hubert-classifier

This model is a fine-tuned version of facebook/hubert-base-ls960 on an unknown dataset. It achieves the following results on the evaluation set:

• Loss: 3.9330
• Accuracy: 0.0674
• Precision: 0.0116
• Recall: 0.0674
• F1: 0.0182
• Binary: 0.3423

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: 32
• eval_batch_size: 32
• seed: 42
• gradient_accumulation_steps: 4
• total_train_batch_size: 128
• optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
• lr_scheduler_type: linear
• num_epochs: 10
• mixed_precision_training: Native AMP

Training results

Training Loss	Epoch	Step	Validation Loss	Accuracy	Precision	Recall	F1	Binary
No log	0.96	50	4.4099	0.0647	0.0191	0.0647	0.0221	0.2396
No log	1.91	100	4.3523	0.0593	0.0190	0.0593	0.0194	0.3019
No log	2.87	150	4.2416	0.0701	0.0246	0.0701	0.0235	0.3358
No log	3.83	200	4.1412	0.0701	0.0265	0.0701	0.0214	0.3437
No log	4.78	250	4.0716	0.0593	0.0069	0.0593	0.0122	0.3334
No log	5.74	300	4.0195	0.0701	0.0124	0.0701	0.0186	0.3453
No log	6.7	350	3.9850	0.0593	0.0073	0.0593	0.0126	0.3350
No log	7.66	400	3.9610	0.0647	0.0097	0.0647	0.0162	0.3388
No log	8.61	450	3.9420	0.0674	0.0113	0.0674	0.0180	0.3396
4.2019	9.57	500	3.9330	0.0674	0.0116	0.0674	0.0182	0.3423

Framework versions

• Transformers 4.38.2
• Pytorch 2.3.0
• Datasets 2.19.1
• Tokenizers 0.15.1