test-avatar-02

This model is a fine-tuned version of facebook/wav2vec2-xls-r-300m on the None dataset. It achieves the following results on the evaluation set:

• Loss: 0.1194
• Wer: 2.0537

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: 0.001
• train_batch_size: 16
• eval_batch_size: 8
• seed: 42
• gradient_accumulation_steps: 2
• total_train_batch_size: 32
• optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
• lr_scheduler_type: linear
• lr_scheduler_warmup_steps: 1227
• num_epochs: 15
• mixed_precision_training: Native AMP

Training results

Training Loss	Epoch	Step	Validation Loss	Wer
2.2743	1.0	1227	0.1604	2.5532
0.01	2.0	2454	-0.0493	2.2163
-0.1007	3.0	3681	-0.0902	2.0385
-0.1512	4.0	4908	-0.1111	2.0659
-0.1885	5.0	6135	-0.1170	2.0161
-0.2193	6.0	7362	-0.1221	2.0787
-0.2472	7.0	8589	-0.1163	2.0158
-0.2727	8.0	9816	-0.1060	1.9938
-0.2973	9.0	11043	-0.0917	2.0298
-0.3204	10.0	12270	-0.0744	2.0447
-0.3419	11.0	13497	-0.0437	2.0499
-0.3629	12.0	14724	-0.0118	2.0490
-0.3823	13.0	15951	0.0212	2.0427
-0.3997	14.0	17178	0.0738	2.0484
-0.4138	15.0	18405	0.1194	2.0537

Framework versions

• Transformers 4.36.0.dev0
• Pytorch 1.12.1+cu113
• Datasets 2.14.7
• Tokenizers 0.15.0