rubert-tiny2-ner-finetuned

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: 0.3182
• Precision: 0.8107
• Recall: 0.8650
• F1: 0.8370
• Accuracy: 0.9176

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: Use OptimizerNames.ADAMW_TORCH with betas=(0.9,0.999) and epsilon=1e-08 and optimizer_args=No additional optimizer arguments
• lr_scheduler_type: linear
• num_epochs: 15

Training results

Training Loss	Epoch	Step	Validation Loss	Precision	Recall	F1	Accuracy
No log	1.0	50	1.6494	0.0901	0.0341	0.0494	0.4330
No log	2.0	100	1.1966	0.4701	0.4223	0.4449	0.6464
No log	3.0	150	0.8823	0.5788	0.6122	0.5950	0.7567
No log	4.0	200	0.6968	0.6465	0.7151	0.6791	0.8211
No log	5.0	250	0.5777	0.7052	0.7696	0.7360	0.8569
No log	6.0	300	0.4985	0.7534	0.8129	0.7820	0.8863
No log	7.0	350	0.4437	0.7764	0.8305	0.8026	0.8958
No log	8.0	400	0.4044	0.7862	0.8413	0.8129	0.9030
No log	9.0	450	0.3762	0.7953	0.8498	0.8216	0.9080
0.8776	10.0	500	0.3559	0.8003	0.8558	0.8271	0.9105
0.8776	11.0	550	0.3407	0.7995	0.8562	0.8269	0.9131
0.8776	12.0	600	0.3295	0.8055	0.8594	0.8316	0.9155
0.8776	13.0	650	0.3232	0.8059	0.8618	0.8329	0.9161
0.8776	14.0	700	0.3196	0.8088	0.8610	0.8341	0.9170
0.8776	15.0	750	0.3182	0.8107	0.8650	0.8370	0.9176

Framework versions

• Transformers 4.50.0
• Pytorch 2.6.0
• Datasets 3.5.0
• Tokenizers 0.21.1