router-mmBERT-base-v2-text-only

This model is a fine-tuned version of jhu-clsp/mmBERT-base on an unknown dataset. It achieves the following results on the evaluation set:

• Loss: 0.8326
• Accuracy: 0.6875
• Precision: 0.6827
• Recall: 0.6875
• F1: 0.6846

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: 0.0001
• train_batch_size: 8
• eval_batch_size: 32
• seed: 42
• gradient_accumulation_steps: 4
• total_train_batch_size: 32
• 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: cosine
• num_epochs: 5

Training results

Training Loss	Epoch	Step	Validation Loss	Accuracy	Precision	Recall	F1
2.6118	0.4545	20	0.6795	0.6193	0.7300	0.6193	0.6174
2.4002	0.9091	40	0.5451	0.7386	0.7327	0.7386	0.7270
2.1361	1.3636	60	0.8168	0.6705	0.6832	0.6705	0.6747
2.589	1.8182	80	0.5943	0.6364	0.6591	0.6364	0.6426
2.8187	2.2727	100	0.8248	0.5852	0.7404	0.5852	0.5725
2.08	2.7273	120	0.5875	0.7386	0.7409	0.7386	0.7166
1.1623	3.1818	140	0.6528	0.7159	0.7080	0.7159	0.7084
1.5022	3.6364	160	0.6610	0.6932	0.6828	0.6932	0.6834
0.898	4.0909	180	0.7076	0.6705	0.6683	0.6705	0.6693
0.7971	4.5455	200	0.8023	0.7045	0.6969	0.7045	0.6983
0.582	5.0	220	0.8326	0.6875	0.6827	0.6875	0.6846

Framework versions

• Transformers 4.57.1
• Pytorch 2.8.0+cu128
• Datasets 4.2.0
• Tokenizers 0.22.1