Modelpublic

SucharithaS/wav2vec2-large-xls-r-300m-dm32

sourceHugging Faceapache-2.0updated 5mo agoView on Hugging Face

0likes6downloads

Model Card

wav2vec2-large-xls-r-300m-dm32

This model is a fine-tuned version of facebook/wav2vec2-xls-r-300m on an unknown dataset. It achieves the following results on the evaluation set:

—Loss: 0.6008
—Accuracy: 0.7292
—F1: 0.7300
—Precision: 0.7315
—Recall: 0.7292

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: 0.0001
—trainbatchsize: 4
—evalbatchsize: 4
—seed: 42
—gradientaccumulationsteps: 4
—totaltrainbatch_size: 16
—optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
—lrschedulertype: linear
—num_epochs: 22
—mixedprecisiontraining: Native AMP

Training results

Training Loss	Epoch	Step	Validation Loss	Accuracy	F1	Precision	Recall
No log	2.39	34	0.6920	0.4167	0.2451	0.1736	0.4167
No log	4.77	68	0.6586	0.6875	0.6268	0.7965	0.6875
No log	7.16	102	0.7262	0.4167	0.2451	0.1736	0.4167
0.6954	9.54	136	0.6758	0.5833	0.4298	0.3403	0.5833
0.6954	11.93	170	0.5389	0.7917	0.7834	0.8032	0.7917
0.6954	14.32	204	0.6160	0.7292	0.7309	0.7486	0.7292
0.6119	16.7	238	0.5537	0.7708	0.7534	0.8045	0.7708
0.6119	19.09	272	0.5732	0.7292	0.7300	0.7315	0.7292
0.6119	21.47	306	0.6008	0.7292	0.7300	0.7315	0.7292

Framework versions

—Transformers 4.37.2
—Pytorch 2.10.0+cu128
—Datasets 4.0.0
—Tokenizers 0.15.2