MedRuRobertaLarge_neg

This model is a fine-tuned version of DmitryPogrebnoy/MedRuRobertaLarge on the None dataset. It achieves the following results on the evaluation set:

• Loss: 0.6996
• Precision: 0.5225
• Recall: 0.5788
• F1: 0.5492
• Accuracy: 0.8955

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	1.0	50	0.6906	0.0175	0.0019	0.0035	0.7724
No log	2.0	100	0.7240	0.0526	0.0019	0.0037	0.7756
No log	3.0	150	0.5668	0.0407	0.0289	0.0338	0.7668
No log	4.0	200	0.4358	0.1326	0.1522	0.1417	0.8236
No log	5.0	250	0.3509	0.1932	0.2177	0.2047	0.8573
No log	6.0	300	0.2961	0.3339	0.3699	0.3510	0.8862
No log	7.0	350	0.3715	0.4073	0.3642	0.3845	0.8820
No log	8.0	400	0.2778	0.4511	0.4528	0.4519	0.9040
No log	9.0	450	0.3318	0.4576	0.4778	0.4675	0.8997
0.4025	10.0	500	0.3198	0.5278	0.5299	0.5288	0.9049
0.4025	11.0	550	0.3157	0.4297	0.6358	0.5128	0.8909
0.4025	12.0	600	0.3024	0.5548	0.5954	0.5743	0.9188
0.4025	13.0	650	0.3670	0.6091	0.6185	0.6138	0.9149
0.4025	14.0	700	0.4036	0.5088	0.6127	0.5559	0.8998
0.4025	15.0	750	0.4116	0.5542	0.6012	0.5767	0.9085
0.4025	16.0	800	0.3971	0.5301	0.6455	0.5821	0.9095
0.4025	17.0	850	0.4887	0.5535	0.5183	0.5353	0.8977
0.4025	18.0	900	0.4385	0.5563	0.6474	0.5984	0.9106
0.4025	19.0	950	0.4007	0.6316	0.6012	0.6160	0.9219
0.0841	20.0	1000	0.3720	0.5709	0.5896	0.5801	0.9165
0.0841	21.0	1050	0.5100	0.6393	0.6012	0.6197	0.9150
0.0841	22.0	1100	0.5028	0.5319	0.6590	0.5886	0.8972
0.0841	23.0	1150	0.4347	0.5656	0.5896	0.5774	0.9149
0.0841	24.0	1200	0.4721	0.5861	0.6031	0.5945	0.9122
0.0841	25.0	1250	0.5677	0.6457	0.5549	0.5969	0.9116
0.0841	26.0	1300	0.4095	0.6278	0.6435	0.6356	0.9189
0.0841	27.0	1350	0.4633	0.5088	0.6686	0.5779	0.8989
0.0841	28.0	1400	0.3649	0.5617	0.6493	0.6023	0.9105
0.0841	29.0	1450	0.4653	0.5633	0.6262	0.5931	0.9111
0.0464	30.0	1500	0.5159	0.5581	0.6474	0.5995	0.9119
0.0464	31.0	1550	0.4562	0.5248	0.6513	0.5813	0.9090
0.0464	32.0	1600	0.4424	0.5665	0.5742	0.5703	0.9173
0.0464	33.0	1650	0.4866	0.5617	0.5703	0.5660	0.9164
0.0464	34.0	1700	0.4313	0.3760	0.4586	0.4132	0.8986
0.0464	35.0	1750	0.3786	0.5218	0.5761	0.5476	0.9093

Framework versions

• Transformers 4.40.1
• Pytorch 2.2.1+cu121
• Datasets 2.19.1
• Tokenizers 0.19.1