wav2vec2-XLS-R-Fleurs-demo-google-colab-Ezra_William_Prod8

This model is a fine-tuned version of facebook/wav2vec2-large-xlsr-53 on the xtreme_s dataset. It achieves the following results on the evaluation set:

• Loss: 0.9283
• Wer: 0.4204

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: 8
• eval_batch_size: 8
• seed: 42
• gradient_accumulation_steps: 4
• 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: 600
• num_epochs: 180
• mixed_precision_training: Native AMP

Training results

Training Loss	Epoch	Step	Validation Loss	Wer
4.7974	15.58	300	2.8438	1.0
1.0661	31.17	600	0.7520	0.5781
0.1697	46.75	900	0.8342	0.5181
0.1081	62.34	1200	0.8508	0.5010
0.0833	77.92	1500	0.9216	0.5014
0.0616	93.51	1800	0.9659	0.4859
0.0463	109.09	2100	0.9606	0.4646
0.0356	124.68	2400	0.9255	0.4578
0.0291	140.26	2700	0.9889	0.4503
0.0218	155.84	3000	0.9336	0.4371
0.0182	171.43	3300	0.9283	0.4204

Framework versions

• Transformers 4.39.1
• Pytorch 2.2.1+cu121
• Datasets 2.18.0
• Tokenizers 0.15.2