README.md

metadata

license: mit
library_name: peft
tags:
  - generated_from_trainer
base_model: facebook/bart-large-mnli
metrics:
  - f1
  - precision
  - recall
  - accuracy
model-index:
  - name: results
    results: []

results

This model is a fine-tuned version of facebook/bart-large-mnli on an unknown dataset. It achieves the following results on the evaluation set:

• Loss: 0.1726
• F1: 0.9412
• Precision: 0.9524
• Recall: 0.9302
• Accuracy: 0.9333

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: 0.0002
• train_batch_size: 16
• eval_batch_size: 16
• seed: 42
• optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
• lr_scheduler_type: linear
• num_epochs: 5

Training results

Training Loss	Epoch	Step	Validation Loss	F1	Precision	Recall	Accuracy
0.5809	0.14	10	0.3912	0.8671	0.8621	0.8721	0.8467
0.3659	0.28	20	0.3689	0.8790	0.9718	0.8023	0.8733
0.3805	0.42	30	0.2890	0.9133	0.9080	0.9186	0.9
0.4068	0.56	40	0.3249	0.9068	0.9733	0.8488	0.9
0.3183	0.69	50	0.2801	0.9133	0.9080	0.9186	0.9
0.1929	0.83	60	0.2832	0.9123	0.9176	0.9070	0.9
0.2861	0.97	70	0.2883	0.9195	0.9091	0.9302	0.9067
0.209	1.11	80	0.3000	0.9222	0.9506	0.8953	0.9133
0.2192	1.25	90	0.2845	0.9176	0.9286	0.9070	0.9067
0.3116	1.39	100	0.2520	0.9249	0.9195	0.9302	0.9133
0.2512	1.53	110	0.2650	0.9222	0.9506	0.8953	0.9133
0.1774	1.67	120	0.2571	0.9231	0.9398	0.9070	0.9133
0.1126	1.81	130	0.2668	0.9364	0.9310	0.9419	0.9267
0.2379	1.94	140	0.3075	0.9012	0.9605	0.8488	0.8933
0.2753	2.08	150	0.2254	0.9240	0.9294	0.9186	0.9133
0.1727	2.22	160	0.2707	0.9310	0.9205	0.9419	0.92
0.224	2.36	170	0.3118	0.9057	0.9863	0.8372	0.9
0.2056	2.5	180	0.2673	0.9302	0.9302	0.9302	0.92
0.2274	2.64	190	0.2515	0.9302	0.9302	0.9302	0.92
0.1193	2.78	200	0.2250	0.9357	0.9412	0.9302	0.9267
0.2806	2.92	210	0.2268	0.9286	0.9512	0.9070	0.92
0.1272	3.06	220	0.2031	0.9349	0.9518	0.9186	0.9267
0.1879	3.19	230	0.1730	0.9480	0.9425	0.9535	0.94
0.1341	3.33	240	0.1867	0.9419	0.9419	0.9419	0.9333
0.1376	3.47	250	0.2628	0.9341	0.9630	0.9070	0.9267
0.1599	3.61	260	0.2484	0.9405	0.9634	0.9186	0.9333
0.1899	3.75	270	0.1847	0.9480	0.9425	0.9535	0.94
0.0828	3.89	280	0.1869	0.9412	0.9524	0.9302	0.9333
0.1025	4.03	290	0.1876	0.9349	0.9518	0.9186	0.9267
0.118	4.17	300	0.1811	0.9419	0.9419	0.9419	0.9333
0.1475	4.31	310	0.1901	0.9294	0.9405	0.9186	0.92
0.1354	4.44	320	0.1805	0.9357	0.9412	0.9302	0.9267
0.1444	4.58	330	0.1706	0.9540	0.9432	0.9651	0.9467
0.1068	4.72	340	0.1693	0.9480	0.9425	0.9535	0.94
0.0875	4.86	350	0.1715	0.9412	0.9524	0.9302	0.9333
0.0922	5.0	360	0.1726	0.9412	0.9524	0.9302	0.9333

Framework versions

• PEFT 0.9.0
• Transformers 4.39.1
• Pytorch 2.2.1+cu121
• Datasets 2.18.0
• Tokenizers 0.15.2