ModernBERT-large-llm-router

This model is a fine-tuned version of answerdotai/ModernBERT-large on DevQuasar/llm_router_dataset-synth.

It achieves the following results on the test set:

• Loss: 0.0536
• F1: 0.9933

Model description

See original answerdotai/ModernBERT-large model card for additional information. This model is intended to classify queries for LLM routing. More advanced queries get labeled 1 for large_llm and simpler queries get 0 for small_llm.

Training procedure

Annotated training procedure available in this notebook. Methodology and code credits to Phillip Schmid from his Fine-tune classifier with ModernBERT in 2025 blog post.

Training hyperparameters

The following hyperparameters were used during training:

• learning_rate: 5e-05
• train_batch_size: 32
• eval_batch_size: 16
• seed: 42
• 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
• num_epochs: 5

Training results

Training Loss	Epoch	Step	Validation Loss	F1
0.0303	1.0	479	0.0317	0.9881
0.014	2.0	958	0.0374	0.9927
0.0044	3.0	1437	0.0502	0.9921
0.0004	4.0	1916	0.0554	0.9927
0.0003	5.0	2395	0.0536	0.9933

Framework versions

• Transformers 4.48.0.dev0
• Pytorch 2.5.1+cu124
• Datasets 3.1.0
• Tokenizers 0.21.0