README.md

metadata

tags:
  - generated_from_trainer
model-index:
  - name: rubert-base-srl-seqlabeling
    results: []

rubert-base-srl-seqlabeling

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

• Loss: 0.2417
• Predicate Precision: 0.9323
• Predicate Recall: 0.9612
• Predicate F1: 0.9466
• Predicate Number: 129
• Инструмент Precision: 0.0
• Инструмент Recall: 0.0
• Инструмент F1: 0.0
• Инструмент Number: 1
• Каузатор Precision: 0.7667
• Каузатор Recall: 0.6301
• Каузатор F1: 0.6917
• Каузатор Number: 73
• Экспериенцер Precision: 0.6939
• Экспериенцер Recall: 0.8293
• Экспериенцер F1: 0.7556
• Экспериенцер Number: 41
• Overall Precision: 0.8430
• Overall Recall: 0.8361
• Overall F1: 0.8395
• Overall Accuracy: 0.9584

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: 16
• eval_batch_size: 8
• seed: 42
• optimizer: Adam with betas=(0.9,0.98) and epsilon=1e-06
• lr_scheduler_type: cosine
• lr_scheduler_warmup_ratio: 0.06
• num_epochs: 10.0

Training results

Training Loss	Epoch	Step	Validation Loss	Predicate Precision	Predicate Recall	Predicate F1	Predicate Number	Инструмент Precision	Инструмент Recall	Инструмент F1	Инструмент Number	Каузатор Precision	Каузатор Recall	Каузатор F1	Каузатор Number	Экспериенцер Precision	Экспериенцер Recall	Экспериенцер F1	Экспериенцер Number	Overall Precision	Overall Recall	Overall F1	Overall Accuracy
0.2462	1.0	54	0.1554	0.8897	1.0	0.9416	129	0.0	0.0	0.0	1	0.7708	0.5068	0.6116	73	0.6047	0.6341	0.6190	41	0.8136	0.7869	0.8	0.9486
0.1863	2.0	108	0.1268	0.9014	0.9922	0.9446	129	0.0	0.0	0.0	1	0.8444	0.5205	0.6441	73	0.6829	0.6829	0.6829	41	0.8509	0.7951	0.8220	0.9557
0.0668	3.0	162	0.1288	0.9338	0.9845	0.9585	129	0.0	0.0	0.0	1	0.8148	0.6027	0.6929	73	0.6957	0.7805	0.7356	41	0.8602	0.8320	0.8458	0.9600
0.039	4.0	216	0.1695	0.9007	0.9845	0.9407	129	0.0	0.0	0.0	1	0.8298	0.5342	0.6500	73	0.6441	0.9268	0.76	41	0.8259	0.8361	0.8310	0.9557
0.0187	5.0	270	0.1955	0.9323	0.9612	0.9466	129	0.0	0.0	0.0	1	0.75	0.5753	0.6512	73	0.7105	0.6585	0.6835	41	0.8502	0.7910	0.8195	0.9551
0.0216	6.0	324	0.2083	0.9394	0.9612	0.9502	129	0.0	0.0	0.0	1	0.7586	0.6027	0.6718	73	0.6829	0.6829	0.6829	41	0.8485	0.8033	0.8253	0.9562
0.0176	7.0	378	0.2203	0.9323	0.9612	0.9466	129	0.0	0.0	0.0	1	0.7273	0.6575	0.6906	73	0.68	0.8293	0.7473	41	0.8273	0.8443	0.8357	0.9578
0.0037	8.0	432	0.2313	0.9323	0.9612	0.9466	129	0.0	0.0	0.0	1	0.7541	0.6301	0.6866	73	0.6809	0.7805	0.7273	41	0.8382	0.8279	0.8330	0.9567
0.0089	9.0	486	0.2409	0.9323	0.9612	0.9466	129	0.0	0.0	0.0	1	0.7705	0.6438	0.7015	73	0.6939	0.8293	0.7556	41	0.8436	0.8402	0.8419	0.9589
0.0043	10.0	540	0.2417	0.9323	0.9612	0.9466	129	0.0	0.0	0.0	1	0.7667	0.6301	0.6917	73	0.6939	0.8293	0.7556	41	0.8430	0.8361	0.8395	0.9584

Framework versions

• Transformers 4.13.0.dev0
• Pytorch 1.10.0+cu102
• Datasets 1.15.1
• Tokenizers 0.10.3