hubert-base-ls960-finetuned-ic-slurp-no-pretrain

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: 3.2153
• Accuracy: 0.2587

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: 5e-05
• train_batch_size: 24
• eval_batch_size: 24
• seed: 42
• gradient_accumulation_steps: 4
• total_train_batch_size: 96
• optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
• lr_scheduler_type: linear
• lr_scheduler_warmup_ratio: 0.1
• num_epochs: 50

Training results

Training Loss	Epoch	Step	Validation Loss	Accuracy
3.9001	1.0	527	3.9015	0.0736
3.8219	2.0	1055	3.8454	0.0766
3.7453	3.0	1582	3.7615	0.0837
3.7202	4.0	2110	3.7143	0.0912
3.649	5.0	2637	3.6899	0.0868
3.6459	6.0	3165	3.6261	0.1077
3.5103	7.0	3692	3.5303	0.1216
3.4177	8.0	4220	3.4234	0.1503
3.3008	9.0	4747	3.3969	0.1586
3.0881	10.0	5275	3.2262	0.1993
2.9312	11.0	5802	3.1606	0.2214
2.7669	12.0	6330	3.1171	0.2364
2.5412	13.0	6857	3.1180	0.2495
2.4121	14.0	7385	3.1714	0.2458
2.1346	15.0	7912	3.2153	0.2587
2.0515	16.0	8440	3.3048	0.2564
1.7885	17.0	8967	3.3968	0.2558
1.6461	18.0	9495	3.5184	0.2511
1.4339	19.0	10022	3.7439	0.2549
1.2975	20.0	10550	3.8629	0.2549

Framework versions

• Transformers 4.37.2
• Pytorch 2.2.0+cu121
• Datasets 2.17.1
• Tokenizers 0.15.2