metadata

license: apache-2.0
base_model: facebook/wav2vec2-large-xlsr-53
tags:
  - generated_from_trainer
metrics:
  - wer
model-index:
  - name: wav2vec2-xlsr-53-ft-btb-ccv-cy
    results: []

wav2vec2-xlsr-53-ft-btb-ccv-cy

This model is a fine-tuned version of facebook/wav2vec2-large-xlsr-53 on an unknown dataset. It achieves the following results on the evaluation set:

Loss: 0.4290
Wer: 0.3378

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: 0.0003
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_steps: 500
training_steps: 2600
mixed_precision_training: Native AMP

Training results

Training Loss	Epoch	Step	Validation Loss	Wer
No log	0.0774	100	3.5325	1.0
No log	0.1549	200	2.9652	1.0
No log	0.2323	300	2.8521	1.0
No log	0.3097	400	1.2473	0.8265
3.7403	0.3871	500	0.9730	0.7234
3.7403	0.4646	600	0.8328	0.6178
3.7403	0.5420	700	0.7426	0.5505
3.7403	0.6194	800	0.7127	0.5540
3.7403	0.6969	900	0.6692	0.5080
0.7271	0.7743	1000	0.6376	0.5256
0.7271	0.8517	1100	0.6119	0.4706
0.7271	0.9292	1200	0.5987	0.4651
0.7271	1.0066	1300	0.5614	0.4267
0.7271	1.0840	1400	0.5463	0.4229
0.5511	1.1614	1500	0.5232	0.4079
0.5511	1.2389	1600	0.5185	0.4029
0.5511	1.3163	1700	0.5090	0.4042
0.5511	1.3937	1800	0.4785	0.3851
0.5511	1.4712	1900	0.4775	0.3803
0.4529	1.5486	2000	0.4677	0.3722
0.4529	1.6260	2100	0.4574	0.3544
0.4529	1.7034	2200	0.4473	0.3562
0.4529	1.7809	2300	0.4437	0.3470
0.4529	1.8583	2400	0.4353	0.3450
0.4149	1.9357	2500	0.4300	0.3401
0.4149	2.0132	2600	0.4290	0.3378

Framework versions

Transformers 4.40.2
Pytorch 2.3.0+cu121
Datasets 2.19.1
Tokenizers 0.19.1