wav2vec2-large-xlsr-53-dialect-classifier

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

• Loss: 1.3693
• Accuracy: 0.3605

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: 1e-05
• train_batch_size: 4
• eval_batch_size: 16
• seed: 42
• gradient_accumulation_steps: 4
• total_train_batch_size: 16
• optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
• lr_scheduler_type: linear
• lr_scheduler_warmup_ratio: 0.1
• num_epochs: 30

Training results

Training Loss	Epoch	Step	Validation Loss	Accuracy
1.3719	0.98	21	1.3693	0.3605
1.371	2.0	43	1.3652	0.3605
1.3638	2.98	64	1.3588	0.3605
1.3571	4.0	86	1.3508	0.3605
1.3446	4.98	107	1.3440	0.3605
1.3399	6.0	129	1.3388	0.3605
1.3303	6.98	150	1.3340	0.3605
1.3305	8.0	172	1.3305	0.3605
1.334	8.98	193	1.3267	0.3605
1.3303	10.0	215	1.3232	0.3605
1.3241	10.98	236	1.3205	0.3605
1.317	12.0	258	1.3183	0.3605
1.3239	12.98	279	1.3161	0.3605
1.3169	14.0	301	1.3144	0.3605
1.3129	14.98	322	1.3128	0.3605
1.3112	16.0	344	1.3115	0.3605
1.3231	16.98	365	1.3107	0.3605
1.3115	18.0	387	1.3099	0.3605
1.3082	18.98	408	1.3092	0.3605
1.3228	20.0	430	1.3085	0.3605
1.2872	20.98	451	1.3080	0.3605
1.2744	22.0	473	1.3075	0.3605
1.2835	22.98	494	1.3070	0.3605
1.3213	24.0	516	1.3068	0.3605
1.3112	24.98	537	1.3065	0.3605
1.3047	26.0	559	1.3062	0.3605
1.29	26.98	580	1.3061	0.3605
1.3114	28.0	602	1.3060	0.3605
1.3123	28.98	623	1.3060	0.3605
1.3217	29.3	630	1.3060	0.3605

Framework versions

• Transformers 4.38.2
• Pytorch 2.0.1+cu118
• Datasets 2.18.0
• Tokenizers 0.15.2