<!-- This model card has been generated automatically according to the information the Trainer had access to. You should probably proofread and complete it, then remove this comment. -->
wav2vec2-10epochs-3e5
This model is a fine-tuned version of facebook/wav2vec2-large-xlsr-53 on the None dataset. It achieves the following results on the evaluation set:
- Loss: 0.3076
- Wer: 0.2157
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: 4
- eval_batch_size: 4
- seed: 42
- gradient_accumulation_steps: 2
- total_train_batch_size: 8
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
- lr_scheduler_type: linear
- lr_scheduler_warmup_steps: 500
- num_epochs: 10
- mixed_precision_training: Native AMP
Training results
Training Loss | Epoch | Step | Validation Loss | Wer |
---|---|---|---|---|
0.1184 | 0.36 | 100 | 0.4169 | 0.3008 |
0.0747 | 0.72 | 200 | 0.3413 | 0.2677 |
0.046 | 1.08 | 300 | 0.3245 | 0.2496 |
0.0263 | 1.44 | 400 | 0.3094 | 0.2450 |
0.0326 | 1.8 | 500 | 0.3031 | 0.2366 |
0.0179 | 2.16 | 600 | 0.2966 | 0.2324 |
0.0209 | 2.52 | 700 | 0.2920 | 0.2287 |
0.0095 | 2.88 | 800 | 0.2966 | 0.2267 |
0.0113 | 3.24 | 900 | 0.2997 | 0.2259 |
0.0107 | 3.6 | 1000 | 0.3011 | 0.2221 |
0.009 | 3.96 | 1100 | 0.3050 | 0.2219 |
0.0131 | 4.32 | 1200 | 0.3034 | 0.2195 |
0.0147 | 4.68 | 1300 | 0.3042 | 0.2204 |
0.0195 | 5.04 | 1400 | 0.3187 | 0.2204 |
0.0087 | 5.4 | 1500 | 0.3205 | 0.2198 |
0.0083 | 5.76 | 1600 | 0.3177 | 0.2197 |
0.016 | 6.12 | 1700 | 0.3110 | 0.2175 |
0.0152 | 6.47 | 1800 | 0.3078 | 0.2171 |
0.0173 | 6.83 | 1900 | 0.3134 | 0.2169 |
0.046 | 7.19 | 2000 | 0.3134 | 0.2164 |
0.0264 | 7.55 | 2100 | 0.3095 | 0.2174 |
0.0143 | 7.91 | 2200 | 0.3066 | 0.2170 |
0.0186 | 8.27 | 2300 | 0.3086 | 0.2181 |
0.0308 | 8.63 | 2400 | 0.3103 | 0.2175 |
0.0246 | 8.99 | 2500 | 0.3083 | 0.2157 |
0.0126 | 9.35 | 2600 | 0.3073 | 0.2154 |
0.0134 | 9.71 | 2700 | 0.3076 | 0.2157 |
Framework versions
- Transformers 4.11.3
- Pytorch 1.10.1
- Datasets 2.9.0
- Tokenizers 0.10.3