mms-oral-cancer1

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: 0.7520
• Wer: 0.2858

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: 0.0004
• train_batch_size: 4
• eval_batch_size: 2
• seed: 42
• gradient_accumulation_steps: 4
• total_train_batch_size: 16
• optimizer: Use OptimizerNames.ADAMW_TORCH 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: 40
• num_epochs: 10
• mixed_precision_training: Native AMP

Training results

Training Loss	Epoch	Step	Validation Loss	Wer
1.0076	0.9825	28	0.8407	0.3384
0.8473	1.9825	56	0.7604	0.3173
0.755	2.9825	84	0.7681	0.3114
0.7099	3.9825	112	0.7701	0.3093
0.7202	4.9825	140	0.7671	0.2989
0.6763	5.9825	168	0.7696	0.2923
0.6529	6.9825	196	0.7441	0.2917
0.6359	7.9825	224	0.7506	0.2917
0.6049	8.9825	252	0.7640	0.2854
0.5831	9.9825	280	0.7520	0.2858

Framework versions

• PEFT 0.14.0
• Transformers 4.51.1
• Pytorch 2.5.1+cu124
• Datasets 3.5.0
• Tokenizers 0.21.0