--- license: apache-2.0 base_model: facebook/wav2vec2-large tags: - generated_from_trainer metrics: - wer model-index: - name: wav2vec2-large-sw-cv-20hr-v1 results: [] --- [Visualize in Weights & Biases](https://wandb.ai/asr-africa-research-team/ASR%20Africa/runs/w0c4nymx) # wav2vec2-large-sw-cv-20hr-v1 This model is a fine-tuned version of [facebook/wav2vec2-large](https://huggingface.co/facebook/wav2vec2-large) on an unknown dataset. It achieves the following results on the evaluation set: - Loss: inf - Model Preparation Time: 0.0059 - Wer: 0.3464 - Cer: 0.1302 ## 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: 32 - eval_batch_size: 8 - seed: 42 - gradient_accumulation_steps: 2 - total_train_batch_size: 64 - optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08 - lr_scheduler_type: cosine - lr_scheduler_warmup_steps: 500 - num_epochs: 100 - mixed_precision_training: Native AMP ### Training results | Training Loss | Epoch | Step | Validation Loss | Model Preparation Time | Wer | Cer | |:-------------:|:-------:|:-----:|:---------------:|:----------------------:|:------:|:------:| | 4.0167 | 0.9976 | 210 | 1.3039 | 0.0059 | 0.9301 | 0.3195 | | 0.7784 | 2.0 | 421 | 0.7306 | 0.0059 | 0.5818 | 0.1788 | | 0.5359 | 2.9976 | 631 | 0.5886 | 0.0059 | 0.5048 | 0.1517 | | 0.427 | 4.0 | 842 | 0.5274 | 0.0059 | 0.4460 | 0.1345 | | 0.3657 | 4.9976 | 1052 | 0.5617 | 0.0059 | 0.4620 | 0.1430 | | 0.3219 | 6.0 | 1263 | 0.5162 | 0.0059 | 0.408 | 0.1240 | | 0.2922 | 6.9976 | 1473 | 0.4861 | 0.0059 | 0.4074 | 0.1256 | | 0.2681 | 8.0 | 1684 | 0.5076 | 0.0059 | 0.404 | 0.1253 | | 0.2459 | 8.9976 | 1894 | 0.5042 | 0.0059 | 0.3915 | 0.1205 | | 0.2332 | 10.0 | 2105 | 0.5051 | 0.0059 | 0.3706 | 0.1120 | | 0.2181 | 10.9976 | 2315 | 0.5370 | 0.0059 | 0.3750 | 0.1149 | | 0.2073 | 12.0 | 2526 | 0.5231 | 0.0059 | 0.3860 | 0.1249 | | 0.1982 | 12.9976 | 2736 | 0.5290 | 0.0059 | 0.4045 | 0.1239 | | 0.1875 | 14.0 | 2947 | 0.5184 | 0.0059 | 0.3755 | 0.1153 | | 0.1782 | 14.9976 | 3157 | 0.5215 | 0.0059 | 0.3587 | 0.1100 | | 0.1684 | 16.0 | 3368 | 0.5395 | 0.0059 | 0.371 | 0.1142 | | 0.1629 | 16.9976 | 3578 | 0.5499 | 0.0059 | 0.3608 | 0.1101 | | 0.1563 | 18.0 | 3789 | 0.5478 | 0.0059 | 0.3577 | 0.1107 | | 0.1516 | 18.9976 | 3999 | 0.5290 | 0.0059 | 0.3649 | 0.1148 | | 0.1431 | 20.0 | 4210 | 0.5765 | 0.0059 | 0.3657 | 0.1167 | | 0.1366 | 20.9976 | 4420 | 0.5604 | 0.0059 | 0.3617 | 0.1137 | | 0.1345 | 22.0 | 4631 | 0.5546 | 0.0059 | 0.3604 | 0.1118 | | 0.1303 | 22.9976 | 4841 | 0.5284 | 0.0059 | 0.3511 | 0.1089 | | 0.122 | 24.0 | 5052 | 0.5668 | 0.0059 | 0.3555 | 0.1111 | | 0.1183 | 24.9976 | 5262 | 0.5874 | 0.0059 | 0.3521 | 0.1088 | | 0.1151 | 26.0 | 5473 | 0.5539 | 0.0059 | 0.3379 | 0.1044 | | 0.1108 | 26.9976 | 5683 | 0.6110 | 0.0059 | 0.3375 | 0.1051 | | 0.1089 | 28.0 | 5894 | 0.5582 | 0.0059 | 0.3397 | 0.1029 | | 0.1064 | 28.9976 | 6104 | 0.5774 | 0.0059 | 0.3432 | 0.1062 | | 0.1026 | 30.0 | 6315 | 0.6042 | 0.0059 | 0.3420 | 0.1062 | | 0.0983 | 30.9976 | 6525 | 0.5793 | 0.0059 | 0.3402 | 0.1046 | | 0.0952 | 32.0 | 6736 | 0.6083 | 0.0059 | 0.3423 | 0.1074 | | 0.0927 | 32.9976 | 6946 | 0.6015 | 0.0059 | 0.3363 | 0.1035 | | 0.0895 | 34.0 | 7157 | 0.5790 | 0.0059 | 0.3368 | 0.1041 | | 0.0889 | 34.9976 | 7367 | 0.5530 | 0.0059 | 0.3338 | 0.1023 | | 0.0865 | 36.0 | 7578 | 0.5598 | 0.0059 | 0.3267 | 0.1009 | | 0.0828 | 36.9976 | 7788 | 0.5699 | 0.0059 | 0.3249 | 0.1001 | | 0.0814 | 38.0 | 7999 | 0.5756 | 0.0059 | 0.3237 | 0.0996 | | 0.0819 | 38.9976 | 8209 | 0.5878 | 0.0059 | 0.3363 | 0.1052 | | 0.077 | 40.0 | 8420 | 0.5852 | 0.0059 | 0.3216 | 0.0984 | | 0.075 | 40.9976 | 8630 | 0.5940 | 0.0059 | 0.3295 | 0.1022 | | 0.0725 | 42.0 | 8841 | 0.5779 | 0.0059 | 0.3219 | 0.0997 | | 0.0701 | 42.9976 | 9051 | 0.5962 | 0.0059 | 0.3144 | 0.0965 | | 0.0693 | 44.0 | 9262 | 0.6192 | 0.0059 | 0.317 | 0.0975 | | 0.0659 | 44.9976 | 9472 | 0.5989 | 0.0059 | 0.3126 | 0.0964 | | 0.0662 | 46.0 | 9683 | 0.6069 | 0.0059 | 0.3112 | 0.0975 | | 0.0646 | 46.9976 | 9893 | 0.6309 | 0.0059 | 0.3164 | 0.0986 | | 0.0626 | 48.0 | 10104 | 0.6266 | 0.0059 | 0.3199 | 0.1007 | | 0.062 | 48.9976 | 10314 | 0.6403 | 0.0059 | 0.3116 | 0.0963 | | 0.0591 | 50.0 | 10525 | 0.6140 | 0.0059 | 0.3133 | 0.0965 | | 0.0568 | 50.9976 | 10735 | 0.5947 | 0.0059 | 0.3078 | 0.0950 | | 0.0538 | 52.0 | 10946 | 0.6202 | 0.0059 | 0.3029 | 0.0939 | | 0.0544 | 52.9976 | 11156 | 0.6215 | 0.0059 | 0.312 | 0.0966 | | 0.0526 | 54.0 | 11367 | 0.6637 | 0.0059 | 0.3093 | 0.0959 | | 0.05 | 54.9976 | 11577 | 0.6513 | 0.0059 | 0.3079 | 0.0955 | | 0.0518 | 56.0 | 11788 | 0.6611 | 0.0059 | 0.3070 | 0.0948 | | 0.0493 | 56.9976 | 11998 | 0.6415 | 0.0059 | 0.3041 | 0.0941 | | 0.0482 | 58.0 | 12209 | 0.6386 | 0.0059 | 0.3042 | 0.0939 | | 0.0461 | 58.9976 | 12419 | 0.6664 | 0.0059 | 0.316 | 0.0995 | | 0.0445 | 60.0 | 12630 | 0.6472 | 0.0059 | 0.3057 | 0.0963 | | 0.0449 | 60.9976 | 12840 | 0.6510 | 0.0059 | 0.3103 | 0.0972 | | 0.0437 | 62.0 | 13051 | 0.6696 | 0.0059 | 0.3166 | 0.1005 | ### Framework versions - Transformers 4.43.1 - Pytorch 2.2.0+cu121 - Datasets 2.20.0 - Tokenizers 0.19.1