all_mini_lm_text_classifier_331h_whisper

This model is a fine-tuned version of KasuleTrevor/all-MiniLM-L6-v2_intents on the None dataset. It achieves the following results on the evaluation set:

• Loss: 0.3747
• Accuracy: 0.927
• Precision: 0.932
• Recall: 0.927
• F1: 0.927

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: 0.0001
• train_batch_size: 8
• eval_batch_size: 8
• seed: 42
• gradient_accumulation_steps: 2
• total_train_batch_size: 16
• 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_steps: 500
• num_epochs: 20
• mixed_precision_training: Native AMP

Training results

Training Loss	Epoch	Step	Validation Loss	Accuracy	Precision	Recall	F1
2.7544	1.0	232	1.9815	0.441	0.327	0.441	0.316
1.58	2.0	464	0.9619	0.693	0.72	0.693	0.631
0.8075	3.0	696	0.4160	0.914	0.904	0.914	0.906
0.4276	4.0	928	0.2829	0.935	0.925	0.935	0.928
0.2676	5.0	1160	0.2485	0.928	0.919	0.928	0.921
0.2006	6.0	1392	0.2550	0.941	0.945	0.941	0.942
0.1547	7.0	1624	0.2331	0.951	0.955	0.951	0.952
0.1236	8.0	1856	0.2301	0.953	0.955	0.953	0.953
0.0985	9.0	2088	0.2265	0.948	0.951	0.948	0.948
0.0789	10.0	2320	0.2813	0.941	0.944	0.941	0.942
0.0649	11.0	2552	0.2906	0.943	0.947	0.943	0.944
0.0536	12.0	2784	0.2782	0.948	0.951	0.948	0.948
0.0444	13.0	3016	0.3242	0.932	0.937	0.932	0.932
0.0394	14.0	3248	0.2991	0.937	0.94	0.937	0.937
0.0301	15.0	3480	0.3409	0.928	0.934	0.928	0.929
0.0247	16.0	3712	0.3552	0.93	0.935	0.93	0.93
0.0205	17.0	3944	0.3668	0.93	0.935	0.93	0.93
0.0162	18.0	4176	0.3730	0.927	0.932	0.927	0.927
0.0168	19.0	4408	0.3745	0.927	0.932	0.927	0.927
0.014	20.0	4640	0.3747	0.927	0.932	0.927	0.927

Framework versions

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