spellcorrector_1709_v6

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

• Loss: 0.0158
• Precision: 0.9871
• Recall: 0.9831
• F1: 0.9851
• Accuracy: 0.9951

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: 4
• eval_batch_size: 4
• 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
0.2794	1.0	1951	0.2144	0.7935	0.7534	0.7729	0.9414
0.2179	2.0	3902	0.1587	0.8333	0.7994	0.8160	0.9555
0.1752	3.0	5853	0.1272	0.8639	0.8239	0.8434	0.9642
0.1507	4.0	7804	0.1063	0.8844	0.8510	0.8674	0.9699
0.1273	5.0	9755	0.0859	0.9044	0.8753	0.8896	0.9755
0.1106	6.0	11706	0.0762	0.9197	0.8838	0.9014	0.9779
0.0984	7.0	13657	0.0670	0.9255	0.9035	0.9144	0.9805
0.0861	8.0	15608	0.0571	0.9402	0.9150	0.9274	0.9832
0.0754	9.0	17559	0.0517	0.9469	0.9269	0.9368	0.9844
0.07	10.0	19510	0.0442	0.9559	0.9361	0.9459	0.9866
0.0636	11.0	21461	0.0387	0.9604	0.9477	0.9540	0.9882
0.0597	12.0	23412	0.0315	0.9673	0.9559	0.9616	0.9903
0.0523	13.0	25363	0.0296	0.9723	0.9612	0.9667	0.9909
0.0449	14.0	27314	0.0258	0.9744	0.9683	0.9713	0.9920
0.0407	15.0	29265	0.0226	0.9797	0.9715	0.9756	0.9930
0.0395	16.0	31216	0.0212	0.9817	0.9758	0.9788	0.9934
0.0357	17.0	33167	0.0181	0.9843	0.9783	0.9813	0.9943
0.0356	18.0	35118	0.0172	0.9859	0.9812	0.9836	0.9947
0.0328	19.0	37069	0.0162	0.9865	0.9829	0.9847	0.9950
0.0316	20.0	39020	0.0158	0.9871	0.9831	0.9851	0.9951

Framework versions

• Transformers 4.28.0
• Pytorch 2.0.1+cu118
• Datasets 2.14.5
• Tokenizers 0.13.3