README.md

metadata

license: apache-2.0
base_model: facebook/hubert-base-ls960
tags:
  - generated_from_trainer
metrics:
  - accuracy
  - f1
  - recall
  - precision
model-index:
  - name: hubert-base-ls960-finetuned-common_voice
    results: []

hubert-base-ls960-finetuned-common_voice

This model is a fine-tuned version of facebook/hubert-base-ls960 on the None dataset. It achieves the following results on the evaluation set:

• Loss: 0.4616
• Accuracy: 0.9375
• F1: 0.9377
• Recall: 0.9375
• Precision: 0.9403
• Mcc: 0.9225
• Auc: 0.9925

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: 3e-05
• train_batch_size: 8
• eval_batch_size: 8
• seed: 42
• gradient_accumulation_steps: 16
• total_train_batch_size: 128
• optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
• lr_scheduler_type: linear
• lr_scheduler_warmup_ratio: 0.1
• num_epochs: 10
• mixed_precision_training: Native AMP

Training results

Training Loss	Epoch	Step	Validation Loss	Accuracy	F1	Recall	Precision	Mcc	Auc
1.601	0.96	12	1.5594	0.385	0.3217	0.3850	0.6064	0.2666	0.7895
1.5467	2.0	25	1.3344	0.67	0.6516	0.6700	0.7185	0.6030	0.9009
1.4062	2.96	37	1.0521	0.8	0.7964	0.8	0.8014	0.7521	0.9436
1.0881	4.0	50	0.8340	0.8525	0.8502	0.8525	0.8677	0.8201	0.9759
0.9348	4.96	62	0.7227	0.89	0.8894	0.89	0.8939	0.8639	0.9801
0.8596	6.0	75	0.5873	0.9275	0.9276	0.9275	0.9300	0.9100	0.9908
0.7917	6.96	87	0.5208	0.93	0.9298	0.93	0.9310	0.9128	0.9940
0.6721	8.0	100	0.4784	0.9475	0.9476	0.9475	0.9491	0.9348	0.9935
0.6297	8.96	112	0.4734	0.9325	0.9326	0.9325	0.9363	0.9166	0.9916
0.6127	9.6	120	0.4616	0.9375	0.9377	0.9375	0.9403	0.9225	0.9925

Framework versions

• Transformers 4.40.2
• Pytorch 2.2.1+cu121
• Datasets 2.19.1
• Tokenizers 0.19.1