spellcorrector_1709_v7

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: 0.0885
• Precision: 0.9610
• Recall: 0.9739
• F1: 0.9674
• Accuracy: 0.9742

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
0.2583	1.0	1951	0.2210	0.8849	0.9710	0.9260	0.9443
0.2244	2.0	3902	0.1992	0.8954	0.9685	0.9305	0.9466
0.2099	3.0	5853	0.1868	0.9021	0.9684	0.9341	0.9482
0.1989	4.0	7804	0.1789	0.9109	0.9634	0.9364	0.9500
0.1926	5.0	9755	0.1694	0.9141	0.9652	0.9389	0.9518
0.1837	6.0	11706	0.1605	0.9160	0.9697	0.9421	0.9540
0.1817	7.0	13657	0.1547	0.9209	0.9656	0.9427	0.9549
0.1723	8.0	15608	0.1484	0.9226	0.9687	0.9451	0.9564
0.1692	9.0	17559	0.1442	0.9269	0.9649	0.9455	0.9571
0.1642	10.0	19510	0.1386	0.9269	0.9697	0.9478	0.9587
0.1578	11.0	21461	0.1327	0.9310	0.9693	0.9497	0.9600
0.148	12.0	23412	0.1231	0.9393	0.9674	0.9532	0.9630
0.1455	13.0	25363	0.1172	0.9413	0.9711	0.9560	0.9653
0.1402	14.0	27314	0.1123	0.9476	0.9673	0.9573	0.9662
0.1323	15.0	29265	0.1056	0.9511	0.9694	0.9602	0.9685
0.1269	16.0	31216	0.0989	0.9521	0.9740	0.9629	0.9709
0.1225	17.0	33167	0.0953	0.9575	0.9716	0.9645	0.9720
0.1186	18.0	35118	0.0907	0.9582	0.9744	0.9662	0.9735
0.1169	19.0	37069	0.0897	0.9606	0.9734	0.9670	0.9738
0.113	20.0	39020	0.0885	0.9610	0.9739	0.9674	0.9742

Framework versions

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