gpt2-ner-invoiceSenderRecipient_all_inv_03_01

This model was trained from scratch on an unknown dataset. It achieves the following results on the evaluation set:

• Loss: 0.0307
• Precision: 0.7932
• Recall: 0.8488
• F1: 0.8201
• Accuracy: 0.9895

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: 4
• eval_batch_size: 4
• seed: 42
• optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
• lr_scheduler_type: linear
• num_epochs: 1
• mixed_precision_training: Native AMP

Training results

Training Loss	Epoch	Step	Validation Loss	Precision	Recall	F1	Accuracy
0.0363	0.01	500	0.0338	0.7846	0.7969	0.7907	0.9884
0.0392	0.02	1000	0.0346	0.7665	0.8211	0.7929	0.9881
0.0363	0.04	1500	0.0347	0.7701	0.8075	0.7884	0.9880
0.0396	0.05	2000	0.0347	0.7454	0.8375	0.7888	0.9879
0.0366	0.06	2500	0.0350	0.7519	0.8345	0.7911	0.9879
0.0382	0.07	3000	0.0356	0.7500	0.8434	0.7939	0.9877
0.0424	0.09	3500	0.0358	0.7517	0.8287	0.7883	0.9877
0.0385	0.1	4000	0.0352	0.7605	0.8225	0.7903	0.9880
0.0382	0.11	4500	0.0361	0.7494	0.8159	0.7813	0.9874
0.0372	0.12	5000	0.0345	0.7817	0.8044	0.7929	0.9885
0.0377	0.14	5500	0.0346	0.7749	0.8238	0.7986	0.9884
0.0383	0.15	6000	0.0359	0.7568	0.8341	0.7936	0.9879
0.0372	0.16	6500	0.0356	0.7548	0.8356	0.7932	0.9879
0.0371	0.17	7000	0.0352	0.7540	0.8477	0.7981	0.9880
0.0368	0.19	7500	0.0349	0.7662	0.8310	0.7973	0.9881
0.0388	0.2	8000	0.0339	0.7648	0.8336	0.7977	0.9883
0.0368	0.21	8500	0.0336	0.7729	0.8305	0.8006	0.9886
0.0389	0.22	9000	0.0340	0.7750	0.8208	0.7972	0.9884
0.0384	0.24	9500	0.0349	0.7549	0.8499	0.7996	0.9880
0.0376	0.25	10000	0.0358	0.7531	0.8390	0.7938	0.9875
0.0354	0.26	10500	0.0346	0.7650	0.8318	0.7970	0.9882
0.0358	0.27	11000	0.0338	0.7694	0.8397	0.8030	0.9886
0.0389	0.28	11500	0.0341	0.7586	0.8502	0.8018	0.9882
0.0383	0.3	12000	0.0342	0.7688	0.8275	0.7971	0.9881
0.0355	0.31	12500	0.0337	0.7783	0.8281	0.8024	0.9885
0.0372	0.32	13000	0.0338	0.7703	0.8399	0.8036	0.9884
0.0369	0.33	13500	0.0331	0.7683	0.8427	0.8038	0.9886
0.0361	0.35	14000	0.0336	0.7699	0.8322	0.7999	0.9885
0.0361	0.36	14500	0.0336	0.7735	0.8390	0.8049	0.9885
0.0372	0.37	15000	0.0333	0.7747	0.8343	0.8034	0.9887
0.0366	0.38	15500	0.0343	0.7646	0.8468	0.8036	0.9883
0.0345	0.4	16000	0.0333	0.7790	0.8334	0.8053	0.9887
0.0363	0.41	16500	0.0329	0.7783	0.8301	0.8034	0.9887
0.0348	0.42	17000	0.0341	0.7626	0.8533	0.8054	0.9884
0.0391	0.43	17500	0.0324	0.7873	0.8295	0.8079	0.9889
0.0344	0.45	18000	0.0334	0.7769	0.8369	0.8058	0.9887
0.0378	0.46	18500	0.0337	0.7741	0.8394	0.8054	0.9886
0.035	0.47	19000	0.0328	0.7827	0.8323	0.8067	0.9888
0.0351	0.48	19500	0.0327	0.7815	0.8371	0.8083	0.9889
0.037	0.5	20000	0.0328	0.7793	0.8388	0.8079	0.9888
0.0346	0.51	20500	0.0325	0.7804	0.8416	0.8099	0.9890
0.0364	0.52	21000	0.0323	0.7861	0.8339	0.8093	0.9889
0.0356	0.53	21500	0.0327	0.7729	0.8510	0.8101	0.9889
0.0346	0.54	22000	0.0325	0.7791	0.8407	0.8087	0.9889
0.0342	0.56	22500	0.0334	0.7790	0.8443	0.8104	0.9889
0.0368	0.57	23000	0.0322	0.7869	0.8323	0.8089	0.9890
0.0371	0.58	23500	0.0320	0.7890	0.8356	0.8116	0.9891
0.0344	0.59	24000	0.0321	0.7910	0.8321	0.8110	0.9892
0.0342	0.61	24500	0.0319	0.7881	0.8356	0.8111	0.9892
0.0339	0.62	25000	0.0320	0.7889	0.8317	0.8097	0.9892
0.0347	0.63	25500	0.0316	0.7909	0.8347	0.8122	0.9892
0.034	0.64	26000	0.0318	0.7887	0.8324	0.8100	0.9891
0.0347	0.66	26500	0.0317	0.7791	0.8525	0.8141	0.9891
0.0345	0.67	27000	0.0318	0.7870	0.8384	0.8119	0.9892
0.0347	0.68	27500	0.0317	0.7903	0.8426	0.8157	0.9893
0.0371	0.69	28000	0.0311	0.7965	0.8332	0.8144	0.9894
0.0338	0.71	28500	0.0316	0.7863	0.8442	0.8142	0.9892
0.0352	0.72	29000	0.0315	0.7810	0.8537	0.8157	0.9892
0.0344	0.73	29500	0.0314	0.7953	0.8353	0.8148	0.9894
0.0322	0.74	30000	0.0320	0.7836	0.8449	0.8131	0.9891
0.0355	0.76	30500	0.0312	0.7877	0.8480	0.8167	0.9894
0.035	0.77	31000	0.0313	0.7864	0.8504	0.8171	0.9893
0.0346	0.78	31500	0.0310	0.7931	0.8424	0.8170	0.9895
0.0339	0.79	32000	0.0316	0.7857	0.8501	0.8166	0.9893
0.033	0.8	32500	0.0311	0.7975	0.8406	0.8185	0.9895
0.0337	0.82	33000	0.0314	0.7886	0.8457	0.8162	0.9894
0.0357	0.83	33500	0.0311	0.7923	0.8437	0.8172	0.9894
0.0348	0.84	34000	0.0312	0.7909	0.8490	0.8189	0.9894
0.0343	0.85	34500	0.0311	0.7856	0.8528	0.8179	0.9893
0.0323	0.87	35000	0.0311	0.7884	0.8505	0.8183	0.9894
0.0329	0.88	35500	0.0307	0.7981	0.8399	0.8185	0.9896
0.0324	0.89	36000	0.0313	0.7830	0.8576	0.8186	0.9893
0.0336	0.9	36500	0.0312	0.7836	0.8566	0.8185	0.9893
0.0327	0.92	37000	0.0309	0.7887	0.8501	0.8182	0.9895
0.0338	0.93	37500	0.0312	0.7887	0.8514	0.8188	0.9894
0.0327	0.94	38000	0.0311	0.7873	0.8534	0.8190	0.9894
0.0326	0.95	38500	0.0308	0.7953	0.8459	0.8198	0.9895
0.0338	0.97	39000	0.0307	0.7932	0.8488	0.8201	0.9895
0.0354	0.98	39500	0.0308	0.7916	0.8502	0.8198	0.9895
0.0313	0.99	40000	0.0309	0.7897	0.8523	0.8198	0.9895

Framework versions

• Transformers 4.22.0
• Pytorch 1.10.0
• Tokenizers 0.12.1