rubert-base-cased_pos

This model is a fine-tuned version of DeepPavlov/rubert-base-cased on the None dataset. It achieves the following results on the evaluation set:

• Loss: 0.4707
• Precision: 0.5814
• Recall: 0.5824
• F1: 0.5819
• Accuracy: 0.9021

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: 4
• eval_batch_size: 8
• seed: 42
• optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
• lr_scheduler_type: linear
• num_epochs: 100

Training results

Training Loss	Epoch	Step	Validation Loss	Precision	Recall	F1	Accuracy
No log	1.11	50	0.6626	0.0119	0.0019	0.0033	0.7605
No log	2.22	100	0.5508	0.0187	0.0193	0.0190	0.7994
No log	3.33	150	0.4129	0.1039	0.1139	0.1087	0.8451
No log	4.44	200	0.3683	0.1473	0.1448	0.1461	0.8555
No log	5.56	250	0.3231	0.2346	0.2934	0.2607	0.8687
No log	6.67	300	0.3154	0.3034	0.4498	0.3624	0.8752
No log	7.78	350	0.2665	0.3306	0.4595	0.3845	0.8931
No log	8.89	400	0.2724	0.3532	0.5734	0.4371	0.8929
No log	10.0	450	0.2694	0.4179	0.6042	0.4941	0.8997
0.3748	11.11	500	0.2984	0.4609	0.5347	0.4951	0.9051
0.3748	12.22	550	0.2823	0.4063	0.6448	0.4985	0.8926
0.3748	13.33	600	0.2894	0.4514	0.5463	0.4943	0.9022
0.3748	14.44	650	0.3150	0.5043	0.5676	0.5341	0.9043
0.3748	15.56	700	0.3608	0.4828	0.5965	0.5337	0.9018
0.3748	16.67	750	0.4073	0.4824	0.6332	0.5476	0.9003
0.3748	17.78	800	0.4018	0.5271	0.5811	0.5528	0.9083
0.3748	18.89	850	0.4261	0.528	0.6371	0.5774	0.9047
0.3748	20.0	900	0.4229	0.5187	0.6429	0.5741	0.9049
0.3748	21.11	950	0.4429	0.4511	0.6506	0.5328	0.8942
0.0787	22.22	1000	0.3901	0.5289	0.6004	0.5624	0.9102
0.0787	23.33	1050	0.3880	0.5	0.5965	0.5440	0.9122
0.0787	24.44	1100	0.3993	0.5462	0.6042	0.5738	0.9109
0.0787	25.56	1150	0.5128	0.5087	0.5097	0.5092	0.9010
0.0787	26.67	1200	0.3716	0.5492	0.6139	0.5798	0.9144
0.0787	27.78	1250	0.4023	0.5087	0.6197	0.5587	0.9093
0.0787	28.89	1300	0.4473	0.5461	0.6178	0.5797	0.9106
0.0787	30.0	1350	0.4449	0.5739	0.5772	0.5756	0.9067
0.0787	31.11	1400	0.5749	0.5466	0.5656	0.5560	0.9025
0.0787	32.22	1450	0.4480	0.5571	0.6120	0.5833	0.9125
0.0352	33.33	1500	0.4459	0.4877	0.6120	0.5428	0.9004
0.0352	34.44	1550	0.4265	0.5554	0.6293	0.5900	0.9148
0.0352	35.56	1600	0.3678	0.5983	0.5347	0.5647	0.9031
0.0352	36.67	1650	0.4182	0.5777	0.6100	0.5934	0.9120
0.0352	37.78	1700	0.4270	0.5767	0.6390	0.6062	0.9146
0.0352	38.89	1750	0.4614	0.5283	0.6486	0.5823	0.9045
0.0352	40.0	1800	0.4962	0.5553	0.5714	0.5633	0.9124
0.0352	41.11	1850	0.4498	0.5981	0.6178	0.6078	0.9169
0.0352	42.22	1900	0.4311	0.4920	0.6564	0.5624	0.9009
0.0352	43.33	1950	0.5744	0.5065	0.6023	0.5503	0.8935
0.0269	44.44	2000	0.4810	0.5325	0.5849	0.5575	0.9093
0.0269	45.56	2050	0.5074	0.5918	0.5541	0.5723	0.9064
0.0269	46.67	2100	0.5291	0.5081	0.6622	0.5750	0.8938
0.0269	47.78	2150	0.4815	0.5904	0.6429	0.6155	0.9092
0.0269	48.89	2200	0.5512	0.6294	0.5869	0.6074	0.9126
0.0269	50.0	2250	0.4686	0.5876	0.6023	0.5949	0.9094
0.0269	51.11	2300	0.4961	0.5942	0.5965	0.5954	0.9135
0.0269	52.22	2350	0.5033	0.5471	0.6390	0.5895	0.9046
0.0269	53.33	2400	0.5583	0.6308	0.5212	0.5708	0.9021
0.0269	54.44	2450	0.4128	0.5904	0.5927	0.5915	0.9003
0.0268	55.56	2500	0.4908	0.5219	0.5985	0.5576	0.9059
0.0268	56.67	2550	0.4669	0.5068	0.6506	0.5697	0.9004
0.0268	57.78	2600	0.4738	0.5692	0.6351	0.6004	0.9004
0.0268	58.89	2650	0.4933	0.6197	0.6197	0.6197	0.9185
0.0268	60.0	2700	0.4623	0.6060	0.6236	0.6147	0.9141
0.0268	61.11	2750	0.4472	0.5981	0.6178	0.6078	0.9163
0.0268	62.22	2800	0.4988	0.6410	0.6274	0.6341	0.9132
0.0268	63.33	2850	0.4900	0.6176	0.5830	0.5998	0.9065
0.0268	64.44	2900	0.4380	0.6410	0.6100	0.6251	0.9143
0.0268	65.56	2950	0.5097	0.6115	0.6564	0.6331	0.9130
0.0251	66.67	3000	0.3672	0.5845	0.6409	0.6114	0.9088
0.0251	67.78	3050	0.4630	0.5213	0.6371	0.5734	0.9106
0.0251	68.89	3100	0.4418	0.6375	0.6042	0.6204	0.9134
0.0251	70.0	3150	0.4648	0.5705	0.6718	0.6170	0.9122
0.0251	71.11	3200	0.4314	0.6388	0.5907	0.6138	0.9138
0.0251	72.22	3250	0.4000	0.5575	0.6178	0.5861	0.9121
0.0251	73.33	3300	0.3548	0.5039	0.6216	0.5566	0.9145
0.0251	74.44	3350	0.4886	0.6616	0.5849	0.6209	0.9124
0.0251	75.56	3400	0.5781	0.4402	0.6602	0.5282	0.8803
0.0251	76.67	3450	0.3379	0.5704	0.6178	0.5931	0.9088
0.0271	77.78	3500	0.4527	0.5303	0.6429	0.5812	0.9016
0.0271	78.89	3550	0.5293	0.6211	0.5792	0.5994	0.9085
0.0271	80.0	3600	0.6486	0.3684	0.6351	0.4663	0.8263

Framework versions

• Transformers 4.38.2
• Pytorch 2.1.2
• Datasets 2.1.0
• Tokenizers 0.15.2