hubert-base-ls960-finetuned-ie

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:

• Accuracy: 0.6130
• Loss: 1.1686

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: 16
• eval_batch_size: 16
• seed: 42
• gradient_accumulation_steps: 4
• total_train_batch_size: 64
• optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
• lr_scheduler_type: linear
• lr_scheduler_warmup_ratio: 0.1
• num_epochs: 15

Training results

Training Loss	Epoch	Step	Accuracy	Validation Loss
1.3145	1.0	51	0.3647	1.2880
1.137	2.0	102	0.4316	1.1491
1.0227	3.0	153	0.5829	0.9724
0.9822	4.0	204	0.5645	0.9873
0.9084	5.0	255	0.5742	1.0029
0.8217	6.0	306	0.5887	1.0273
0.779	7.0	357	0.6120	0.9774
0.7444	8.0	408	0.6208	1.0336
0.6894	9.0	459	0.6140	0.9925
0.6486	10.0	510	0.6043	1.0733
0.6669	11.0	561	0.6305	1.0746
0.6184	12.0	612	0.6072	1.1670
0.5231	13.0	663	0.6052	1.1792
0.5381	14.0	714	0.6198	1.1432
0.5251	15.0	765	0.6130	1.1686

Framework versions

• Transformers 4.26.1
• Pytorch 1.13.0
• Datasets 2.10.1
• Tokenizers 0.13.2