my_fyp_test13

This model is a fine-tuned version of facebook/wav2vec2-base on the None dataset. It achieves the following results on the evaluation set:

• Loss: 2.8594
• Accuracy: 0.17

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: 64
• eval_batch_size: 64
• seed: 42
• gradient_accumulation_steps: 8
• total_train_batch_size: 512
• optimizer: Use OptimizerNames.ADAMW_TORCH with betas=(0.9,0.999) and epsilon=1e-08 and optimizer_args=No additional optimizer arguments
• lr_scheduler_type: cosine
• lr_scheduler_warmup_ratio: 0.05
• num_epochs: 30

Training results

Training Loss	Epoch	Step	Validation Loss	Accuracy
No log	1.0	7	3.8041	0.0638
3.5516	2.0	14	3.5967	0.0625
3.329	3.0	21	3.4894	0.0625
3.329	4.0	28	3.4042	0.07
2.9364	5.0	35	3.3234	0.065
3.0836	6.0	42	3.2488	0.07
3.0836	7.0	49	3.1819	0.07
2.7451	8.0	56	3.1250	0.07
2.8969	9.0	63	3.0800	0.07
2.6078	10.0	70	3.0414	0.07
2.6078	11.0	77	3.0133	0.0838
2.7883	12.0	84	2.9887	0.0963
2.7495	13.0	91	2.9683	0.105
2.7495	14.0	98	2.9482	0.13
2.4974	15.0	105	2.9343	0.1237
2.6843	16.0	112	2.9164	0.135
2.6843	17.0	119	2.9068	0.145
2.4408	18.0	126	2.8867	0.16
2.6254	19.0	133	2.8676	0.1675
2.4018	20.0	140	2.8673	0.1737
2.4018	21.0	147	2.8651	0.1713
2.5887	22.0	154	2.8591	0.1688
2.5823	23.0	161	2.8593	0.1725
2.5823	24.0	168	2.8578	0.1675
2.3684	25.0	175	2.8593	0.1713
2.571	25.8	180	2.8594	0.17

Framework versions

• Transformers 4.50.2
• Pytorch 2.6.0+cu124
• Datasets 3.5.0
• Tokenizers 0.21.1