f1c0fcadc7f3a2165eb6f6194914c9f6

This model is a fine-tuned version of studio-ousia/luke-large-lite on the fancyzhx/dbpedia_14 dataset. It achieves the following results on the evaluation set:

• Loss: 2.6537
• Data Size: 0.5
• Epoch Runtime: 1556.8986
• Accuracy: 0.0714
• F1 Macro: 0.0095
• Rouge1: 0.0714
• Rouge2: 0.0
• Rougel: 0.0715
• Rougelsum: 0.0715

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	2.6894	0	106.9529	0.0531	0.0134	0.0531	0.0	0.0532	0.0531
0.116	1	17500	0.0852	0.0078	129.4599	0.9825	0.9825	0.9825	0.0	0.9825	0.9825
0.0701	2	35000	0.1344	0.0156	152.3257	0.9779	0.9779	0.9779	0.0	0.9779	0.9779
0.0851	3	52500	0.0850	0.0312	197.5330	0.9854	0.9854	0.9855	0.0	0.9854	0.9854
0.163	4	70000	0.0961	0.0625	289.3644	0.9840	0.9839	0.9840	0.0	0.9840	0.9840
0.1108	5	87500	0.0957	0.125	468.1463	0.9854	0.9854	0.9854	0.0	0.9854	0.9855
1.9298	6	105000	2.4605	0.25	834.7894	0.1431	0.1041	0.1431	0.0	0.1430	0.1431
0.015	7	122500	2.6537	0.5	1556.8986	0.0714	0.0095	0.0714	0.0	0.0715	0.0715

Framework versions

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