078	0.6895	0.0	0.6885	0.6885
0.6203	11.0	2937	0.6220	1.0	50.1492	0.6885	0.4078	0.6895	0.0	0.6885	0.6885
0.6233	12.0	3204	0.6215	1.0	50.8919	0.6885	0.4078	0.6895	0.0	0.6885	0.6885

Framework versions

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