wav2vec2-large-xls-r-300m-test1yakutsi-colab

This model is a fine-tuned version of facebook/wav2vec2-xls-r-300m on the common_voice_17_0 dataset. It achieves the following results on the evaluation set:

• Loss: 0.4509
• Wer: 0.4245

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: 0.0002
• 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: 250
• num_epochs: 10
• mixed_precision_training: Native AMP

Training results

Training Loss	Epoch	Step	Validation Loss	Wer
3.1089	1.1707	120	2.9271	1.0
2.1217	2.3415	240	0.8076	0.7261
0.5442	3.5122	360	0.4935	0.5490
0.3041	4.6829	480	0.4464	0.4832
0.2184	5.8537	600	0.4263	0.4554
0.1675	7.0244	720	0.4416	0.4488
0.138	8.1951	840	0.4512	0.4380
0.1167	9.3659	960	0.4509	0.4245

Framework versions

• Transformers 4.41.0.dev0
• Pytorch 2.2.1+cu121
• Datasets 2.19.1
• Tokenizers 0.19.1