README.md

metadata

license: apache-2.0
base_model: facebook/wav2vec2-xls-r-300m
tags:
  - generated_from_trainer
datasets:
  - timit_asr
metrics:
  - wer
model-index:
  - name: timit-xls-r-300m
    results:
      - task:
          name: Automatic Speech Recognition
          type: automatic-speech-recognition
        dataset:
          name: timit_asr
          type: timit_asr
          config: clean
          split: None
          args: clean
        metrics:
          - name: Wer
            type: wer
            value: 0.2466404796361381

timit-xls-r-300m

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

• Loss: 0.4457
• Wer: 0.2466

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: 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: 1000
• num_epochs: 50
• mixed_precision_training: Native AMP

Training results

Training Loss	Epoch	Step	Validation Loss	Wer
4.0049	1.72	500	1.9735	1.0655
1.033	3.45	1000	0.6172	0.5115
0.4499	5.17	1500	0.5231	0.4395
0.2551	6.9	2000	0.4768	0.3772
0.1724	8.62	2500	0.4699	0.3626
0.133	10.34	3000	0.4346	0.3329
0.1082	12.07	3500	0.4479	0.3163
0.0886	13.79	4000	0.4393	0.3167
0.0766	15.52	4500	0.4920	0.3100
0.0637	17.24	5000	0.4510	0.3013
0.0607	18.97	5500	0.4284	0.2808
0.0495	20.69	6000	0.4270	0.2820
0.0479	22.41	6500	0.4294	0.2852
0.0444	24.14	7000	0.4456	0.2816
0.0378	25.86	7500	0.4236	0.2763
0.0325	27.59	8000	0.4365	0.2849
0.031	29.31	8500	0.4482	0.2862
0.0285	31.03	9000	0.4388	0.2691
0.0252	32.76	9500	0.4253	0.2692
0.0229	34.48	10000	0.4598	0.2641
0.0223	36.21	10500	0.4462	0.2533
0.0188	37.93	11000	0.4350	0.2673
0.0163	39.66	11500	0.4460	0.2608
0.0167	41.38	12000	0.4441	0.2683
0.0138	43.1	12500	0.4290	0.2528
0.0127	44.83	13000	0.4360	0.2508
0.0124	46.55	13500	0.4406	0.2511
0.0107	48.28	14000	0.4482	0.2477
0.0108	50.0	14500	0.4457	0.2466

Framework versions

• Transformers 4.39.3
• Pytorch 2.0.1+cu117
• Datasets 2.18.0
• Tokenizers 0.15.2