metadata

library_name: transformers
license: apache-2.0
base_model: openai/whisper-small
tags:
  - generated_from_trainer
datasets:
  - balbus-classifier
metrics:
  - accuracy
  - precision
  - recall
  - f1
model-index:
  - name: miosipof/whisper-small-ft-balbus-sep28k-v1.5
    results:
      - task:
          name: Audio Classification
          type: audio-classification
        dataset:
          name: Apple dataset
          type: balbus-classifier
          config: default
          split: train
          args: default
        metrics:
          - name: Accuracy
            type: accuracy
            value:
              accuracy: 0.8111877154497023
          - name: Precision
            type: precision
            value:
              precision: 0.8133174791914387
          - name: Recall
            type: recall
            value:
              recall: 0.7365398420674802
          - name: F1
            type: f1
            value:
              f1: 0.7730269353927294

miosipof/whisper-small-ft-balbus-sep28k-v1.5

This model is a fine-tuned version of openai/whisper-small on the Apple dataset dataset. It achieves the following results on the evaluation set:

Loss: 0.1083
Accuracy: {'accuracy': 0.8111877154497023}
Precision: {'precision': 0.8133174791914387}
Recall: {'recall': 0.7365398420674802}
F1: {'f1': 0.7730269353927294}

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-06
train_batch_size: 16
eval_batch_size: 8
seed: 42
gradient_accumulation_steps: 2
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: 0.5
training_steps: 1000
mixed_precision_training: Native AMP

Training results

Training Loss	Epoch	Step	Validation Loss	Accuracy	Precision	Recall	F1
0.1718	0.1253	100	0.1705	{'accuracy': 0.564243183954873}	{'precision': 0.6190476190476191}	{'recall': 0.00466618808327351}	{'f1': 0.009262557890986818}
0.1683	0.2506	200	0.1653	{'accuracy': 0.6118771544970228}	{'precision': 0.7677642980935875}	{'recall': 0.15900933237616655}	{'f1': 0.26345524829021705}
0.1595	0.3759	300	0.1494	{'accuracy': 0.6847383265434033}	{'precision': 0.6486175115207373}	{'recall': 0.6062455132806892}	{'f1': 0.6267161410018552}
0.1299	0.5013	400	0.1266	{'accuracy': 0.7608900031338138}	{'precision': 0.7008928571428571}	{'recall': 0.7889447236180904}	{'f1': 0.7423167848699763}
0.1174	0.6266	500	0.1140	{'accuracy': 0.7977123158884363}	{'precision': 0.7800674409891345}	{'recall': 0.747307968413496}	{'f1': 0.7633363886342804}
0.1117	0.7519	600	0.1155	{'accuracy': 0.7919147602632404}	{'precision': 0.7362281270252754}	{'recall': 0.8155061019382628}	{'f1': 0.773841961852861}
0.1072	0.8772	700	0.1074	{'accuracy': 0.8096208085239737}	{'precision': 0.8282490597576264}	{'recall': 0.7114142139267767}	{'f1': 0.765398725622707}
0.106	1.0025	800	0.1078	{'accuracy': 0.8077405202130994}	{'precision': 0.8175152749490835}	{'recall': 0.7203876525484566}	{'f1': 0.7658843732112193}
0.1001	1.1278	900	0.1079	{'accuracy': 0.810404261986838}	{'precision': 0.8174858984689767}	{'recall': 0.7282842785355348}	{'f1': 0.7703113135914958}
0.092	1.2531	1000	0.1083	{'accuracy': 0.8111877154497023}	{'precision': 0.8133174791914387}	{'recall': 0.7365398420674802}	{'f1': 0.7730269353927294}

Framework versions

Transformers 4.45.2
Pytorch 2.2.0
Datasets 3.2.0
Tokenizers 0.20.3