Luganda_speech_to_intent_multilingual_xlsr

This model is a fine-tuned version of KasuleTrevor/wav2vec2-xls-r-300m-multilingual_filtered-yogera-v3 on an unknown dataset. It achieves the following results on the evaluation set:

Model description

More information needed

More information needed

More information needed

The following hyperparameters were used during training:

learning_rate: 0.0001
train_batch_size: 32
eval_batch_size: 8
seed: 42
gradient_accumulation_steps: 2
total_train_batch_size: 64
optimizer: Use 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: 30
mixed_precision_training: Native AMP

Training Loss	Epoch	Step	Validation Loss	Accuracy	Precision	Recall	F1
2.9358	1.0	131	2.2976	0.5239	0.4807	0.5239	0.4429
1.913	2.0	262	0.1712	0.9870	0.9873	0.9870	0.9870
0.3859	3.0	393	0.0777	0.9902	0.9904	0.9902	0.9902
0.1297	4.0	524	0.0721	0.9902	0.9904	0.9902	0.9902
0.1239	5.0	655	0.0669	0.9892	0.9893	0.9892	0.9892
0.1111	6.0	786	0.0633	0.9902	0.9904	0.9902	0.9902
0.0805	7.0	917	0.0574	0.9902	0.9904	0.9902	0.9902
0.0726	8.0	1048	0.0629	0.9913	0.9915	0.9913	0.9913
0.0602	9.0	1179	0.0646	0.9870	0.9872	0.9870	0.9869
0.0544	10.0	1310	0.0659	0.9892	0.9893	0.9892	0.9891
0.0472	11.0	1441	0.0639	0.9892	0.9893	0.9892	0.9891
0.0346	12.0	1572	0.0626	0.9913	0.9915	0.9913	0.9913
0.0338	13.0	1703	0.0720	0.9881	0.9885	0.9881	0.9881
0.0283	14.0	1834	0.0665	0.9913	0.9915	0.9913	0.9913
0.0236	15.0	1965	0.0711	0.9913	0.9915	0.9913	0.9913
0.0192	16.0	2096	0.0683	0.9913	0.9915	0.9913	0.9913
0.0232	17.0	2227	0.0637	0.9913	0.9915	0.9913	0.9913
0.0186	18.0	2358	0.0674	0.9902	0.9904	0.9902	0.9902

Safetensors

Model size

0.3B params

Tensor type

F32

Base model

Finetuned

Finetuned

(3)

this model