Modelpublic

dhasmana/Angika-Chhattisgarhi-Kumaoni-Magadhi-Garhwali-Bhojpuri-w2v-bert-2.0

sourceHugging Facemitupdated 8mo agoView on Hugging Face

0likes8downloads

Model Card

Angika-Chhattisgarhi-Kumaoni-Magadhi-Garhwali-Bhojpuri-w2v-bert-2.0

This model is a fine-tuned version of facebook/w2v-bert-2.0 on the None dataset. It achieves the following results on the evaluation set:

—Loss: 2.2516
—Cer: 0.1901

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
—trainbatchsize: 16
—evalbatchsize: 8
—seed: 42
—gradientaccumulationsteps: 2
—totaltrainbatch_size: 32
—optimizer: Use OptimizerNames.ADAMWTORCHFUSED with betas=(0.9,0.999) and epsilon=1e-08 and optimizer_args=No additional optimizer arguments
—lrschedulertype: linear
—lrschedulerwarmup_steps: 500
—num_epochs: 10
—mixedprecisiontraining: Native AMP

Training results

Training Loss	Epoch	Step	Validation Loss	Cer
5.4668	0.5021	300	1.0444	0.2297
1.7381	1.0033	600	0.9856	0.2355
1.2914	1.5054	900	0.8753	0.2055
1.0549	2.0067	1200	0.9217	0.1924
0.7774	2.5088	1500	0.9469	0.1940
0.6229	3.0100	1800	1.0222	0.1919
0.4280	3.5121	2100	1.0945	0.1952
0.3410	4.0134	2400	1.1561	0.1967
0.2191	4.5155	2700	1.3199	0.1977
0.1739	5.0167	3000	1.4493	0.1956
0.1106	5.5188	3300	1.4783	0.1950
0.0941	6.0201	3600	1.6971	0.1938
0.0602	6.5222	3900	1.7262	0.1969
0.0489	7.0234	4200	1.9145	0.1934
0.0314	7.5255	4500	1.8897	0.1925
0.0238	8.0268	4800	2.0750	0.1915
0.0142	8.5289	5100	2.0837	0.1914
0.0090	9.0301	5400	2.1940	0.1906
0.0052	9.5322	5700	2.2516	0.1901

Framework versions

—Transformers 5.0.0
—Pytorch 2.9.0+cu128
—Datasets 4.0.0
—Tokenizers 0.22.2