4a93bcf2da11b654e629f50729ec1c13

This model is a fine-tuned version of facebook/opt-6.7b on the fancyzhx/dbpedia_14 dataset. It achieves the following results on the evaluation set:

• Loss: 0.0970
• Data Size: 1.0
• Epoch Runtime: 8475.3662
• Accuracy: 0.9828
• F1 Macro: 0.9828
• Rouge1: 0.9828
• Rouge2: 0.0
• Rougel: 0.9828
• Rougelsum: 0.9828

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	3.2732	0	226.1149	0.0779	0.0424	0.0779	0.0	0.0779	0.0778
0.4846	1	17500	0.5364	0.0078	290.8559	0.8792	0.8752	0.8792	0.0	0.8791	0.8792
0.2645	2	35000	0.2341	0.0156	368.5432	0.9366	0.9366	0.9367	0.0	0.9366	0.9366
0.1584	3	52500	0.1671	0.0312	497.0515	0.9651	0.9652	0.9651	0.0	0.9652	0.9652
0.1021	4	70000	0.0839	0.0625	753.1418	0.9774	0.9773	0.9774	0.0	0.9774	0.9774
0.0753	5	87500	0.0971	0.125	1267.0456	0.9775	0.9775	0.9775	0.0	0.9775	0.9775
0.0754	6	105000	0.0759	0.25	2295.9749	0.9816	0.9816	0.9816	0.0	0.9816	0.9816
0.0004	7	122500	0.0930	0.5	4356.0073	0.9767	0.9766	0.9767	0.0	0.9767	0.9767
0.0337	8.0	140000	0.0584	1.0	8488.3989	0.9874	0.9874	0.9874	0.0	0.9874	0.9874
0.021	9.0	157500	0.0650	1.0	8483.5547	0.9862	0.9862	0.9862	0.0	0.9862	0.9862
0.0249	10.0	175000	0.0782	1.0	8478.5897	0.9853	0.9853	0.9853	0.0	0.9853	0.9853
0.0413	11.0	192500	0.0964	1.0	8476.6622	0.9845	0.9845	0.9845	0.0	0.9845	0.9845
0.0565	12.0	210000	0.0970	1.0	8475.3662	0.9828	0.9828	0.9828	0.0	0.9828	0.9828

Framework versions

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