mms-1b-zdj

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: 0.7884
• Wer: 0.3296

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: 4
• eval_batch_size: 16
• seed: 42
• optimizer: Use OptimizerNames.ADAMW_TORCH_FUSED with betas=(0.9,0.999) and epsilon=1e-08 and optimizer_args=No additional optimizer arguments
• lr_scheduler_type: linear
• lr_scheduler_warmup_steps: 500
• num_epochs: 4
• mixed_precision_training: Native AMP

Training results

Training Loss	Epoch	Step	Validation Loss	Wer
5.8521	0.6623	500	1.1876	0.6160
1.0851	1.3245	1000	0.9700	0.4729
0.9955	1.9868	1500	0.8846	0.4126
0.8827	2.6490	2000	0.8372	0.3677
0.8029	3.3113	2500	0.8070	0.3437
0.7793	3.9735	3000	0.7883	0.3298
0.7793	4.0	3020	0.7884	0.3296

Framework versions

• Transformers 5.13.0.dev0
• Pytorch 2.10.0+cu128
• Datasets 5.0.0
• Tokenizers 0.22.2