wav2vec2-large-mms-1b-chich-colab

This model is a fine-tuned version of facebook/mms-1b-all on the None dataset. It achieves the following results on the evaluation set:

• Loss: inf
• Wer: 0.6583

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: 0.001
• train_batch_size: 8
• eval_batch_size: 8
• seed: 42
• optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
• lr_scheduler_type: linear
• lr_scheduler_warmup_steps: 100
• num_epochs: 8
• mixed_precision_training: Native AMP

Training results

Training Loss	Epoch	Step	Validation Loss	Wer
6.9371	0.44	100	inf	0.7521
1.4647	0.88	200	inf	0.7314
1.2928	1.32	300	inf	0.7388
1.3529	1.75	400	inf	0.7320
1.1927	2.19	500	inf	0.6984
1.34	2.63	600	inf	0.6992
1.2707	3.07	700	inf	0.6922
1.105	3.51	800	inf	0.7185
1.1334	3.95	900	inf	0.6838
1.0915	4.39	1000	inf	0.6751
1.133	4.82	1100	inf	0.6934
1.1295	5.26	1200	inf	0.6752
1.0214	5.7	1300	inf	0.6700
0.9687	6.14	1400	inf	0.6692
1.0553	6.58	1500	inf	0.6664
0.9686	7.02	1600	inf	0.6618
0.9565	7.46	1700	inf	0.6605
1.001	7.89	1800	inf	0.6583

Framework versions

• Transformers 4.35.0.dev0
• Pytorch 2.1.0+cu118
• Datasets 2.14.6
• Tokenizers 0.14.1