eternis_router_encoder_sft_11Sep

This model is a fine-tuned version of answerdotai/ModernBERT-base on the None dataset. It achieves the following results on the evaluation set:

• Loss: 0.4944
• Model Accuracy: 0.7861

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: 32
• eval_batch_size: 64
• seed: 42
• gradient_accumulation_steps: 2
• total_train_batch_size: 64
• 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: cosine
• lr_scheduler_warmup_ratio: 0.01
• num_epochs: 20

Training results

Training Loss	Epoch	Step	Validation Loss	Model Accuracy
0.535	1.0602	300	0.5205	0.7730
0.5295	2.1204	600	0.5053	0.7764
0.5166	3.1805	900	0.5038	0.7811
0.5138	4.2407	1200	0.5009	0.7807
0.5196	5.3009	1500	0.5048	0.7761
0.4977	6.3611	1800	0.5034	0.7795
0.4924	7.4212	2100	0.4970	0.7803
0.4915	8.4814	2400	0.5084	0.7749
0.4903	9.5416	2700	0.4960	0.7830
0.4942	10.6018	3000	0.4945	0.7873
0.4926	11.6619	3300	0.4955	0.7807
0.5057	12.7221	3600	0.4929	0.7826
0.4916	13.7823	3900	0.4939	0.7838
0.4953	14.8425	4200	0.4948	0.7807
0.4919	15.9027	4500	0.4947	0.7834
0.5043	16.9628	4800	0.4944	0.7854
0.4768	18.0212	5100	0.4944	0.7857
0.4879	19.0814	5400	0.4944	0.7861

Framework versions

• Transformers 4.56.1
• Pytorch 2.8.0+cu128
• Datasets 4.0.0
• Tokenizers 0.22.0