gpt2-finetuned-ner

This model is a fine-tuned version of gpt2 on an unknown dataset. It achieves the following results on the evaluation set:

• Loss: 0.6671
• Precision: 0.6051
• Recall: 0.5135
• F1: 0.5556
• Accuracy: 0.8962

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: 2e-05
• 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: 50

Training results

Training Loss	Epoch	Step	Validation Loss	Precision	Recall	F1	Accuracy
No log	1.0	23	2.0585	0.0448	0.2703	0.0769	0.2815
No log	2.0	46	0.7904	0.0	0.0	0.0	0.8444
No log	3.0	69	0.6555	0.24	0.0324	0.0571	0.8450
No log	4.0	92	0.6181	0.3504	0.2216	0.2715	0.8534
No log	5.0	115	0.5701	0.4483	0.2811	0.3455	0.8671
No log	6.0	138	0.5416	0.4818	0.3568	0.4099	0.8742
No log	7.0	161	0.5311	0.4641	0.3838	0.4201	0.8748
No log	8.0	184	0.5186	0.4938	0.4270	0.4580	0.8800
No log	9.0	207	0.5574	0.5448	0.3946	0.4577	0.8859
No log	10.0	230	0.5245	0.5226	0.4378	0.4765	0.8839
No log	11.0	253	0.5400	0.5436	0.4378	0.4850	0.8898
No log	12.0	276	0.5290	0.4880	0.4378	0.4615	0.8826
No log	13.0	299	0.5426	0.5467	0.4432	0.4896	0.8891
No log	14.0	322	0.5524	0.5714	0.4541	0.5060	0.8898
No log	15.0	345	0.5444	0.5802	0.5081	0.5418	0.8936
No log	16.0	368	0.5383	0.5706	0.5027	0.5345	0.8917
No log	17.0	391	0.5427	0.5839	0.5081	0.5434	0.8936
No log	18.0	414	0.5573	0.5758	0.5135	0.5429	0.8930
No log	19.0	437	0.5695	0.5860	0.4973	0.5380	0.8943
No log	20.0	460	0.5657	0.5802	0.5081	0.5418	0.8943
No log	21.0	483	0.5815	0.6025	0.5243	0.5607	0.8962
0.6194	22.0	506	0.5836	0.5860	0.4973	0.5380	0.8936
0.6194	23.0	529	0.5937	0.6259	0.4973	0.5542	0.8969
0.6194	24.0	552	0.6034	0.6184	0.5081	0.5579	0.8975
0.6194	25.0	575	0.5804	0.6038	0.5189	0.5581	0.8969
0.6194	26.0	598	0.6122	0.625	0.5135	0.5638	0.8988
0.6194	27.0	621	0.6078	0.5987	0.5081	0.5497	0.8969
0.6194	28.0	644	0.6027	0.5988	0.5243	0.5591	0.8962
0.6194	29.0	667	0.6138	0.6	0.5189	0.5565	0.8988
0.6194	30.0	690	0.6118	0.6012	0.5297	0.5632	0.8975
0.6194	31.0	713	0.6226	0.6051	0.5135	0.5556	0.8962
0.6194	32.0	736	0.6149	0.5975	0.5135	0.5523	0.8962
0.6194	33.0	759	0.6184	0.5949	0.5081	0.5481	0.8923
0.6194	34.0	782	0.6251	0.6025	0.5243	0.5607	0.8975
0.6194	35.0	805	0.6379	0.6129	0.5135	0.5588	0.8969
0.6194	36.0	828	0.6439	0.6062	0.5243	0.5623	0.8969
0.6194	37.0	851	0.6494	0.6026	0.5081	0.5513	0.8962
0.6194	38.0	874	0.6426	0.6090	0.5135	0.5572	0.8969
0.6194	39.0	897	0.6328	0.6012	0.5297	0.5632	0.8975
0.6194	40.0	920	0.6517	0.6078	0.5027	0.5503	0.8949
0.6194	41.0	943	0.6582	0.6115	0.5189	0.5614	0.8962
0.6194	42.0	966	0.6556	0.5901	0.5135	0.5491	0.8943
0.6194	43.0	989	0.6571	0.6	0.5189	0.5565	0.8956
0.0843	44.0	1012	0.6609	0.5988	0.5243	0.5591	0.8962
0.0843	45.0	1035	0.6661	0.6087	0.5297	0.5665	0.8975
0.0843	46.0	1058	0.6597	0.5951	0.5243	0.5575	0.8962
0.0843	47.0	1081	0.6582	0.5963	0.5189	0.5549	0.8949
0.0843	48.0	1104	0.6609	0.5938	0.5135	0.5507	0.8943
0.0843	49.0	1127	0.6665	0.6051	0.5135	0.5556	0.8962
0.0843	50.0	1150	0.6671	0.6051	0.5135	0.5556	0.8962

Framework versions

• Transformers 4.37.2
• Pytorch 2.1.0+cu121
• Datasets 2.16.1
• Tokenizers 0.15.1