6a2530e6c868df18680f41c14d5ef6f0

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

• Loss: 0.6165
• Data Size: 1.0
• Epoch Runtime: 1143.7028
• Accuracy: 0.7916
• F1 Macro: 0.7923
• Rouge1: 0.7913
• Rouge2: 0.0
• Rougel: 0.7916
• Rougelsum: 0.7916

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	1.1016	0	8.4589	0.3288	0.1972	0.3290	0.0	0.3288	0.3290
1.1039	1	12271	1.0478	0.0078	18.3370	0.4819	0.4702	0.4821	0.0	0.4817	0.4817
0.833	2	24542	0.7488	0.0156	26.2226	0.6867	0.6845	0.6868	0.0	0.6868	0.6869
0.7266	3	36813	0.6842	0.0312	43.0689	0.7215	0.7178	0.7215	0.0	0.7214	0.7216
0.6783	4	49084	0.6413	0.0625	76.7296	0.7430	0.7426	0.7430	0.0	0.7431	0.7430
0.6121	5	61355	0.5804	0.125	143.1177	0.7653	0.7639	0.7653	0.0	0.7655	0.7652
0.5942	6	73626	0.6098	0.25	277.7222	0.7555	0.7558	0.7556	0.0	0.7555	0.7553
0.5069	7	85897	0.5428	0.5	568.1862	0.7815	0.7803	0.7814	0.0	0.7812	0.7816
0.4843	8.0	98168	0.5297	1.0	1148.5348	0.7963	0.7962	0.7962	0.0	0.7964	0.7962
0.4279	9.0	110439	0.5177	1.0	1145.6533	0.8040	0.8036	0.8041	0.0	0.8042	0.8041
0.3805	10.0	122710	0.5437	1.0	1141.6211	0.7984	0.7976	0.7984	0.0	0.7984	0.7984
0.3417	11.0	134981	0.5861	1.0	1141.4024	0.7929	0.7932	0.7929	0.0	0.7929	0.7928
0.2841	12.0	147252	0.6009	1.0	1143.8513	0.8018	0.8012	0.8016	0.0	0.8018	0.8017
0.3038	13.0	159523	0.6165	1.0	1143.7028	0.7916	0.7923	0.7913	0.0	0.7916	0.7916

Framework versions

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