w2v-bert-cv-grain-lg_both

This model is a fine-tuned version of facebook/w2v-bert-2.0 on the None dataset. It achieves the following results on the evaluation set:

• Loss: 16.2243
• Wer: 1.0
• Cer: 1.0

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: 4
• seed: 42
• gradient_accumulation_steps: 2
• total_train_batch_size: 16
• optimizer: Use adamw_torch with betas=(0.9,0.999) and epsilon=1e-08 and optimizer_args=No additional optimizer arguments
• lr_scheduler_type: linear
• num_epochs: 100
• mixed_precision_training: Native AMP

Training results

Training Loss	Epoch	Step	Validation Loss	Wer	Cer
0.4609	1.0	5406	0.1400	0.1423	0.0296
0.2829	2.0	10812	0.1133	0.0968	0.0213
0.2369	3.0	16218	0.1033	0.0883	0.0193
0.2106	4.0	21624	0.0848	0.0681	0.0162
0.197	5.0	27030	0.0871	0.0681	0.0159
0.2459	6.0	32436	0.1335	0.1022	0.0203
0.3563	7.0	37842	0.1809	0.1254	0.0267
0.6033	8.0	43248	0.5575	0.7032	0.1768
4.656	9.0	48654	16.9063	0.9980	0.9837
10.5595	10.0	54060	12.4706	1.0	1.0
17.1148	11.0	59466	16.2280	1.0	1.0
17.4223	12.0	64872	16.2273	1.0	1.0
17.4172	13.0	70278	16.2222	1.0	1.0
17.4159	14.0	75684	16.2243	1.0	1.0

Framework versions

• Transformers 4.46.1
• Pytorch 2.1.0+cu118
• Datasets 3.1.0
• Tokenizers 0.20.1