mms_severity_classifier

This model is a fine-tuned version of facebook/mms-1b-all on an unknown dataset. It achieves the following results on the evaluation set:

• Loss: 0.0834
• Accuracy: 0.9787
• F1 Macro: 0.9750
• F1 Weighted: 0.9786
• F1 Mild: 0.9581
• F1 Moderate: 0.9812
• F1 Normal: 0.9719
• F1 Severe: 0.9886

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: 3e-05
• train_batch_size: 8
• eval_batch_size: 8
• seed: 42
• gradient_accumulation_steps: 4
• total_train_batch_size: 32
• optimizer: Use 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_ratio: 0.1
• num_epochs: 10

Training results

Training Loss	Epoch	Step	Validation Loss	Accuracy	F1 Macro	F1 Weighted	F1 Mild	F1 Moderate	F1 Normal	F1 Severe
0.3997	1.0	1210	0.3725	0.8652	0.8510	0.8637	0.7821	0.8791	0.8618	0.8809
0.1947	2.0	2420	0.2191	0.9193	0.9067	0.9195	0.8594	0.9330	0.8828	0.9517
0.1382	3.0	3630	0.1758	0.9433	0.9336	0.9432	0.8950	0.9519	0.9185	0.9690
0.1204	4.0	4840	0.1432	0.9573	0.9517	0.9575	0.9365	0.9653	0.9324	0.9726
0.0815	5.0	6050	0.0922	0.9726	0.9699	0.9727	0.9639	0.9748	0.9567	0.9841
0.0647	6.0	7260	0.0896	0.9753	0.9710	0.9753	0.9561	0.9819	0.9632	0.9829
0.0864	7.0	8470	0.1191	0.9726	0.9699	0.9726	0.9581	0.9734	0.9652	0.9829
0.0243	8.0	9680	0.0870	0.9787	0.9776	0.9787	0.9821	0.9810	0.9635	0.9840
0.0455	9.0	10890	0.0909	0.9760	0.9715	0.9759	0.9501	0.9795	0.9701	0.9863
0.0452	10.0	12100	0.0834	0.9787	0.9750	0.9786	0.9581	0.9812	0.9719	0.9886

Framework versions

• Transformers 4.57.6
• Pytorch 2.8.0+cu128
• Datasets 4.5.0
• Tokenizers 0.22.2