rubert-tiny2-srl

This model is a fine-tuned version of cointegrated/rubert-tiny2 on an unknown dataset. It achieves the following results on the evaluation set:

• Loss: 0.1428
• Addressee Precision: 0.6364
• Addressee Recall: 0.875
• Addressee F1: 0.7368
• Addressee Number: 8
• Benefactive Precision: 0.0
• Benefactive Recall: 0.0
• Benefactive F1: 0.0
• Benefactive Number: 2
• Causator Precision: 0.9286
• Causator Recall: 0.8125
• Causator F1: 0.8667
• Causator Number: 16
• Cause Precision: 0.6
• Cause Recall: 0.25
• Cause F1: 0.3529
• Cause Number: 12
• Contrsubject Precision: 0.6364
• Contrsubject Recall: 0.4118
• Contrsubject F1: 0.5
• Contrsubject Number: 17
• Deliberative Precision: 1.0
• Deliberative Recall: 0.6667
• Deliberative F1: 0.8
• Deliberative Number: 6
• Destinative Precision: 1.0
• Destinative Recall: 0.5
• Destinative F1: 0.6667
• Destinative Number: 4
• Directivefinal Precision: 1.0
• Directivefinal Recall: 1.0
• Directivefinal F1: 1.0
• Directivefinal Number: 2
• Experiencer Precision: 0.8018
• Experiencer Recall: 0.9368
• Experiencer F1: 0.8641
• Experiencer Number: 95
• Instrument Precision: 0.0
• Instrument Recall: 0.0
• Instrument F1: 0.0
• Instrument Number: 3
• Limitative Precision: 0.0
• Limitative Recall: 0.0
• Limitative F1: 0.0
• Limitative Number: 1
• Object Precision: 0.7589
• Object Recall: 0.8
• Object F1: 0.7789
• Object Number: 240
• Overall Precision: 0.7724
• Overall Recall: 0.7857
• Overall F1: 0.7790
• Overall Accuracy: 0.9589

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: 8.017672397578385e-05
• train_batch_size: 4
• eval_batch_size: 1
• seed: 678943
• gradient_accumulation_steps: 4
• total_train_batch_size: 16
• optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
• lr_scheduler_type: linear
• lr_scheduler_warmup_ratio: 0.04
• num_epochs: 4

Training results

Training Loss	Epoch	Step	Validation Loss	Addressee Precision	Addressee Recall	Addressee F1	Addressee Number	Benefactive Precision	Benefactive Recall	Benefactive F1	Benefactive Number	Causator Precision	Causator Recall	Causator F1	Causator Number	Cause Precision	Cause Recall	Cause F1	Cause Number	Contrsubject Precision	Contrsubject Recall	Contrsubject F1	Contrsubject Number	Deliberative Precision	Deliberative Recall	Deliberative F1	Deliberative Number	Destinative Precision	Destinative Recall	Destinative F1	Destinative Number	Directivefinal Precision	Directivefinal Recall	Directivefinal F1	Directivefinal Number	Experiencer Precision	Experiencer Recall	Experiencer F1	Experiencer Number	Instrument Precision	Instrument Recall	Instrument F1	Instrument Number	Limitative Precision	Limitative Recall	Limitative F1	Limitative Number	Object Precision	Object Recall	Object F1	Object Number	Overall Precision	Overall Recall	Overall F1	Overall Accuracy
0.2206	1.0	490	0.1959	0.6667	0.25	0.3636	8	0.0	0.0	0.0	2	0.8667	0.8125	0.8387	16	0.0	0.0	0.0	12	1.0	0.0588	0.1111	17	0.0	0.0	0.0	6	0.0	0.0	0.0	4	0.0	0.0	0.0	2	0.7203	0.8947	0.7981	95	0.0	0.0	0.0	3	0.0	0.0	0.0	1	0.6692	0.725	0.696	240	0.6927	0.6773	0.6849	0.9445
0.1507	2.0	981	0.1492	0.5556	0.625	0.5882	8	0.0	0.0	0.0	2	0.8667	0.8125	0.8387	16	0.6	0.25	0.3529	12	0.75	0.3529	0.48	17	1.0	0.1667	0.2857	6	1.0	0.25	0.4	4	1.0	1.0	1.0	2	0.8646	0.8737	0.8691	95	0.0	0.0	0.0	3	0.0	0.0	0.0	1	0.7635	0.7667	0.7651	240	0.7884	0.7340	0.7602	0.9566
0.1146	3.0	1472	0.1437	0.6364	0.875	0.7368	8	0.0	0.0	0.0	2	0.9286	0.8125	0.8667	16	0.6	0.25	0.3529	12	0.6429	0.5294	0.5806	17	1.0	0.5	0.6667	6	1.0	0.5	0.6667	4	1.0	1.0	1.0	2	0.8	0.9263	0.8585	95	0.0	0.0	0.0	3	0.0	0.0	0.0	1	0.7443	0.8125	0.7769	240	0.7612	0.7931	0.7768	0.9584
0.0842	3.99	1960	0.1428	0.6364	0.875	0.7368	8	0.0	0.0	0.0	2	0.9286	0.8125	0.8667	16	0.6	0.25	0.3529	12	0.6364	0.4118	0.5	17	1.0	0.6667	0.8	6	1.0	0.5	0.6667	4	1.0	1.0	1.0	2	0.8018	0.9368	0.8641	95	0.0	0.0	0.0	3	0.0	0.0	0.0	1	0.7589	0.8	0.7789	240	0.7724	0.7857	0.7790	0.9589

Framework versions

• Transformers 4.33.2
• Pytorch 2.0.1+cu117
• Datasets 2.14.5
• Tokenizers 0.13.3