rubert-tiny2-odonata-f3-ner

This model is a fine-tuned version of cointegrated/rubert-tiny2 on the None dataset. It achieves the following results on the evaluation set:

• Loss: 0.0188
• Precision: 0.6653
• Recall: 0.6157
• F1: 0.6395
• Accuracy: 0.9944

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: 16
• eval_batch_size: 16
• seed: 42
• optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
• lr_scheduler_type: linear
• num_epochs: 20

Training results

Training Loss	Epoch	Step	Validation Loss	Precision	Recall	F1	Accuracy
No log	1.0	32	0.1309	0.0	0.0	0.0	0.9903
No log	2.0	64	0.0672	0.0	0.0	0.0	0.9903
No log	3.0	96	0.0623	0.0	0.0	0.0	0.9903
No log	4.0	128	0.0576	0.0	0.0	0.0	0.9903
No log	5.0	160	0.0488	0.0	0.0	0.0	0.9903
No log	6.0	192	0.0353	0.0	0.0	0.0	0.9903
No log	7.0	224	0.0288	0.7921	0.5529	0.6513	0.9935
No log	8.0	256	0.0256	0.7987	0.4824	0.6015	0.9931
No log	9.0	288	0.0235	0.7975	0.5098	0.6220	0.9933
No log	10.0	320	0.0221	0.7310	0.5647	0.6372	0.9938
No log	11.0	352	0.0212	0.6912	0.5529	0.6144	0.9938
No log	12.0	384	0.0205	0.6746	0.5529	0.6078	0.9937
No log	13.0	416	0.0201	0.6774	0.5765	0.6229	0.9938
No log	14.0	448	0.0196	0.6712	0.5843	0.6247	0.9940
No log	15.0	480	0.0194	0.6581	0.6039	0.6299	0.9941
0.0722	16.0	512	0.0192	0.6681	0.6	0.6322	0.9942
0.0722	17.0	544	0.0190	0.6624	0.6078	0.6339	0.9943
0.0722	18.0	576	0.0189	0.6542	0.6157	0.6343	0.9943
0.0722	19.0	608	0.0188	0.6624	0.6157	0.6382	0.9944
0.0722	20.0	640	0.0188	0.6653	0.6157	0.6395	0.9944

Framework versions

• Transformers 4.41.2
• Pytorch 2.3.1+cpu
• Datasets 2.19.2
• Tokenizers 0.19.1