f48eb80acb48c65a84c3f964e75342b4

This model is a fine-tuned version of albert/albert-base-v2 on the fancyzhx/dbpedia_14 dataset. It achieves the following results on the evaluation set:

• Loss: 0.0695
• Data Size: 1.0
• Epoch Runtime: 654.4659
• Accuracy: 0.9875
• F1 Macro: 0.9875
• Rouge1: 0.9875
• Rouge2: 0.0
• Rougel: 0.9875
• Rougelsum: 0.9875

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	2.7644	0	26.9719	0.0932	0.0265	0.0932	0.0	0.0931	0.0933
0.1916	1	17500	0.1133	0.0078	32.0201	0.9774	0.9776	0.9774	0.0	0.9775	0.9775
0.0981	2	35000	0.1615	0.0156	36.7241	0.9673	0.9671	0.9673	0.0	0.9674	0.9673
0.1107	3	52500	0.1060	0.0312	46.1673	0.9812	0.9812	0.9812	0.0	0.9812	0.9812
0.0757	4	70000	0.0837	0.0625	66.7804	0.9821	0.9821	0.9822	0.0	0.9822	0.9821
0.0547	5	87500	0.0957	0.125	108.1098	0.9803	0.9802	0.9803	0.0	0.9803	0.9803
0.0775	6	105000	0.0829	0.25	184.8853	0.9853	0.9854	0.9854	0.0	0.9853	0.9853
0.0006	7	122500	0.0811	0.5	341.2127	0.9811	0.9811	0.9812	0.0	0.9811	0.9812
0.0838	8.0	140000	0.0620	1.0	654.9239	0.9877	0.9877	0.9877	0.0	0.9877	0.9877
0.0376	9.0	157500	0.0833	1.0	657.6147	0.9844	0.9844	0.9844	0.0	0.9843	0.9844
0.0754	10.0	175000	0.0774	1.0	660.7885	0.9849	0.9850	0.9849	0.0	0.9849	0.9849
0.0491	11.0	192500	0.0668	1.0	659.9687	0.9861	0.9861	0.9861	0.0	0.9862	0.9861
0.0458	12.0	210000	0.0695	1.0	654.4659	0.9875	0.9875	0.9875	0.0	0.9875	0.9875

Framework versions

• Transformers 4.57.0
• Pytorch 2.8.0+cu128
• Datasets 4.0.0
• Tokenizers 0.22.1