tiny-wav2vec2-no-tokenizer-ft-keyword-spotting

This model is a fine-tuned version of patrickvonplaten/tiny-wav2vec2-no-tokenizer on the Panga-Azazia/Bambara-Keyword-Spotting dataset. It achieves the following results on the evaluation set:

• Loss: 0.6901
• Accuracy: 0.7143

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: 32
• eval_batch_size: 32
• seed: 0
• gradient_accumulation_steps: 4
• total_train_batch_size: 128
• 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: linear
• lr_scheduler_warmup_ratio: 0.1
• num_epochs: 15.0
• mixed_precision_training: Native AMP

Training results

Training Loss	Epoch	Step	Validation Loss	Accuracy
No log	1.0	3	0.6930	0.4643
No log	2.0	6	0.6923	0.5
No log	3.0	9	0.6917	0.5714
0.5373	4.0	12	0.6911	0.6071
0.5373	5.0	15	0.6906	0.6429
0.5373	6.0	18	0.6901	0.7143
0.5357	7.0	21	0.6898	0.7143
0.5357	8.0	24	0.6895	0.6786
0.5357	9.0	27	0.6893	0.6786
0.4832	10.0	30	0.6890	0.6786
0.4832	11.0	33	0.6889	0.6786
0.4832	12.0	36	0.6888	0.6786
0.4832	13.0	39	0.6886	0.6786
0.5343	14.0	42	0.6886	0.6786
0.5343	15.0	45	0.6885	0.6786

Framework versions

• Transformers 4.52.4
• Pytorch 2.6.0+cu124
• Datasets 3.6.0
• Tokenizers 0.21.1