metadata

license: apache-2.0
base_model: facebook/wav2vec2-xls-r-300m
tags:
  - generated_from_trainer
datasets:
  - common_voice_15_0
metrics:
  - wer
model-index:
  - name: wav2vec2-xls-r-300m-br
    results:
      - task:
          name: Automatic Speech Recognition
          type: automatic-speech-recognition
        dataset:
          name: common_voice_15_0
          type: common_voice_15_0
          config: br
          split: None
          args: br
        metrics:
          - name: Wer
            type: wer
            value: 49.79811574697174

wav2vec2-xls-r-300m-br

This model is a fine-tuned version of facebook/wav2vec2-xls-r-300m on the common_voice_15_0 dataset. It achieves the following results on the evaluation set:

Loss: 0.8887
Wer: 49.7981
Cer: 17.3877

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: 5e-05
train_batch_size: 8
eval_batch_size: 8
seed: 42
gradient_accumulation_steps: 2
total_train_batch_size: 16
optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
lr_scheduler_type: linear
lr_scheduler_warmup_steps: 500
num_epochs: 40
mixed_precision_training: Native AMP

Training results

Training Loss	Epoch	Step	Validation Loss	Wer	Cer
5.1153	2.18	1000	2.8854	100.0	100.0
1.4117	4.36	2000	0.9161	71.2786	25.3180
0.7888	6.54	3000	0.7753	62.7456	22.0767
0.6316	8.71	4000	0.7550	58.1786	20.5383
0.5434	10.89	5000	0.7508	56.5096	20.1168
0.4672	13.07	6000	0.7844	54.9125	19.3835
0.4237	15.25	7000	0.7786	53.2705	18.5765
0.3899	17.43	8000	0.8050	53.0552	18.6105
0.3607	19.61	9000	0.8280	51.9874	18.3024
0.3355	21.79	10000	0.7967	51.5388	17.9811
0.3098	23.97	11000	0.8296	51.2876	17.9547
0.2937	26.14	12000	0.8544	50.9915	17.7827
0.2793	28.32	13000	0.8909	51.5478	18.1286
0.2641	30.5	14000	0.8740	50.4800	17.6561
0.2552	32.68	15000	0.8832	49.9776	17.4463
0.2467	34.86	16000	0.8753	50.3096	17.4765
0.2378	37.04	17000	0.8895	49.8789	17.3952
0.2337	39.22	18000	0.8887	49.7981	17.3877

Framework versions

Transformers 4.39.1
Pytorch 2.0.1+cu117
Datasets 2.18.0
Tokenizers 0.15.2