metadata

license: apache-2.0
base_model: facebook/hubert-base-ls960
tags:
  - generated_from_trainer
datasets:
  - audiofolder
metrics:
  - accuracy
model-index:
  - name: urdu-emotions-hubert-large-Emotion
    results:
      - task:
          name: Audio Classification
          type: audio-classification
        dataset:
          name: audiofolder
          type: audiofolder
          config: default
          split: train
          args: default
        metrics:
          - name: Accuracy
            type: accuracy
            value: 0.8166666666666667

urdu-emotions-hubert-large-Emotion

This model is a fine-tuned version of facebook/hubert-base-ls960 on the audiofolder dataset. It achieves the following results on the evaluation set:

Loss: 1.3374
Accuracy: 0.8167

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: 4
total_train_batch_size: 32
optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
lr_scheduler_type: linear
lr_scheduler_warmup_ratio: 0.2
num_epochs: 60

Training results

Training Loss	Epoch	Step	Validation Loss	Accuracy
No log	0.9333	7	0.9539	0.7833
0.1337	2.0	15	0.8845	0.8167
0.1048	2.9333	22	1.0824	0.75
0.0406	4.0	30	1.0672	0.8
0.0406	4.9333	37	1.2542	0.75
0.0931	6.0	45	0.9778	0.8
0.0646	6.9333	52	1.2557	0.75
0.0966	8.0	60	1.0541	0.7833
0.0966	8.9333	67	1.5521	0.75
0.1301	10.0	75	0.9688	0.8333
0.2704	10.9333	82	1.2517	0.7667
0.1625	12.0	90	1.3938	0.7667
0.1625	12.9333	97	1.4804	0.7667
0.1278	14.0	105	0.9219	0.8167
0.2052	14.9333	112	1.2735	0.75
0.2487	16.0	120	1.0251	0.7833
0.2487	16.9333	127	1.1808	0.8
0.1784	18.0	135	1.2522	0.7333
0.2182	18.9333	142	0.8958	0.8333
0.1688	20.0	150	1.1747	0.75
0.1688	20.9333	157	1.3938	0.8
0.2948	22.0	165	0.6410	0.8833
0.0945	22.9333	172	0.8846	0.8333
0.0738	24.0	180	0.7653	0.8333
0.0738	24.9333	187	0.7587	0.8333
0.0909	26.0	195	1.1861	0.8
0.0721	26.9333	202	0.8185	0.8333
0.1215	28.0	210	1.4169	0.7333
0.1215	28.9333	217	1.1844	0.8
0.0454	30.0	225	1.1273	0.7833
0.0915	30.9333	232	1.3536	0.8
0.0274	32.0	240	1.1561	0.7667
0.0274	32.9333	247	1.2680	0.7833
0.0251	34.0	255	1.3334	0.8
0.1263	34.9333	262	1.2555	0.8167
0.0389	36.0	270	1.0567	0.8
0.0389	36.9333	277	1.5755	0.7667
0.109	38.0	285	1.5332	0.7667
0.0599	38.9333	292	1.0758	0.85
0.0064	40.0	300	1.1251	0.85
0.0064	40.9333	307	1.3546	0.8
0.003	42.0	315	1.4129	0.8
0.0303	42.9333	322	1.3925	0.8
0.0016	44.0	330	1.3129	0.7833
0.0016	44.9333	337	1.2522	0.8
0.0308	46.0	345	1.3130	0.8167
0.002	46.9333	352	1.3005	0.8333
0.003	48.0	360	1.3434	0.8
0.003	48.9333	367	1.3762	0.8
0.0024	50.0	375	1.4090	0.8
0.0778	50.9333	382	1.3769	0.8167
0.0024	52.0	390	1.3748	0.8167
0.0024	52.9333	397	1.3649	0.8167
0.0132	54.0	405	1.3365	0.8167
0.0102	54.9333	412	1.3363	0.8167
0.0012	56.0	420	1.3374	0.8167

Framework versions

Transformers 4.40.2
Pytorch 2.2.1+cu121
Datasets 2.19.1
Tokenizers 0.19.1